Thursday, May 3, 2007

Linux Tips!!

TIP 1:

Is NTP Working?

STEP 1 (Test the current server):

Try issuing the following command:

$ ntpq -pn

remote refid st t when poll reach delay offset jitter
=================================================== 16 u - 64 0 0.000 0.000 4000.00

The above is an example of a problem.
Compare it to a working configuration.

$ ntpq -pn

remote refid st t when poll reach delay offset jitter
+ 2 u 107 128 377 25.642 3.350 1.012 10 l 40 64 377 0.000 0.000 0.008
+ 3 u 34 128 377 21.138 6.118 0.398
* .USNO. 1 u 110 128 377 33.69 9.533 3.534

STEP 2 (Configure the /etc/ntp.conf):

$ cat /etc/ntp.conf

# My simple client-only ntp configuration.
# ping -a shows the IP address
# which is used in the restrict below
server # local clock
fudge stratum 10
driftfile /etc/ntp/drift
restrict default ignore
restrict mask
authenticate no

STEP 3 (Configure /etc/ntp/step-tickers):

The values for server above are placed in the "/etc/ntp/step-tickers" file

$ cat /etc/ntp/step-tickers

The startup script /etc/rc.d/init.d/ntpd will grab the servers in this
file and execute the ntpdate command as follows:

/usr/sbin/ntpdate -s -b -p 8

Why? Because if the time is off ntpd will not start. The command above set the
clock. If System Time deviates from true time by more than 1000 seconds, then,
the ntpd daemon will enter panic mode and exit.

STEP 4 (Restart the service and check):

Issue the restart command

/etc/init.d/ntpd restart

check the values for "ntpq -pn",
which should match step 1.

ntpq -pn


Time is always stored in the kernel as the number of seconds since
midnight of the 1st of January 1970 UTC, regardless of whether the
hardware clock is stored as UTC or not. Conversions to local time
are done at run-time. So, it's easy to get the time in different
timezones for only the current session as follows:

$ export TZ=EST
$ date
Mon Aug 2 10:34:04 EST 2004

$ export TZ=NET
$ date
Mon Aug 2 15:34:18 NET 2004

The following are possible values for TZ:

Hours From Greenwich Mean Time (GMT) Value Description
0 GMT Greenwich Mean Time
+1 ECT European Central Time
+2 EET European Eastern Time
+2 ART
+3 EAT Saudi Arabia
+3.5 MET Iran
+4 NET
+5 PLT West Asia
+5.5 IST India
+6 BST Central Asia
+7 VST Bangkok
+8 CTT China
+9 JST Japan
+9.5 ACT Central Australia
+10 AET Eastern Australia
+11 SST Central Pacific
+12 NST New Zealand
-11 MIT Samoa
-10 HST Hawaii
-9 AST Alaska
-8 PST Pacific Standard Time
-7 PNT Arizona
-7 MST Mountain Standard Time
-6 CST Central Standard Time
-5 EST Eastern Standard Time
-5 IET Indiana East
-4 PRT Atlantic Standard Time
-3.5 CNT Newfoundland
-3 AGT Eastern South America
-3 BET Eastern South America
-1 CAT Azores

DST timezone

0 BST for British Summer.
+400 ADT for Atlantic Daylight.
+500 EDT for Eastern Daylight.
+600 CDT for Central Daylight.
+700 MDT for Mountain Daylight.
+800 PDT for Pacific Daylight.
+900 YDT for Yukon Daylight.
+1000 HDT for Hawaii Daylight.
-100 MEST for Middle European Summer,
MESZ for Middle European Summer,
SST for Swedish Summer and FST for French Summer.
-700 WADT for West Australian Daylight.
-1000 EADT for Eastern Australian Daylight.
-1200 NZDT for New Zealand Daylight.

The following is an example of setting the TZ environment variable
for the timezone, only when timezone changes go into effect.

$ export TZ=EST+5EDT,M4.1.0/2,M10.5.0/2

Take a look at the last line "M10.5.0/2". What does it mean? Here is the

Mm.w.d This specifies day d (0 <= d <= 6) of week w (1 <= w <= 5) of
month m (1 <= m <= 12). Week 1 is the first week in which day d
occurs and week 5 is the last week in which day d occurs. Day 0
is a Sunday.

The time fields specify when, in the local time currently in
effect, the change to the other time occurs. If omitted,
the default is 02:00:00.

So this is what it means. M10 stands for October, the 5 is the fifth week
that includes a Sunday (note 0 in M10.5.0/2 is Sunday). To see that it is
the fifth week see the calendar below. The time change occurs a 2am in
the morning. (Special Note: In 2007, DST was extended. See TIP 230).

Su Mo Tu We Th Fr Sa
1 2
3 4 5 6 7 8 9
10 11 12 13 14 15 16
17 18 19 20 21 22 23
24 25 26 27 28 29 30

Prove it. Take the following program sunrise, which can calcuates sunrise
and sunset for an latitude and longitude. This program can be downloaded
from the following location:

Below is a bash script that will run the program for the next 100 days.

# program: next100days Mike Chirico
# download:
# This will calculate the sunrise and sunset for
# latitude 39.95 Note must convert to degrees
# longitude 75.15 Note must convert to degrees
for (( i=0; i <= 100; i++))
sunrise `date -d "+$i day" "+%Y %m %d"` $lat $long

Take a look at the following sample output.

$ export TZ=EST+5EDT,M4.1.0/2,M10.5.0/2
$ ./next100days

Sunrise 08-24-2004 06:21:12 Sunset 08-24-2004 19:43:42
Sunrise 08-25-2004 06:22:09 Sunset 08-25-2004 19:42:12
Sunrise 08-26-2004 06:23:06 Sunset 08-26-2004 19:40:41
Sunrise 08-27-2004 06:24:03 Sunset 08-27-2004 19:39:09
Sunrise 08-28-2004 06:25:00 Sunset 08-28-2004 19:37:37
Sunrise 08-29-2004 06:25:56 Sunset 08-29-2004 19:36:04
Sunrise 08-30-2004 06:26:53 Sunset 08-30-2004 19:34:31
Sunrise 08-31-2004 06:27:50 Sunset 08-31-2004 19:32:57
Sunrise 09-01-2004 06:28:46 Sunset 09-01-2004 19:31:22
Sunrise 09-02-2004 06:29:43 Sunset 09-02-2004 19:29:47
..[values omitted ]
Sunrise 10-28-2004 07:25:31 Sunset 10-28-2004 18:02:34
Sunrise 10-29-2004 07:26:38 Sunset 10-29-2004 18:01:19
Sunrise 10-30-2004 07:27:46 Sunset 10-30-2004 18:00:06
Sunrise 10-31-2004 06:28:53 Sunset 10-31-2004 16:58:54
Sunrise 11-01-2004 06:30:01 Sunset 11-01-2004 16:57:44
Sunrise 11-02-2004 06:31:10 Sunset 11-02-2004 16:56:35

Compare 10-30-2004 with 10-31-2004. Sunrise is an hour earlier because
daylight saving time has ended, just as predicted.

There is an easier way to switch between timezones. Take a look at the
directory zoneinfo as follows:

$ ls /usr/share/zoneinfo

Africa Chile Factory Iceland Mexico posix UCT
America CST6CDT GB Indian Mideast posixrules Universal
Antarctica Cuba GB-Eire Iran MST PRC US
Asia Egypt GMT0 Israel Navajo right WET
Atlantic Eire GMT-0 Jamaica NZ ROC W-SU
Australia EST GMT+0 Japan NZ-CHAT ROK
Brazil EST5EDT Greenwich Kwajalein Pacific Singapore Zulu
Canada Etc Hongkong Libya Poland SystemV
CET Europe HST MET Portugal Turkey

TZ can be set to any one of these files. Some of these are directories and contain
subdirectories, such as ./posix/America. This way you don not have to enter the
timezone, offset, and range for dst, since it has already been calculated.

$ export TZ=:/usr/share/zoneinfo/posix/America/Aruba
$ export TZ=:/usr/share/zoneinfo/Egypt


Also see (TIP 27).
Also see (TIP 103) using chrony which is very similiar to ntpd.
Note time settings can usually be found in /etc/sysconfig/clock

TIP 2:

cpio works like tar, only better.

STEP 1 (Create two directories with data ../dir1 an ../dir2)

mkdir -p ../dir1
mkdir -p ../dir2
cp /etc/*.conf ../dir1/.
cp /etc/*.cnf ../dir2/.

Which will backup all your cnf and conf files.

STEP 2 (Piping the files to tar)

cpio works like tar but can take input
from the "find" command.

$ find ../dir1/ | cpio -o --format=tar > test.tar
$ find ../dir1/ | cpio -o -H tar > test2.tar

Same command without the ">"

$ find ../dir1/ | cpio -o --format=tar -F test.tar
$ find ../dir1/ | cpio -o -H tar -F test2.tar

Using append

$ find ../dir1/ | cpio -o --format=tar -F test.tar
$ find ../dir2/ | cpio -o --format=tar --append -F test.tar

STEP 3 (List contents of the tar file)

$ cpio -it < test.tar
$ cpio -it -F test.tar

STEP 4 (Extract the contents)

$ cpio -i -F test.tar

TIP 3:

Working with tar. The basics with encryption.

STEP 1 (Using the tar command on the directory /stuff)

Suppose you have a directory /stuff
To tar everything in stuff to create a ".tar" file.

$ tar -cvf stuff.tar stuff

Which will create "stuff.tar".

STEP 2 (Using the tar command to create a ".tar.gz" of /stuff)

$ tar -czf stuff.tar.gz stuff

STEP 3 (List the files in the archive)

$ tar -tzf stuff.tar.gz
$ tar -tf stuff.tar

STEP 4 (A way to list specific files)

Note, pipe the results to a file and edit

$ tar -tzf stuff.tar.gz > mout

Then, edit mout to only include the files you want

$ tar -T mout -xzf stuff.tar.gz

The above command will only get the files in mout.
Of couse, if you want them all

$ tar -xzf stuff.tar.gz


$ tar -zcvf - stuff|openssl des3 -salt -k secretpassword | dd of=stuff.des3

This will create stuff.des3...don't forget the password you
put in place of secretpassword. This can be done interactively as

$ dd if=stuff.des3 |openssl des3 -d -k secretpassword|tar zxf -

NOTE: above there is a "-" at the end... this will
extract everything.

TIP 4:

Creating a Virtual File System and Mounting it with a Loopback Device.

STEP 1 (Construct a 10MB file)

$ dd if=/dev/zero of=/tmp/disk-image count=20480

By default dd uses block of 512 so the size will be 20480*512

STEP 2 (Make an ext2 or ext3 file system) -- ext2 shown here.

$ mke2fs -q

or if you want ext3

$ mkfs -t ext3 -q /tmp/disk-image

yes, you can even use reiser, but you'll need to create a bigger
disk image. Something like "dd if=/dev/zero of=/tmp/disk-image count=50480".

$ mkfs -t reiserfs -q /tmp/disk-image

Hit yes for confirmation. It only asks this because it's a file

STEP 3 (Create a directory "virtual-fs" and mount. This has to be done as root)

$ mkdir /virtual-fs
$ mount -o loop=/dev/loop0 /tmp/disk-image /virtual-fs

SPECIAL NOTE: if you mount a second device you will have to increase the
loop count: loop=/dev/loop1, loop=/dev/loop2, ... loop=/dev/loopn

Now it operates just like a disk. This virtual filesystem can be mounted
when the system boots by adding the following to the "/etc/fstab" file. Then,
to mount, just type "mount /virtual-fs".

/tmp/disk-image /virtual-fs ext2 rw,loop=/dev/loop0 0 0

STEP 4 (When done, umount it)

$ umount /virtual-fs

SPECIAL NOTE: If you are using Fedora core 2, in the /etc/fstab you can take
advantage of acl properties for this mount. Note the acl next to the
rw entry. This is shown here with ext3.

/tmp/disk-image /virtual-fs ext3 rw,acl,loop=/dev/loop1 0 0

Also, if you are using Fedora core 2 and above, you can mount the file
on a cryptoloop.

$ dd if=/dev/urandom of=disk-aes count=20480

$ modprobe loop
$ modprobe cryptoloop
$ modprobe aes

$ losetup -e aes /dev/loop0 disk-aes
$ mkfs -t ext2 /dev/loop0
$ mount -o loop,encryption=aes disk-aes

If you do not have Fedora core 2, then, you can build the kernel from source
with some of the following options (not complete, yet)

Cryptographic API Support (CONFIG_CRYPTO)
generic loop cryptographic (CONFIG_CRYPTOLOOP)
Cryptographic ciphers (CONFIG_CIPHERS)
Enable one or more ciphers (CONFIG CIPHER .*) such as AES.

HELPFUL INFORMATION: It is possible to bind mount partitions, or associate the
mounted partition to a directory name.

# mount --bind /virtual-fs /home/mchirico/vfs

Also, if you want to see what filesystems are currently mounted, "cat" the
file "/etc/mtab"

$ cat /etc/mtab

Also see TIP 91.

TIP 5:

Setting up 2 IP address on "One" NIC. This example is on ethernet.

STEP 1 (The settings for the initial IP address)

$ cat /etc/sysconfig/network-scripts/ifcfg-eth0


STEP 2 (2nd IP address: )

$ cat /etc/sysconfig/network-scripts/ifcfg-eth0:1


SUMMARY Note, in STEP 1 the filename is "ifcfg-eth0", whereas in
STEP 2 it's "ifcfg-eth0:1" and also not the matching
entries for "DEVICE=...". Also, obviously, the
"IPADDR" is different as well.

TIP 6:

Sharing Directories Among Several Users.

Several people are working on a project in "/home/share"
and they need to create documents and programs so that
others in the group can edit and execute these documents
as needed. Also see (TIP 186) for adding existing users
to groups.

$ /usr/sbin/groupadd share
$ chown -R root.share /home/share
$ /usr/bin/gpasswd -a share
$ chmod 2775 /home/share

$ ls -ld /home/share
drwxrwsr-x 2 root share 4096 Nov 8 16:19 /home/share
^---------- Note the s bit, which was set with the chmod 2775

$ cat /etc/group
... ^------- users are added to this group.

The user may need to login again to get access. Or, if the user is currently
logged in, they can run the following command:

$ su -

Note, the above step is recommended over "newgrp - share" since currently
newgrp in FC2,FC3, and FC4 gets access to the group but the umask is not
correctly formed.

As root you can test their account.

$ su - "You need to '-' to pickup thier environment '$ su - chirico' "

Note: SUID, SGID, Sticky bit. Only the left most octet is examined, and "chmod 755" is used
as an example of the full command. But, anything else could be used as well. Normally
you'd want executable permissions.

Octal digit Binary value Meaning Example usage
0 000 all cleared $ chmod 0755 or chmod 755
1 001 sticky $ chmod 1755
2 010 setgid $ chmod 2755
3 011 setgid, sticky $ chmod 3755
4 100 setuid $ chmod 4755
5 101 setuid, sticky $ chmod 5755
6 110 setuid, setgid $ chmod 6755
7 111 setuid, setgid, sticky $ chmod 7755

A few examples applied to a directory below. In the first example all users in the group can
add files to directory "dirA" and they can delete their own files. Users cannot delete other
user's files.

Sticky bit:
$ chmod 1770 dirA

Below files created within the directory have the group ID of the directory, rather than that
of the default group setting for the user who created the file.

Set group ID bit:
$ chmod 2755 dirB

TIP 7:

Getting Infomation on Commands

The "info" is a great utility for getting information about the system.
Here's a quick key on using "info" from the terminal prompt.

'q' exits.
'u' moves up to the table of contents of the current section.
'n' moves to the next chapter.
'p' moves to the previous chapter.
'space' goes into the selected section.

The following is a good starting point:

$ info coreutils

Need to find out what a certain program does?

$ whatis open
open (2) - open and possibly create a file or device
open (3) - perl pragma to set default PerlIO layers for input and output
open (3pm) - perl pragma to set default PerlIO layers for input and output
open (n) - Open a file-based or command pipeline channel

To get specific information about the open commmand

$ man 2 open

also try 'keyword' search, which is the same as the apropos command.
For example, to find all the man pages on selinux, type the following:

$ man -k selinux

or the man full word search. Same as whatis command.

$ man -f

This is a hint once you are inside man.

space moves forward one page
b moves backward
y scrolls up one line "yikes, I missed it!"
g goes to the beginning
q quits
/ search, repeat seach n
m mark, enter a letter like "a", then, ' to go back
' enter a letter that is marked.

To get section numbers

$ man 8 ping

Note the numbers are used as follows
(This is OpenBSD)

1 General Commands
2 System Calls and Error Numbers
3 C Libraries
3p perl
4 Devices and device drivers
5 File Formats and config files
6 Game instructions
7 Miscellaneous information
8 System maintenance
9 Kernel internals

To find the man page directly, "ls" command:

$ whereis -m ls
ls: /usr/share/man/man1/ls.1.gz /usr/share/man/man1/ls.1 /usr/share/man/man1p/ls.1p

To read this file directly, do the following:

$ man /usr/share/man/man1/ls.1.gz

If you want to know the manpath, execute manpath.

$ manpath

TIP 8:

How to Put a "Running Job" in the Background.

You're running a job at the terminal prompt, and it's taking
a very long time. You want to put the job in the backgroud.

"CTL - z" Temporarily suspends the job
$ jobs This will list all the jobs
$ bg %jobnumber (bg %1) To run in the background
$ fg %jobnumber To bring back in the foreground

Need to kill all jobs -- say you're using several suspended
emacs sessions and you just want everything to exit.

$ kill -9 `jobs -p`

The "jobs -p" gives the process number of each job, and the
kill -9 kills everything. Yes, sometimes "kill -9" is excessive
and you should issue a "kill -15" that allows jobs to clean-up.
However, for exacs session, I prefer "kill -9" and haven't had
a problem.

Sometimes you need to list the process id along with job
information. For instance, here's process id with the listing.

$ jobs -l

Note you can also renice a job, or give it lower priority.

$ nice -n +15 find . -ctime 2 -type f -exec ls {} \; > last48hours
$ bg

So above that was a ctl-z to suppend. Then, bg to run it in
the background. Now, if you want to change the priority lower
you just renice it, once you know the process id.

$ jobs -pl
[1]+ 29388 Running nice -n +15 find . -ctime 2 -exec ls -l {} \; >mout &

$ renice +30 -p 29388
29388: old priority 15, new priority 19

19 was the lowest priority for this job. You cannot increase
the priority unless you are root.

TIP 9:

Need to Delete a File for Good -- not even GOD can recover.

You have a file "secret". The following makes it so no one
can read it. If the file was 12 bytes, it's now 4096 after it
has been over written 100 times. There's no way to recover this.

$ shred -n 100 -z secret

Want to remove the file? Use the "u" option.

$ shred -n 100 -z -u test2

It can be applied to a device

$ shred -n 100 -z -u /dev/fd0

CAUTION: Note that shred relies on a very important assumption: that the file system overwrites data
in place. This is the traditional way to do things, but many modern file system designs do not sat-
isfy this assumption. The following are examples of file systems on which shred is not effective, or
is not guaranteed to be effective in all file system modes:

* log-structured or journaled file systems, such as those supplied with

AIX and Solaris (and JFS, ReiserFS, XFS, Ext3, etc.)

Also see (TIP 52).

TIP 10:

Who and What is doing What on Your System - finding open sockets,
files etc.

$ lsof
or as root
$ watch lsof -i

To list all open Internet files, use:

$ lsof -i -U

You can also get very specific about ports. Do this as root for low

$ lsof -i TCP:3306

Or, look at UDP ports as follows:

$ lsof -i UDP:1812

(See TIP 118)

Also try fuser. Suppose you have a mounted file-system, and you need
to umount it. To list the users on the file-system /work

$ fuser -u /work

To kill all processes accessing the file system /work in any way.

$ fuser -km /work

Or better yet, maybe you want to eject a cdrom on /mnt/cdrom

$ fuser -km /mnt/cdrom

If you need IO load information about your system, you can execute
iostat. But note, the very first iostat gives a snapshot since
the last boot. You typically want the following command, which gives
you 3 outputs every 5 seconds.

$ iostat -xtc 5 3
Linux 2.6.12-1.1376_FC3smp ( 10/05/2005

Time: 07:05:04 PM
avg-cpu: %user %nice %system %iowait %idle
0.97 0.06 1.94 0.62 96.41

Time: 07:05:09 PM
avg-cpu: %user %nice %system %iowait %idle
0.60 0.00 1.70 0.00 97.70

Time: 07:05:14 PM
avg-cpu: %user %nice %system %iowait %idle
1.00 0.00 1.60 0.00 97.39

vmstat reports memory statistics.

$ vmstat
$ ifconfig
$ cat /proc/sys/vm/.. (entries under here)

*NOTE: (TIP 77) shows sample usage of "ifconfig". Also
(TIP 84) shows sample output of "$ cat /proc/cpuinfo". You can download iostat
and other packages from (
You also may want to look at iozone (TIP 178).


$ cat /proc/meminfo
$ cat /proc/stat

$ cat /proc/uptime
1078623.55 1048008.34 First number is the number of seconds since boot.
The second number is the number of idle seconds.

$ cat /proc/loadavg
0.25 0.14 0.10 1/166 7778 This shows load at 1,5, and 15 minutes,
a total of 1 current running process out
from a total of 166. The 7778 is the last
process id used.

Or current process open file descriptors

$ ls -l /proc/self/fd/0
lrwx------ 1 chirico chirico 64 Jun 29 13:17 0 -> /dev/pts/51
lrwx------ 1 chirico chirico 64 Jun 29 13:17 1 -> /dev/pts/51
lrwx------ 1 chirico chirico 64 Jun 29 13:17 2 -> /dev/pts/51
lr-x------ 1 chirico chirico 64 Jun 29 13:17 3 -> /proc/26667/fd

So you could, $ echo "stuff" > /dev/pts/51, to get output. Note, tree is also
helpful here:

$ tree /proc/self

|-- auxv
|-- cmdline
|-- cwd -> /work/souptonuts/documentation/theBook
|-- environ
|-- exe -> /usr/bin/tree
|-- fd
| |-- 0 -> /dev/pts/51
| |-- 1 -> /dev/pts/51
| |-- 2 -> /dev/pts/51
| `-- 3 -> /proc/26668/fd
|-- maps
|-- mem
|-- mounts
|-- root -> /
|-- stat
|-- statm
|-- status
|-- task
| `-- 26668
| |-- auxv
| |-- cmdline
| |-- cwd -> /work/souptonuts/documentation/theBook
| |-- environ
| |-- exe -> /usr/bin/tree
| |-- fd
| | |-- 0 -> /dev/pts/51
| | |-- 1 -> /dev/pts/51
| | |-- 2 -> /dev/pts/51
| | `-- 3 -> /proc/26668/task/26668/fd
| |-- maps
| |-- mem
| |-- mounts
| |-- root -> /
| |-- stat
| |-- statm
| |-- status
| `-- wchan
`-- wchan

10 directories, 28 files

Need a listing of the system settings?

$ sysctl -a

Need IPC (Shared Memory Segments, Semaphore Arrays, Message Queue) status

$ ipcs
$ ipcs -l "This gives limits"

Need to "watch" everything a user does? The following watches donkey.

$ watch lsof -u donkey

Or, to see what in going on in directory "/work/junk"

$ watch lsof +D /work/junk

TIP 11:

How to make a File "immutable" or "unalterable" -- it cannot be changed
or deleted even by root. Note this works on (ext2/ext3) filesystems.
And, yes, root can delete after it's changed back.

As root:

$ chattr +i filename

And to change it back:

$ chattr -i filename

List attributes

$ lsattr filename

TIP 12:

SSH - How to Generate the Key Pair.

On the local server

$ ssh-keygen -t dsa -b 2048

This will create the two files:

.ssh/id_dsa (Private key)
.ssh/ (Public key you can share)

Next insert ".ssh/" on the remote server
in the file ".ssh/authorized_keys" and ".ssh/authorized_keys2"
and change the permission of each file to (chmod 600). Plus, make
sure the directory ".ssh" exists on the remote computer with 700 rights.
Ok, assuming is the remote server and "donkey" is the
account on that remote server.

$ ssh donkey@ "mkdir -p .ssh"
$ ssh donkey@ "chmod 700 .ssh"
$ scp ./.ssh/ donkey@

Now connect to that remote server "" and add .ssh/
to both "authorized_keys" and "authorized_keys2". When done, the permission
(This is on the remote server)

$chmod 600 .ssh/authorized_key*

Next, go back to the local server and issue the following:

$ ssh-agent $SHELL
$ ssh-add

The "ssh-add" will allow you to enter the passphrase and it will
save it for the current login session.

You don't have to enter a password when running "ssh-keygen" above. But,
remember anyone with root access can "su - " and then connect
to your computers. It's harder, however, not impossible, for root to do
this if you have a password.

(Reference TIP 151)

TIP 13:

Securing the System: Don't allow root to login remotely. Instead,
the admin could login as another account, then, "su -". However,
root can still login "from the local terminal".

In the "/etc/ssh/sshd_config" file change the following lines:

Protocol 2
PermitRootLogin no
PermitEmptyPasswords no

Then, restart ssh

/etc/init.d/sshd restart

Why would you want to do this? It's not possible for anyone to guess
or keep trying the root account. This is especially good for computers
on the Internet. So, even if the "root" passwords is known, they can't
get access to the system remotely. Only from the terminal, which is locked
in your computer room. However, if anyone has a account on the server,
then, they can login under their account then "su -".

Suppose you only want a limited number of users: "mchirico" and "donkey".
Add the following line to "/etc/ssh/sshd_config". Note, this allows access
for chirico and donkey, but everyone else is denied.

# Once you add AllowUsers - everyone else is denied.
AllowUsers mchirico donkey

TIP 14:

Keep Logs Longer with Less Space.

Normally logs rotate monthly, over writing all the old data. Here's a
sample "/etc/logrotate.conf" that will keep 12 months of backup
compressing the logfiles

$ cat /etc/logrotate.conf

# see "man logrotate" for details
# rotate log files weekly
#chirico changes to monthly

# keep 4 weeks worth of backlogs
# keep 12 months of backup
rotate 12

# create new (empty) log files after rotating old ones

# uncomment this if you want your log files compressed

# RPM packages drop log rotation information into this directory
include /etc/logrotate.d

# no packages own wtmp -- we'll rotate them here
/var/log/wtmp {
create 0664 root utmp
rotate 1

# system-specific logs may be also be configured here.

Note: see tip 1. The clock should always be correctly set.

TIP 15:

What Network Services are Running?

$ netstat -atup


$ netstat -ap|grep LISTEN|less

This can be helpful to determine the services running.

Need stats on dropped UDP packets?

$ netstat -s -u

or TCP

$ netstat -s -t

or summary of everything

$ netstat -s

or looking for error rates on the interface?

$ netstat -i

Listening interfaces?

$ netstat -l

(Tip above provided by Amos Shapira)

Also see TIP 77.

TIP 16:

Apache: Creating and Using an ".htaccess" File

Below is a sample ".htaccess" file which goes in
"/usr/local/apache/htdocs/chirico/alpha/.htaccess" for this

AuthUserFile /usr/local/apache/htdocs/chirico/alpha/.htpasswd
AuthGroupFile /dev/null
AuthName "Your Name and regular password required"
AuthType Basic

require valid-user

In order for this to work /usr/local/apache/conf/httpd.conf must
have the following line in it:


AllowOverride FileInfo AuthConfig Limit
Options MultiViews Indexes SymLinksIfOwnerMatch IncludesNoExec

Order allow,deny
Allow from all

Order deny,allow
Deny from all

Also, a password file must be created

$ /usr/local/apache/bin/htpasswd -c .htpasswd chirico

And enter the user names and passwords.

Next Reload Apache:

$ /etc/init.d/httpd reload

(Reference TIP 213 limit access to certain directories based on IP address).

TIP 17:

Working with "mt" Commands: reading and writing to tape.

The following assumes the tape device is "/dev/st0"

STEP 1 ( rewind the tape)

# mt -f /dev/nst0 rewind

STEP 2 (check to see if you are at block 0)

# mt -f /dev/nst0 tell
At block 0.

STEP 3 (Backup "tar compress" directories "one" and "two")

# tar -czf /dev/nst0 one two

STEP 4 (Check to see what block you are at)

# mt -f /dev/nst0 tell

You should get something like block 2 at this point.

STEP 5 (Rewind the tape)

# mt -f /dev/nst0 rewind

STEP 6 (List the files)

# tar -tzf /dev/nst0

STEP 7 (Restore directory "one" into directory "junk"). Note, you
have to first rewind the tape, since the last operation moved
ahead 2 blocks. Check this with "mt -f /dev/nst0".

# cd junk
# mt -f /dev/nst0 rewind
# mt -f /dev/nst0 tell
At block 0.
# tar -xzf /dev/nst0 one

STEP 8 (Next, take a look to see what block the tape is at)

# mt -f /dev/nst0 tell
At block 2.

STEP 9 (Now backup directories three and four)

# tar -czf /dev/nst0 three four

After backing up the files, the tape should be past block 2.
Check this.

# mt -f /dev/nst0 tell
At block 4.

Currently the following exist:

At block 1:

At block 2:

At block 4:
(* This is empty *)

A few notes. You can set the blocking factor and a label
with tar. For example:

$ tar --label="temp label" --create --blocking-factor=128 --file=/dev/nst0 Notes

But note if you try to read it with the default, incorrect blocking
factor, then, you will get the following error:

$ tar -t --file=/dev/nst0
tar: /dev/nst0: Cannot read: Cannot allocate memory
tar: At beginning of tape, quitting now
tar: Error is not recoverable: exiting now

However this is easily fixed with the correct blocking factor

$ mt -f /dev/nst0 rewind
$ tar -t --blocking-factor=128 --file=/dev/nst0
temp label

Take advantage of the label command.

$ MYCOMMENTS="Big_important_tape"
$ tar --label="$(date +%F)"+"${MYCOMMENTS}"

Writing to tape on a remote computer

$ tar cvzf - ./tmp | ssh -l chirico '(mt -f /dev/nst0 rewind; dd of=/dev/st0 )'

Restoring the contents from tape on a remote computer

$ ssh -l chirico '(mt -f /dev/nst0 rewind; dd if=/dev/st0 )'|tar xzf -

Getting data off of tape with dd command with odd blocking factor. Just set ibs very high

$ mt -f /dev/nst0 rewind
$ tar --label="Contenets of Notes" --create --blocking-factor=128 --file=/dev/nst0 Notes
$ mt -f /dev/nst0 rewind
$ dd ibs=1048576 if=/dev/st0 of=notes.tar

The above will probably work with ibs=64k as well

(Also see TIP 136)

TIP 18:

Encrypting Data to Tape using "tar" and "openssl".

The following shows an example of writing the contents of "tapetest" to tape:

$ tar zcvf - tapetest|openssl des3 -salt -k secretpassword | dd of=/dev/st0

Reading the data back:

$ dd if=/dev/st0|openssl des3 -d -k secretpassword|tar xzf -

TIP 19:

Mounting an ISO Image as a Filesystem -- this is great if you don't have the DVD
hardware, but, need to get at the data. The following show an example of
mounting the Fedora core 2 as a file.

$ mkdir /iso0
$ mount -o loop -t iso9660 /FC2-i386-DVD.iso /iso0

Or to mount automatically at boot, add the following to "/etc/fstab"

/FC2-i386-DVD.iso /iso0 iso9660 rw,loop 0 0


TIP 20:

Getting Information about the Hard drive and list all PCI devices.

$ hdparm /dev/hda

multcount = 16 (on)
IO_support = 0 (default 16-bit)
unmaskirq = 0 (off)
using_dma = 1 (on)
keepsettings = 0 (off)
readonly = 0 (off)
readahead = 256 (on)
geometry = 16383/255/63, sectors = 234375000, start = 0

or for SCSI

$ hdparm /dev/sda

Try it with the -i option for information

$ hdparm -i /dev/hda


Model=IC35L120AVV207-1, FwRev=V24OA66A, SerialNo=VNVD09G4CZ6E0T
Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs }
RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=52
BuffType=DualPortCache, BuffSize=7965kB, MaxMultSect=16, MultSect=16
CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=234375000
IORDY=on/off, tPIO={min:240,w/IORDY:120}, tDMA={min:120,rec:120}
PIO modes: pio0 pio1 pio2 pio3 pio4
DMA modes: mdma0 mdma1 mdma2
UDMA modes: udma0 udma1 udma2 udma3 udma4 *udma5
AdvancedPM=yes: disabled (255) WriteCache=enabled
Drive conforms to: ATA/ATAPI-6 T13 1410D revision 3a: 2 3 4 5 6

How fast is your drive?

$ hdparm -tT /dev/hda

Timing buffer-cache reads: 128 MB in 0.41 seconds =315.32 MB/sec
Timing buffered disk reads: 64 MB in 1.19 seconds = 53.65 MB/sec

Need to find your device?

$ mount
$ cat /proc/partitions
$ dmesg | egrep '^(s|h)d'

which for my system lists:

hda: IC35L120AVV207-1, ATA DISK drive
hdc: Lite-On LTN486S 48x Max, ATAPI CD/DVD-ROM drive
hda: max request size: 1024KiB
hda: 234375000 sectors (120000 MB) w/7965KiB Cache, CHS=16383/255/63, UDMA(100)

By the way, if you want to turn on dma

$ hdparm -d1 /dev/hda
setting using_dma to 1 (on)
using_dma = 1 (on)

(Also see TIP 122 )

List all PCI devices

$ lspci -v

00:00.0 Host bridge: Intel Corp. 82845G/GL [Brookdale-G] Chipset Host Bridge (rev
Subsystem: Dell Computer Corporation: Unknown device 0160
Flags: bus master, fast devsel, latency 0
Memory at f0000000 (32-bit, prefetchable) [size=128M]

... lots more ...

Note, there is also lspci -vv for even more information.

(Also see TIP 200)

TIP 21:

Setting up "cron" Jobs.

If you want to use the emacs editor for editing cron jobs, then,
set the following in your "/home/user/.bash_profile"


Then, to edit cron jobs

$ crontab -e

You may want to put in the following header

#MINUTE(0-59) HOUR(0-23) DAYOFMONTH(1-31) MONTHOFYEAR(1-12) DAYOFWEEK(0-6) Note 0=Sun and 7=Sun
#14,15 10 * * 0 /usr/bin/somecommmand >/dev/null 2>&1

The sample "commented out command" will run at 10:14 and 10:15 every Sunday. There will
be no "mail" sent to the user because of the ">/dev/null 2>&1" entry.

$ crontab -l

The above will list all cron jobs. Or if you're root

$ crontab -l -u
$ crontab -e -u

Reference "man 5 crontab":

The time and date fields are:

field allowed values
----- --------------
minute 0-59
hour 0-23
day of month 1-31
month 1-12 (or names, see below)
day of week 0-7 (0 or 7 is Sun, or use names)

A field may be an asterisk (*), which always stands for ``first-last''.

Ranges of numbers are allowed. Ranges are two numbers separated with a
hyphen. The specified range is inclusive. For example, 8-11 for an
``hours'' entry specifies execution at hours 8, 9, 10 and 11.

Lists are allowed. A list is a set of numbers (or ranges) separated by
commas. Examples: ``1,2,5,9'', ``0-4,8-12''.

Ranges can include "steps", so "1-9/2" is the same as "1,3,5,7,9".

Note, you can run just every 5 minutes as follows:

*/5 * * * * /etc/mrtg/domrtg >/dev/null 2>&1

To run jobs hourly,daily,weekly or monthly you can add shell scripts into the
appropriate directory:


Note that the above are pre-configured schedules set in "/etc/crontab", so
if you want, you can change the schedule. Below is my /etc/crontab:

$ cat /etc/crontab

# run-parts
01 * * * * root run-parts /etc/cron.hourly
02 4 * * * root run-parts /etc/cron.daily
22 4 * * 0 root run-parts /etc/cron.weekly
42 4 1 * * root run-parts /etc/cron.monthly

TIP 22:

Keeping Files in Sync Between Servers.

The remote computer is "" and has the account "donkey". You want
to "keep in sync" the files under "/home/cu2000/Logs" on the remote computer
with files on "/home/chirico/dev/MEDIA_Server" on the local computer.

$ rsync -Lae ssh donkey@ /home/chirico/dev/MEDIA_Server

"rsync" is a convient command for keeping files in sync, and as shown here will work
through ssh. The -L option tells rsync to treat symbolic links like ordinary files.

Also see []

TIP 23:

Looking up the Spelling of a Word.

$ look

so the following will list all words that
start with stuff

$ look stuff

It helps to have a large "linuxwords" dictionary. You can download
a much bigger dictionary from the following:

Note: vim users can setup the .vimrc file with the following. Now when you type
CTL-X CTL-T in insert mode, you'll get a thesaurus lookup.

set dictionary+=/usr/share/dict/words
set thesaurus+=/usr/share/dict/words

Or, you can call aspell with the F6 command after putting the folling entry in your
.vimrc file

:nmap :w:!aspell -e -c %:e

Now, hit F6 when you're in vim, and you'll get a spell checker.

There is also an X Windows dictionary that runs with the following command.

$ gnome-dictionary

TIP 24:

Find out if a Command is Aliased.

$ type -all


$ type -all ls
ls is aliased to `ls --color=tty'
ls is /bin/ls

TIP 25:

Create a Terminal Calculator

Put the following in your .bashrc file

function calc
echo "${1}"|bc -l;

Or, run it at the shell prompt. Now
"calc" from the shell will work as follows:

$ calc 3+45

All functions with a "(" or ")" must be enclosed
in quotes. For instance, to get the sin of .4

$ calc "s(.4)"

(See TIP 115 using the expr command)

TIP 26:

Kill a User and All Their Current Processes.

# This program will kill all processes from a
# user. The user name is read from the command line.
# This program also demonstrates reading a bash variable
# into an awk script.
# Usage: kill9user
kill -9 `ps aux|awk -v var=$1 '$1==var { print $2 }'`

or if you want want to create the above script the command
below will kill the user "donkey" and all of his processes.

$ kill -9 `ps aux|awk -v var="donkey" '$1==var { print $2 }'`

Check their cron jobs and "at" jobs, if you have a security issue.

$ crontab -u -e

Lock the account:

$ passwd -l

Remove all authorized_keys

$ rm /home/user/.shosts
$ rm /home/user/.rhosts
$ rm -rf /home/user/.ssh
$ rm /home/user/.forward

or consider

$ mv /home/user /home/safeuser

Change the shell

$ chsh -s /bin/true

Do an inventory

$ find / -user > list_of_user_files

NOTE: Also see (TIP 10).

To see all users, except the current user. Do not use the
dash "ps -aux" is wrong but the following is correct:

$ ps aux| awk '!/'${USER}'/{printf("%s \n",$0)}'

or (ww = wide, wide output)

$ ps auwwx| awk '!/'${USER}'/{printf("%s \n",$0)}'

The following codes may be useful:

D Uninterruptible sleep (usually IO)
R Running or runnable (on run queue)
S Interruptible sleep (waiting for an event to complete)
T Stopped, either by a job control signal or because it is being traced.
W paging (not valid since the 2.6.xx kernel)
X dead (should never be seen)
Z Defunct ("zombie") process, terminated but not reaped by its parent.

For BSD formats and when the stat keyword is used, additional
characters may be displayed:

< high-priority (not nice to other users)
N low-priority (nice to other users)
L has pages locked into memory (for real-time and custom IO)
s is a session leader
l is multi-threaded (using CLONE_THREAD, like NPTL pthreads do)
+ is in the foreground process group

Also see TIP 28. and TIP 89.

TIP 27:

Format Dates for Logs and Files

$ date "+%m%d%y %A,%B %d %Y %X"
061704 Thursday,June 17 2004 07:13:40 PM

$ date "+%m%d%Y"

$ date -d '1 day ago' "+%m%d%Y"

$ date -d '3 months 1 day 2 hour 15 minutes 2 seconds ago'

or to go into the future remove the "ago"

$ date -d '3 months 1 day 2 hour 15 minutes 2 seconds '

Also the following works:

$ date -d '+2 year +1 month -1 week +3 day -8 hour +2 min -5 seconds'

Quick question: If there are 100,000,000 stars in the visible sky, and you can
count them, round the clock, at a rate of a star per second starting now, when
would you finish counting? Would you still be alive?

$ date -d '+100000000 seconds'

Sooner than you think!

This can be assigned to variables

$ mdate=`date -d '3 months 1 day 2 hour 15 minutes 2 seconds ' "+%m%d%Y_%A_%B_%D_%Y_%X" `
$ echo $mdate
09182004_Saturday_September_09/18/04_2004_09:40:41 PM
^---- Easy to sort ^-------^----- Easy to read

See TIP 28 below.

See TIP 87 when working with large delta time changes -40 years, or -200 years ago, or even
1,000,000 days into the future.

Also see (TIP 1) for working with time zones.

TIP 28:

Need Ascii Codes? For instance, for printing quotes:

awk 'BEGIN { msg = "Don\047t Panic!"; printf "%s \n",msg }'
awk 'BEGIN { msg = "Don\x027t Panic!"; printf "%s \n",msg }'

It's better to use \047, because certain characters that follow \x027 may cause problems.

For example, take a look at the following two lines. The first line prints a "}" caused
by the extra D in \x027D. The the line immediately below does not work as expected.

awk 'BEGIN {printf("The D causes problems \x027D\n")}'

However, the line below works fine:

awk 'BEGIN {printf("The D does not cause problems \047D\n")}'

Or if you wanted to use the date command in "awk" to print date.time.nanosecond.timezone for
each line of a file "test".

The following date can be used in awk because the single quotes are enclosed within the
double quotes.

date '+%m%d%Y.%H%M%S.%N%z'

$ awk 'BEGIN { "date '+%m%d%Y.%H%M%S.%N%z'" | getline MyDate } { print MyDate,$0 }' < data

But it's also possible to replace "+" with \x2B, "%" with \x25, and "d" with \x64 as follows:

$ awk 'BEGIN { "date \x27\x2B\x25m\x25\x64\x25Y.\x25H\x25M\x25S.\x25N\x25z\x27" | getline MyDate } { print MyDate,$0 }' < test

07062004.113820.346033000-0400 bob 71
07062004.113820.346033000-0400 tom 43
07062004.113820.346033000-0400 sal 34
07062004.113820.346033000-0400 bob 89
07062004.113820.346033000-0400 tom 66
07062004.113820.346033000-0400 sal 99

For this example it's not needed because single quotes are used inside of double quotes; however, there may be times when
hex replacement is easier.

$ man ascii

Oct Dec Hex Char Oct Dec Hex Char
000 0 00 NUL '\0' 100 64 40 @
001 1 01 SOH 101 65 41 A
002 2 02 STX 102 66 42 B
003 3 03 ETX 103 67 43 C
004 4 04 EOT 104 68 44 D
005 5 05 ENQ 105 69 45 E
006 6 06 ACK 106 70 46 F
007 7 07 BEL '\a' 107 71 47 G
010 8 08 BS '\b' 110 72 48 H
011 9 09 HT '\t' 111 73 49 I
012 10 0A LF '\n' 112 74 4A J
013 11 0B VT '\v' 113 75 4B K
014 12 0C FF '\f' 114 76 4C L
015 13 0D CR '\r' 115 77 4D M
016 14 0E SO 116 78 4E N
017 15 0F SI 117 79 4F O
020 16 10 DLE 120 80 50 P
021 17 11 DC1 121 81 51 Q
022 18 12 DC2 122 82 52 R
023 19 13 DC3 123 83 53 S
024 20 14 DC4 124 84 54 T
025 21 15 NAK 125 85 55 U
026 22 16 SYN 126 86 56 V
027 23 17 ETB 127 87 57 W
030 24 18 CAN 130 88 58 X
031 25 19 EM 131 89 59 Y
032 26 1A SUB 132 90 5A Z
033 27 1B ESC 133 91 5B [
034 28 1C FS 134 92 5C \ '\\'
035 29 1D GS 135 93 5D ]
036 30 1E RS 136 94 5E ^
037 31 1F US 137 95 5F _
040 32 20 SPACE 140 96 60 `
041 33 21 ! 141 97 61 a
042 34 22 " 142 98 62 b
043 35 23 # 143 99 63 c
044 36 24 $ 144 100 64 d
045 37 25 % 145 101 65 e
046 38 26 & 146 102 66 f
047 39 27 ' 147 103 67 g
050 40 28 ( 150 104 68 h
051 41 29 ) 151 105 69 i
052 42 2A * 152 106 6A j
053 43 2B + 153 107 6B k
054 44 2C , 154 108 6C l
055 45 2D - 155 109 6D m
056 46 2E . 156 110 6E n
057 47 2F / 157 111 6F o
060 48 30 0 160 112 70 p
061 49 31 1 161 113 71 q
062 50 32 2 162 114 72 r
063 51 33 3 163 115 73 s
064 52 34 4 164 116 74 t
065 53 35 5 165 117 75 u
066 54 36 6 166 118 76 v
067 55 37 7 167 119 77 w
070 56 38 8 170 120 78 x
071 57 39 9 171 121 79 y
072 58 3A : 172 122 7A z
073 59 3B ; 173 123 7B {
074 60 3C < 174 124 7C |
075 61 3D = 175 125 7D }
076 62 3E > 176 126 7E ~
077 63 3F ? 177 127 7F DEL

TIP 29:

Need a WWW Browser for the Terminal Session? Try lynx or elinks.

$ lynx

Or to read all these tips, with the latest updates

$ lynx

Or, better yet elinks.

$ elinks http://somepage.

You can get elinks at the following site:

TIP 30:

screen - screen manager with VT100/ANSI terminal emulation

This is an excellent utility. But if you work a lot in Emacs,
then, you should place the following in your ~./.bashrc

alias s='screen -e^Pa -D -R'

After loging in again (or source .bashrc) ,
type the following to load "screen":

$ s

If you're using the not using the alias command above, substitute
CTL-a for CTL-p below. :

CTL-p CTL-C To get a new session
CTL-p " To list sessions, and arrow keys to move
CTL-p SHFT-A To name sessions
CTL-p S To split screens
CLT-p Q To unsplit screens
CLT-p TAB To switch between screens
CLT-p :resize n To resize screen to n rows, on split screen

Screen is very powerful. Should you become disconneced, you can
still resume work after loggin in.

$ man screen

The above command will give you more information.

TIP 31:

Need to Find the Factors of a Number?

$ factor 2345678992
2345678992: 2 2 2 2 6581 22277

It's a quick way to find out if a number is prime

$ factor 7867
7867: 7867

TIP 32:

Less is More -- piping to less to scroll backword and forward

For large "ls" listings try the followin, then, use the arrow key
to move up and down the list.

$ ls /some_large_dir/ | less


$ cat some_large_file | less


$ less some_large_file

TIP 33:

C "indent" Settings for Kernel Development

$ indent -kr -i8 program.c

TIP 34:

FTP auto-login. "ftp" to a site and have the password stored.

For instance, here's a sample ".net" file in a user's home
directory for uploading to sourceforge. Note, sourceforge will
take any password, so is used here for login "anonymous".

$ cat ~/.netrc
machine login anonymous password
default login anonymous password user@site

It might be a good idea to change the rights on this file

$ chmod 0400 ~/.netrc

# Sample ftp automated script to download
# file to ${dwnld}
cd ${dwnld}
prompt off
cd /pub/usenet-by-group/news.answers/unix-faq/faq
mget contents
mget diff
mget part*

Sourceforge uses an anonymous login with an email address as
a password. Below is the automated script I use for uploading
binary files.

# ftp sourceforge auto upload
# Usage: ./
# machine user anonymous
ftp -n -u << FTPSTRING
user anonymous
cd incoming
put ${1}

(Also see TIP 114 for ncftpget, which is a very powerful restarting
ftp program)

TIP 35:

Bash Brace Expansion

$ echo f{ee,ie,oe,um}
fee fie foe fum

This works with almost any command

$ mkdir -p /work/junk/{one,two,three,four}

TIP 36:

Getting a List of User Accounts on the System

$ cut -d: -f1 /etc/passwd | sort

TIP 37:

Editing a Bash Command

Try typing a long command say, then, type "fc" for an easy way
to edit the command.

$ find /etc -iname '*.cnf' -exec grep -H 'log' {} \;
$ fc

"fc" will bring the last command typed into an editor, "emacs" if
that's the default editor. Type "fc -l" to list last few commands.

To seach for a command, try typing "CTL-r" at the shell prompt for
searching. "CTL-t" to transpose, say "sl" was typed by you want "ls".

Hints when using "fc: in emacs:

ESC-b move one word backward
ESC-f move one word forward
ESC-DEL kill one word backward
CTL-k kill point to end
CTL-y un-yank killed region at point

TIP 38:

Moving around Directories.

Change to the home directory:
$ cd ~
$ cd

To go back to the last directory
$ cd -

Instead of "cd" to a directory try "pushd" and look
at the can see a list of directories.

$ pushd /etc
$ pushd /usr/local

Then, to get back "popd" or "popd 1"

To list all the directories pushed on the stack
use the "dirs -v" command.

$ dirs -v
0 /usr/local
1 /etc
2 /work/souptonuts/documentation/theBook

Now, if you "pushd +1" you will be moved to "/etc", since
is number "1" on the stack, and this directory will become

$ pwd
$ pushd +1
$ pwd

$ dirs -v
0 /etc
1 /work/souptonuts/documentation/theBook
2 /usr/local

TIP 39:

Need an Underscore after a Variable?

Enclose the variable in "{}".

$echo ${UID}_

Compare to

$echo $UID_

Also try the following:

$ m="my stuff here"
$ echo -e ${m// /'\n'}

TIP 40:

Bash Variable Offset and String Operators

$ r="this is stuff"
$ echo ${r:3}
$ echo ${r:5:2}

Note, ${varname:offset:length}

${varname:?message} If varname exist and isn't null return value, else,
print var and message.

$ r="new stuff"
$ echo ${r:? "that's r for you"}
new stuff
$ unset r
$ echo ${r:? "that's r for you"}
bash: r: that's r for you

${varname:+word} If varname exist and not null return word. Else, return null.

${varname:-word} If varname exist and not null return value. Else, return word.

Working with arrays in bash - bash arrays.

$ unset p
$ p=(one two three)

$ echo -e "${p[@]}"
one two three


$ echo -e "${p[*]}"
one two three

$ echo -e "${#p[@]}"

$ echo -e "${p[0]}"

$ echo -e "${p[1]}"

Also see (TIP 95)

TIP 41:

Loops in Bash

The command below loops through directories listed in $PATH.

$ path=$PATH:
$ while [ $path ]; do echo " ${path%%:*} "; path=${path#*:}; done

The command below will also loop through directories in your path.

$ for dir in $PATH
> do
> ls -ld $dir
> done
drwxr-xr-x 2 root root 4096 Jun 10 20:16 /usr/local/bin
drwxr-xr-x 2 root root 4096 Jun 13 23:12 /bin
drwxr-xr-x 3 root root 40960 Jun 12 08:00 /usr/bin
drwxr-xr-x 2 root root 4096 Feb 14 03:12 /usr/X11R6/bin
drwxrwxr-x 2 chirico chirico 4096 Jun 6 13:06 /home/chirico/bin

Other ways of doing loops:

for (( i=1; i <= 20; i++))
echo -n "$i "

Note, to do it all on one line, do the following:

$ for (( i=1; i <= 20; i++)); do echo -n "$i"; done

Below, is an example of declaring i an integer so that you do not
have to preface with let.

$ declare -i i
$ i=5;
$ while (( $i > 1 )); do
> i=i-1
> echo $i
> done

You can also use "while [ $i -gt 1 ]; do" in place of "while (( $i > 1 )); do"

To get a listing of all declared values

$ declare -i

Try putting a few words in the file "test"

$ while read filename; do echo "- $filename "; done < test |nl -w1

Or, using an array

declare -a Array
for i in `seq ${#Array[@]}`
echo $Array[$i-1]

Also see (TIP 95 and TIP 133).

TIP 42:

"diff" and "patch".

You have created a program "prog.c", saved as this name and also copied
to "prog.c.old". You post "prog.c" to users. Next, you make changes
to prog.c

$ diff -c prog.c.old prog.c > prog.patch

Now, users can get the latest updates by running.

$ patch < prog.patch

By the way, you can make backups of your data easily.

$ cp /etc/fstab{,.bak}

Now, you do your edits to "/etc/fstab" and if you need
to go back to the original, you can find it at

TIP 43:

"cat" the Contents of Files Listed in a File, in That Order.

SETUP (Assume you have the following)

$ cat file_of_files

$ cat file1
This is the data in file1

$ cat file 2
This is the data in file2

So there are 3 files here "file_of_files" which contains the name of
other files. In this case "file1" and "file2". And the contents of
"file1" and "file2" is shown above.

$ cat file_of_files|xargs cat
This is the data in file1
This is the data in file2

Also see (TIP 44, TIP 62 and TIP 235).

TIP 44:

Columns and Rows -- getting anything you want.

Assume you have the following file.

$ cat data
1 2 3
4 5
6 7 8 9 10
11 12
13 14

How to you get everything in 2 columns?

$ cat data|tr ' ' '\n'|xargs -l2
1 2
3 4
5 6
7 8
9 10
11 12
13 14

Three columns?

$ cat data|tr ' ' '\n'|xargs -l3
1 2 3
4 5 6
7 8 9
10 11 12
13 14

What's the row sum of the "three columns?"

$ cat data|tr ' ' '\n'|xargs -l3|tr ' ' '+'|bc


$ tr ' ' '\n' < data |xargs -l3|tr ' ' '+'|bc

NOTE "Steven Heiner's rule":

cat one_file | program

can always be rewritten as

program < one_file

Note: thanks to Steven Heiner ( the above can be
shortened as follows:

$ tr ' ' '\n' < data|xargs -l3|tr ' ' '+'|bc

Need to "tr" from the stdin?

$ tr "xy" "yx"| ... | ...

But there is a the "Stephane CHAZELAS" condition here

"Note that tr, sed, and awk mail fail on files containing '\0'
sed and awk have unspecified behaviors if the input
doesn't end in a '\n' (or to sum up, cat works for
binary and text files, text utilities such as sed or awk
work only for text files).

TIP 45:

Auto Directory Spelling Corrections.

To turn this on:

$ shopt -s cdspell

Now mispell a directory in the cd command.

$ cd /usk/local
^-------- still gets you to --

What other options can you set? The following will list
all the options:

$ shopt -p

TIP 46:

Record Eveything Printed on Your Terminal Screen.

$ script -a

Now start doing stuff and "everything" is appended to .
For example

$ script installation

$ (command)

$ (result)

$ ...

$ ...

$ (command)

$ (result)

$ exit

The whole session log is in the installation file that you can later
read and/or cleanup and add to a documentation.

This command can also be used to redirect the contents to another user,
but you must be root to do this.

Step 1 - find out what pts they are using.

$ w

Step 2 - Run script on that pts. After running this command below
everything you type will appear on their screen.

$ script /dev/pts/4

Thanks to Jacques.GARNIER-EXTERIEUR@EU.RHODIA.COM for his contribution
to this tip.

Also reference TIP 208.

TIP 47:

Monitor all Network Traffic Except Your Current ssh Connection.

$ tcpdump -i eth0 -nN -vvv -xX -s 1500 port not 22

Or to filter out port 123 as well getting the full length of the packet
(-s 0), use the following:

$ tcpdump -i eth0 -nN -vvv -xX -s 0 port not 22 and port not 123

Or to filter only a certain host say

$ tcpdump -i eth0 -nN -vvv -xX port not 22 and host

Just want ip addresses and a little bit of data, then,
use this. The "-c 20" is to stop after 20 packets.

$ tcpdump -i eth0 -nN -s 1500 port not 22 -c 20

If you're looking for sign of DOS attacks, the following show just the SYN
packets on all interfaces:

$ tcpdump 'tcp[13] & 2 == 2'

TIP 48:

Where are the GNU Reference Manuals?

Also worth a look the "Linux Documentation Project"

and Red Hat manuals

TIP 49:

Setting or Changing the Library Path.

The following contains the settings to be added or deleted


After this file is edited, you must run the following:

$ ldconfig

See "man ldconfig" for more information.

TIP 50:

Working with Libraries in C

Assume the following 3 programs:

$ cat ./src/test.c

int test(int t)
return t;

$ cat ./src/prog1.c

program: prog1.c
dependences: test.c

compiling this program:
gcc -o prog test.c prog1.c

Note the libpersonal include
should be remove if NOT using the

main(int argc, char **argv)

$ cat ./include/libpersonal.h

extern int test(int);

Prog1.c needs the test function in test.c
To compile, so that both programs work together, do the following:

$ cd src
$ gcc -o prog test.c prog1.c -I../include

However, if you want to create your own static library, then, run the following:

$ mkdir -p ../lib
$ gcc -c test.c -o ../lib/test.o
$ cd ../lib
$ ar r libpersonal.a test.o
$ ranlib libpersonal.a

or, the ar and ranlib command can be combined as follows:

$ ar rs libpersonal.a test.o

To compile the program with the static library:

$ cd ../src
$ gcc -I../include -L../lib -o prog1 prog1.c -lpersonal

The -I../include tells gcc to look in the ../include directory for
libpersonal.h. and -L../lib, tells gcc to look for the
"libpersonal.a" library.

$ cd ..
$ tree src lib include
|-- prog
|-- prog1
|-- prog1.c
`-- test.c
|-- libpersonal.a
`-- test.o
`-- libpersonal.h

This was a STATIC library. Often times you will want to use a SHARED
or dynamic library.


You must recompile test.c with -fpic option.

$ cd ../lib
$ gcc -c -fpic ../src/test.c -o test.o

Next create the file.

$ gcc -shared -o test.o

Now, compile the source prog1.c as follows:

$ cd ../src
$ gcc -Wl,-R../lib -L../lib -I../include -o prog2 prog1.c -lpersonal

This should work fine. But, take a look at prog2 using the ldd command.

$ ldd prog2 => ../lib/ (0x40017000) => /lib/tls/ (0x42000000)
/lib/ => /lib/ (0x40000000)

If you move the program prog2 to a different location, it will not run.
Instead you will get the following error:

prog2: error while loading shared libraries:
cannot open shared object file: No such file or directory

To fix this, you should specify the direct path to the library. And in my
case it is rather long

$ gcc -Wl,-R/work/souptonuts/documentation/theBook/lib -L../lib -I../include -o prog2 prog1.c -lpersonal

SPECIAL NOTE: The -R must always follow the -Wl. (-Wl,-R) They always go together

TIP 51:

Actively Monitor a File and Send Email when Expression Occurs.

This is a way to monitor "/var/log/messages" or any file for certain changes.
The example below actively monitors "stuff" for the work "now" and as soon as
"now" is added to the file, the contents of msg are sent to the user

$ tail -f stuff | \
awk ' /now/ { system("mail -s \"This is working\" < msg") }'

Or, you can run a program, say get headings on slashdot from the program "getslash.php" which
runs on "" with account "chirico". Assuming you have ssh keys setup, then, the following
will send mail from the output:

$ ssh chirico@ "./bin/getslash.php"|mail -s "Slash cron Headlines"

See (TIP 80) for scraping the headings on slash dot and how to get a copy of "getslash.php". If you still
want to use awk:

$ ssh chirico@ "./bin/getslash.php"| \
awk '{ print $0 | "mail -s \x27 Slash Topics \x27 "}'

Note the "\x27" is a quote. Maybe you only want articles dealing with "Linux":

$ ssh chirico@ "./bin/getslash.php"| \
awk '/Linux/{ print $0 | "mail -s \x27 Slash Topics \x27 "}'

For $60, you can get a numeric display from "delcom engineering" that you can send messages and
data to. I get weather information off the internet and send it to this device.

(Reference TIP 151 for ssh tips)

TIP 52:

Need to Keep Secrets? Encrypt it.

To Encrypt:

$ openssl des3 -salt -in file.txt -out file.des3

The above will prompt for a password, or you can put it in
with a -k option, assuming you're on a trusted server.

To Decrypt

$ openssl des3 -d -salt -in file.des3 -out file.txt -k mypassword

Need to encrypt what you type? Enter the following, then start typing
and ^D to end.

$ openssl des3 -salt -out stuff.txt

TIP 53:

Check that a File has Not Been Tampered With: Use Cryptographic Hashing Function.

The md5sum is popular but dated

$ md5sum file

Instead, use one of the following;

$ openssl dgst -sha1 -c file

$ openssl dgst -ripemd160 -c file

All calls give a fixed length string or "message digest".

TIP 54:

Need to View Information About a Secure Web Server? A SSL/TLS test.

$ openssl s_client -connect

Above will give a long listing of certificates.

Note, it is also possible to get certificate information about a mail server

$ openssl s_client -connect -showcerts

When you do the above command you get two certificates. If you copy
past both certificates by taking the following contents include the
begin and end show below:


Then create files "comcast0.pem" and "comcast1.pem" out of these certificaties and
put them in a directory "/home/donkey/.certs", then, with the openssl src package, in
the "./tools/c_rehash" run

$ c_rehash .certs
Doing .certs
comcast0.pem => 72f90dc0.0
comcast1.pem => f73e89fd.0

Now it's possible to have fetchmail work with these certs.

# Sample .fetchmailrc file for Comcast
# Check mail every 90 seconds
set daemon 90
set syslog
set postmaster donkey
#set bouncemail
# Comcast email is zdonkey but computer account is just donkey
poll with proto POP3 and options no dns
user 'zdonkey' with pass "somethin35" is 'donkey' here options ssl sslcertck sslcertpath '/home/donkey/.certs'
# currently not used
mda '/usr/bin/procmail -d %T'


TIP 55:

cp --parents. What does this option do?

Assume you have the following directory structure

|-- a
| `-- b
| |-- c
| | `-- d
| | |-- file1
| | `-- file2
| `-- x
| `-- y
| `-- file3
`-- newdir

Issue the following command:

$ cp --parents ./a/b/c/d/* ./newdir/

Now you have the following:

|-- a
| `-- b
| |-- c
| | `-- d
| | |-- file1
| | `-- file2
| `-- x
| `-- y
| `-- file3
`-- newdir
`-- a
`-- b
`-- c
`-- d
|-- file1
`-- file2

Note that you can't do this with "cp -r" because you'd pickup
the x directory and its contents.

You probably want to use the "cp --parents" command for directory
surgery, which you need to be very specific on what you cut and

TIP 56:

Quickly Locating files.

The "locate" command quickly searches the indexed database for files. It just
gives the name of the files; but, if you need more information use it as follows

$ locate document|xargs ls -l

The "locate" database may only get updated every 24 hours. For more recent finds,
use the "find" command.

TIP 57:

Using the "find" Command.

List only directories, max 2 nodes down that have "net" in the name

$ find /proc -type d -maxdepth 2 -iname '*net*'

Find all *.c and *.h files starting from the current "." position.

$ find . \( -iname '*.c' -o -iname '*.h' \) -print

Find all, but skip what's in "/CVS" and "/junk". Start from "/work"

$ find /work \( -iregex '.*/CVS' -o -iregex '.*/junk' \) -prune -o -print

Note -regex and -iregex work on the directory as well, which means
you must consider the "./" that comes before all listings.

Here is another example. Find all files except what is under the CVS, including
CVS listings. Also exclude "#" and "~".

$ find . -regex '.*' ! \( -regex '.*CVS.*' -o -regex '.*[#|~].*' \)

Find a *.c file, then run grep on it looking for "stdio.h"

$ find . -iname '*.c' -exec grep -H 'stdio.h' {} \;
sample output --> ./prog1.c:#include

Looking for the disk-hog on the whole system?

$ find / -size +10000k 2>/dev/null

Looking for files changed in the last 24 hours? Make sure you add the
minus sign "-1", otherwise, you will only find files changed exactly
24 hours from now. With the "-1" you get files changed from now to 24

$ find . -ctime -1 -printf "%a %f\n"
Wed Oct 6 12:51:56 2004 .
Wed Oct 6 12:35:16 2004 How_to_Linux_and_Open_Source.txt

Or if you just want files.

$ find . -type f -ctime -1 -printf "%a %f\n"

Details on file status change in the last 48 hours, current directory. Also note "-atime -2").

$ find . -ctime -2 -type f -exec ls -l {} \;

NOTE: if you don't use -type f, you make get "." returned, which
when run through ls "ls ." may list more than what you want.

Also you may only want the current directory

$ find . -ctime -2 -type f -maxdepth 1 -exec ls -l {} \;

To find files modified within the last 5 to 10 minutes

$ find . -mmin +5 -mmin -10

For more example "find" commands, reference the following looking
for the latest version of "bashscripts.x.x.x.tar.gz":

See "TIP 71" for examples of find using the inode feature. " $ find . -inum -exec rm -- '{}' \; "

If you don't want error messages, or need to redirect error messages "> /dev/null 2>&1", or see
"TIP 81".

TIP 58:

Using the "rm" command.

How do you remove a file that has the name "-". For instance, if you run the command
"$ cat > - " and type some text followed by ^d, how does the "-" file get deleted?

$ rm -- -

The "--" nullifies any rm options.

How do you delete the directory "one", all it's sub-directories, and any data?

$ rm -rf ./one

Note, to selectively delete stuff on a directory, use the find command "TIP 57".
To delete by inode, see "TIP 71".

TIP 59:

Giving ownership.

How do you give the user "donkey" ownership to all directories and files under
"./fordonkey" ?

$ chown -R donkey ./fordonkey

TIP 60:

Only Permit root login -- give others a message when they try to login.

Create the file "/etc/nologin" with "nologin" containing the contents
of the message.

TIP 61:

Limits: file size, open files, pipe size, stack size, max memory size
cpu time, plus others.

To get a listing of current limits:

$ ulimit -a
core file size (blocks, -c) 0
data seg size (kbytes, -d) unlimited
file size (blocks, -f) unlimited
max locked memory (kbytes, -l) unlimited
max memory size (kbytes, -m) unlimited
open files (-n) 1024
pipe size (512 bytes, -p) 8
stack size (kbytes, -s) 8192
cpu time (seconds, -t) unlimited
max user processes (-u) 8179
virtual memory (kbytes, -v) unlimited

Note as a user you can decrease your limits in the current
shell session; but, you cannot increase. This can be ideal
for testing programs. But, first you may want to create
another shell "sh" so that you can "go back to where started".

$ ulimit -f 10

Now try

$ yes >> out
File size limit exceeded

To set limits on users, make changes to "/etc/security/limits.conf"

bozo - maxlogins 1

Will keep bozo from loging in more than once.

To list hard limits:

$ ulimit -Ha

To list soft limits:

$ ulimit -Sa

To restrict user access by time, day make changes to

Also take a look at "/etc/profile" to see what other changes
can be made, plus take a look under "/etc/security/*.conf" for
other configuration files.

TIP 62:

Stupid "cat" Tricks.

Also see (TIP 43 and TIP 235).

If you have multiple blank lines that you want to squeeze down to
one line, then, try the following:

$ cat -s

Want to number the lines?

$ cat -n

Want to show tabs?

$ cat -t

Need to mark end of lines by "$"? The following was suggested by (Amos Shapira)

$ cat -e

Want to see all the ctl characters?

/* ctlgen.c
Program to generate ctl characters.


gcc -o ctlgen ctlgen.c


./ctlgen > mout

Now see the characters:

cat -v mout

Here's a sample output:

$ cat -v mout|tail
test M-v
test M-w
test M-x
test M-y
test M-z
test M-{
test M-|
test M-}
test M-~
test M-^?

int main()
int i;

for(i=0; i < 256; ++i)
printf("test %c \n",i);

return 0;

TIP 63:

Guard against SYN attacks and "ping".

As root do the following:

echo 1 > /proc/sys/net/ipv4/tcp_syncookies

Want to disable "ping" ?

echo 1 > /proc/sys/net/ipv4/icmp_echo_ignore_all

Disable broadcast/multicast "ping" ?

echo 1 > /proc/sys/net/ipv4/icmp_echo_ignore_broadcasts

And to enable again:

echo 0 > /proc/sys/net/ipv4/icmp_echo_ignore_all

TIP 64:

Make changes to .bash_profile and need to update the current session?

$ source .bash_profile

With the above command, the user does not have to logout.

TIP 65:

What are the Special Shell Variables?

$# The number of arguments.
$@ All arguments, as separate words.
$* All arguments, as one word.
$$ ID of the current process.
$? Exit status of the last command.
$0,$1,..$9,${10},${11}...${N} Positional parameters. After "9" you must use the ${k} syntax.

Note that 0 is true. For example if you execute the following, which is true you get zero.

$ [[ -f /etc/passwd ]]
$ echo $?
And the following is false, which returns a 1.

$ [[ -f /etc/passwdjabberwisnohere ]]
$ echo $?

So true=0 and false=1.

Sample program "mdo" to show the difference between "$@" and "$*"

function myarg
echo "$# in myarg function"
echo -e "$# parameters on the cmd line\n"
echo -e "calling: myarg \"\$@\" and myarg \"\$*\"\n"
myarg "$@"
myarg "$*"
echo -e "\ncalling: myarg \$@ and myarg \$* without quotes\n"
myarg $@
myarg $*

The result of running "./mdo one two". Note that when quoted, myarg "$*",
returns 1 ... all parameters are smushed together as one word.

[chirico@third-fl-71 theBook]$ ./mdo one two
2 parameters on the cmd line

calling: myarg "$@" and myarg "$*"

2 in myarg function
1 in myarg function

calling: myarg $@ and myarg $* without quotes

2 in myarg function
2 in myarg function

Example program "mdo2" shows how the input separator can be changed.

echo -e "$*\n"
echo -e "$*\n"
echo -e "$*\n"
echo -e "$*\n"

[chirico@third-fl-71 theBook]$ ./mdo2 one two three four five
one two three four five




TIP 66:

Replace all "x" with "y" and all "y" with "x" in file data.

$ cata data
x y
y x

$ tr "xy" "yx" < data
y x
x y

TIP 67:

On a Linux 2.6.x Kernel, how do you directly measure disk activity,
and where is this information documented?

o The information is documented in the kernel source

o The new way of getting this info in 2.6.x is
$ cat /sys/block/hda/stat
151121 5694 1932358 796675 37867 76770 916994 8353762 0 800672 9150437

Field 1 -- # of reads issued
This is the total number of reads completed successfully.
Field 2 -- # of reads merged, field 6 -- # of writes merged
Reads and writes which are adjacent to each other may be merged for
efficiency. Thus two 4K reads may become one 8K read before it is
ultimately handed to the disk, and so it will be counted (and queued)
as only one I/O. This field lets you know how often this was done.
Field 3 -- # of sectors read
This is the total number of sectors read successfully.
Field 4 -- # of milliseconds spent reading
This is the total number of milliseconds spent by all reads (as
measured from __make_request() to end_that_request_last()).
Field 5 -- # of writes completed
This is the total number of writes completed successfully.
Field 7 -- # of sectors written
This is the total number of sectors written successfully.
Field 8 -- # of milliseconds spent writing
This is the total number of milliseconds spent by all writes (as
measured from __make_request() to end_that_request_last()).
Field 9 -- # of I/Os currently in progress
The only field that should go to zero. Incremented as requests are
given to appropriate request_queue_t and decremented as they finish.
Field 10 -- # of milliseconds spent doing I/Os
This field is increases so long as field 9 is nonzero.
Field 11 -- weighted # of milliseconds spent doing I/Os
This field is incremented at each I/O start, I/O completion, I/O
merge, or read of these stats by the number of I/Os in progress
(field 9) times the number of milliseconds spent doing I/O since the
last update of this field. This can provide an easy measure of both
I/O completion time and the backlog that may be accumulating.

Note, this is device specific.

TIP 68:

Passing Outbound Mail, plus Masquerading User and Hostname.

Here's a specific example:

How does one send and receive Comcast email from a home Linux box,
which uses Comcast as the ISP, if the local account on the Linux
box is different from the Comcast email. For instance, the
account on the Linux box is "chirico@third-fl-71" and the Comcast
email account is "". Note both the hostname and
username are different.

So, the user "chirico" using "mutt", "elm" or any email program would
like to send out email to say ""; yet, donkey would
see the email from "" and not "chirico@third-fl-71"
but chirico@third-fl-71 would get the replies.

For a full description of how to solve this problem, including related
"", "site.config.m4", "genericstable", "genericsdomain",
".procmailrc", and ".forward" files, reference the following:

Included in the above link are instructions for building sendmail with

TIP 69:

How do you remove just the last 2 lines from a file and save the result?

$ sed '$d' file | sed '$d' > savefile

Or, as Amos Shapira pointed out, it's much easier with the head command.

$ head -2 file

And, of course, removing just the last line

$ sed '$d' file > savefile


How do you remove extra spaces at the end of a line?

$ sed 's/[ ]*$//g'

How do you remove blank lines, or lines with just spaces and tabs,
saving the origional file as file.backup?

$ perl -pi.backup -e "s/^(\s)*\n//" file

Or, you may want to remove empty spaces and tabs at the end of a line

$ perl -pi.backup -e "s/(\s)*\n/\n/" file

Or, you may want to converts dates of the format 01/23/2007 to the
format 2007-01-23. This is MySQL's common date format.

$ perl -pi.backup -e "s|(\d+)/(\d+)/(\d+)|\$3-\$2-\$1|" file

Note, you need a backslash \$3,\$2,\$1 so as to not get bash shell

TIP 70:

Generating Random Numbers.

$ od -vAn -N4 -tu4 < /dev/urandom

TIP 71:

Deleting a File by it's Inode Value.

See (PROGRAMMING TIP 5) for creating the file, or

$ cat > '\n\n\n\n\n\n\n'
type some text

To list the inode and display the characters.

$ ls -libt *

To remove by inode. Note the "--" option. This
will keep any special characters in the file from being
interpreted at "rm" options.

$ find . -inum -exec rm -- '{}' \;

Or to check contents

$ find . -inum -exec cat '{}' \;


TIP 72:

Sending Attachments Using Mutt -- On the Command Line.

$ mutt -s "See Attachment" -a file.doc < message.txt

or just the message:

$ echo | mutt -a sample.tar.gz


Also see (TIP 51).

TIP 73:

Want to find out what functions a program calls?

$ strace

Try this with "topen.c" (see PROGRAMMING TIP 5)

$ strace ./topen

TIP 74:

RPM Usage Summary.

Install. Full filename is needed.

$ rpm -ivh Fedora/RPMS/postgresql-libs-7.4.2-1.i386.rpm

To view list of files installed with a particular package.

$ rpm -ql postgresql-libs

Or, to get the file listing from a package that is not installed use the
"-p" option.

$ rpm -pql /iso0/Fedora/RPMS/libpcap-0.8.3-7.i386.rpm

For dependencies listing, use the "R" option.

$ rpm -qpR /iso0/Fedora/RPMS/libpcap-0.8.3-7.i386.rpm
kernel >= 2.2.0
rpmlib(CompressedFileNames) <= 3.0.4-1
rpmlib(PayloadFilesHavePrefix) <= 4.0-1

To check the integrity, use the "-K" option.

$ rpm -K /iso0/Fedora/RPMS/libpcap-0.8.3-7.i386.rpm
/iso0/Fedora/RPMS/libpcap-0.8.3-7.i386.rpm: (sha1) dsa sha1 md5 gpg OK

To list all packages installed.

$ rpm -qa

To find out which file a package belongs to.

$ rpm -qf /usr/lib/

To uninstall a package

$ rpm -e

For building rpm packages reference the following:

To verify md5 sum so that you know it downloaded ok

$ rpm -K *.rpm

The following is a good reference:

TIP 75:

Listing Output from a Bash Script.

Add "set -x"

set -x

Will list the files and output as follows:

+ ls
ChangeLog CVS data test
+ date
Thu Jul 1 20:41:04 EDT 2004

TIP 76:

Using wget.

Grap a webpage and pipe it to less. For example suppose you wanted to pipe the
contents of all these tips, directly from the web.

$ wget -O -|less

TIP 77:

Finding IP address and MAC address.

$ /sbin/ifconfig

Note the following output "eth0" and "eth0:1" which means
two IP addresses are tied to 1 NIC (Network Interface Card).

eth0 Link encap:Ethernet HWaddr 00:50:DA:60:5B:AD
inet addr: Bcast: Mask:
RX packets:982757 errors:116 dropped:0 overruns:0 frame:116
TX packets:439297 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:693529078 (661.4 Mb) TX bytes:78400296 (74.7 Mb)
Interrupt:10 Base address:0xa800

eth0:1 Link encap:Ethernet HWaddr 00:50:DA:60:5B:AD
inet addr: Bcast: Mask:
RX packets:982757 errors:116 dropped:0 overruns:0 frame:116
TX packets:439299 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:693529078 (661.4 Mb) TX bytes:78400636 (74.7 Mb)
Interrupt:10 Base address:0xa800

lo Link encap:Local Loopback
inet addr: Mask:
RX packets:785 errors:0 dropped:0 overruns:0 frame:0
TX packets:785 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:2372833 (2.2 Mb) TX bytes:2372833 (2.2 Mb)

TIP 78:

DOS to UNIX and UNIX to DOS.

$ dos2unix file.txt

And to go the other way from UNIX to DOS

$ unix2dos unixfile

See the man page, since there are MAC options.

NOTE: If you're working file DOS files, you'll probably want to use
"zip" instead of "gzip" so users on Windows can unzip them.

$ zip test.txt

TIP 79:

Need to Run Interactive Commands? Try "expect".

This simple example waits for the input "hi", in some form before
returning, immediately, "hello there!". Otherwise, it will wait for
60 seconds, then, return "hello there!".

set timeout 60
expect "hi\n"
send "hello there!\n"


TIP 80:

Using PHP as a Command Line Scripting Language.

The following will grab the complete file from slashdot.

#!/usr/bin/php -q

$fileName = "";
$rss = file($fileName) or die ("Cannot open file $fileName\n");
for ($index=0; $index < count($rss); $index++)
echo $rss[$index];

Note, if you want an example that parses the XML of
slashdot, then, download the following:

TIP 81:

Discarding all output -- including stderr messages.

$ ls > /dev/null 2>&1

Or sending all output to a file

$ someprog > /tmp/file 2>&1

Sometimes, find displays a lot of errors when searching through
directories that the user doesn't have access to. To discard
error messages "stderr", which is normally file descripter "2"
work the following:

$ find / -iname 'stuff' 2>/dev/null

or to pipe results elsewhere

$ find / -iname 'stuff' > /tmp/results_of_find 2>/dev/null

Also see (TIP 118).

TIP 82:

Using MIX. D. Knuth's assembly language/machine-code instruction set used in
his books to illustrate his algorithms.

Download the source:

$ ./configure
$ make
$ make install

Documentation can be found at the following link. The link on
sourceforge is not correct, but, the one below works.

TIP 83:

Gnuplot [ ].

This software is ideal for printing graphs.

gnuplot> set term png
gnuplot> set output 'testcos.png'
gnuplot> plot cos(x)*sin(x)
gnuplot> exit

Or the following command can be put into "file"

$ cat > file
set term png
set output 'testcos.png'
plot cos(x)*sin(x)

Then, run as follows:

$ gnuplot file

Or, suppose you have the following file "/home/chirico/data". Comments
with "#" are not read by gnuplot.

# File /home/chirico/data
2005-07-26 1 2.3 3
2005-07-27 2 3.4 5
2005-07-28 3 4 6.6
2005-07-29 4 6 2.5

And you have the following new "file"

set term png
set xdata time
set timefmt "%Y-%m-%d "
set format x "%Y/%m/%d"
set output '/var/www/html/chirico/gnuplot/data.png'
plot '/home/chirico/data' using 1:2 w linespoints title '1st col', \
'/home/chirico/data' using 1:3 w linespoints title '2nd col', \
'/home/chirico/data' using 1:4 w linespoints title '3rd col'

You can now get a graph of this data running the following:

$ gnuplot file

TIP 84:

CPU Information - speed, processor, cache.

$ cat /proc/cpuinfo

processor : 0
vendor_id : GenuineIntel
cpu family : 15
model : 2
model name : Intel(R) Pentium(R) 4 CPU 2.20GHz
stepping : 9
cpu MHz : 2193.221
cache size : 512 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 2
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr
bogomips : 4325.37

"bogomips" is a rough but good way to quickly compare two computer speeds. True it's a
bogus reading; but, a "good enough" for government work calculation. See (TIP 10) for
"vmstat" and "iostat".

TIP 85:

POVRAY - Making Animated GIFs

To see this in action, reference:

These are the basic command to create:

$ povray orbit.ini -Iorbit.pov
$ convert -delay 20 *.ppm orbit.gif

By the way, convert is a program from imagemagick, and it can
be downloaded from ( ).

The following is "orbit.pov"

#include ""
#include ""
#include ""
#include ""
#include ""
#include ""

camera {
location <>
look_at <>
focal_point <0,>
blur_samples 20

light_source {
color White
area_light <2,0,0>,<0,0,2>, 2, 2
adaptive 1
fade_distance 8
fade_power 1

sky_sphere {

plane { <0,>, -1
texture {
pigment {
checker color Blue, color White
finish {Phong_Glossy}
#declare ball0=
sphere {
<0.5,>, 1
texture {
pigment {Yellow}

#declare ball1=
sphere {
<3,>, 0.5
texture {
pigment {Blue}

#declare ball2=
sphere {
<3,>, 1
texture {
pigment {Green}

object {ball0 rotate 360*clock*y}
object {ball1 rotate 720*clock*y}
object {ball2 rotate 360*(1 - clock)*y}

And, "orbit.ini" follows:






TIP 86:

GPG -- GnuPG

(SCRIPT 4) on following link:

Generage key:

$ gpg --gen-key

Generate public key ID and fingerprint

$ gpg --fingerprint

Get a list of keys:

$ gpg --list-keys

pub 1024D/A11C1499 2004-07-15 Mike Chirico
sub 1024g/E1A3C2B3 2004-07-15


$ gpg -r Mike --encrypt sample.txt

This will produce "sample.txt.asc", which is a binary file. Note, I can use "Mike" because that's the
name on the list of keys. Again, it will be a binary file.

Encrypt using "ASCII-armored text" (--armor), which is probably what you want when sending "in" the body of an
email, or some document.

$ gpg -r Mike --encrypt --armor sample.txt
$ gpg -r Mike -e -a sample.txt
$ gpg --output somefile.asc --armor -r Mike --encrypt --armor sample.txt

The above 3 statements will still produce "sample.txt.asc", but look at it, or "$ cat sample.txt.asc" without
fear, since there are no binary characters. Yes, you could even compile a program "$ g++ -o test test.c" , then,
"$ gpg --output test.asc -r Mike --encrypt --armor test". However, when decrypting make sure to pipe
the results.

$ gpg --decrypt test.asc > test

Export "public" key:

$ gpg --armor --export Mike > m1.asc

Signing the file "message.txt":

$ gpg --clearsign message.txt

Sending the key to the "key-server"

First, list the keys.

$ gpg --list-keys
v------------------ Use this with "0x" in front -------
pub 1024D/A11C1499 2004-07-15 Mike Chirico |
sub 1024g/E1A3C2B3 2004-07-15 |
$ gpg --send-keys 0xA11C1499

The above sends it to the keyserver defined in "/home/chirico/.gnupg/gpg.conf". Other key servers:

When you go to your user-group meetings, you need to bring 2 forms of ID, and
list your Key fingerprint. Shown below is the command for getting this fingerprint.

$ gpg --fingerprint
pub 1024D/A11C1499 2004-07-15
Key fingerprint = 9D7F C80D BB7B 4BAB CCA4 1BE9 9056 5BEC A11C 1499
uid Mike Chirico (
sub 1024g/E1A3C2B3 2004-07-15

Receving keys:

The following will retrieve my key

$ gpg --recv-keys 0xA11C1499

Special Note: If you get the following error "GPG: Warning: Using Insecure Memory" , then,
" chmod 4755 /path/to/gpg" to setuid(root) permissioins on the gpg binary.

NOTE: If using mutt, just before sending with the "y" option, hit "p" to sign or encrypt.

It's possible to create a gpg/pgp email from the command line. For a tutorial on this,
reference (SCRIPT 4) at the following link:

TIP 87:

Working with Dates: Steffen Beyer has developed a Perl and C module for working with dates

This softare can be downloaded from the following location:

$ wget
$ tar -xzvf Date-Calc-5.3.tar.gz
$ cd Date-Calc-5.3
$ cp ./examples/cal.c .
$ gcc cal.c DateCalc.c -o mcal

The file cal.c contains sample function calls from DateCalc.c. Note, "DateCalc.c"
is just a list of functions and includes for "DateCalc.h" and "ToolBox.h".

Or, and this may be easier, just download the following:

The above link contains a few examples.

TIP 88:

Color patterns for mutt.

The colors can be changed in the /home/user/.muttrc file. The first field begins with
color, the second field is the foreground color, and the third field is the background
color, or default.

An example .muttrc for colors:

# color patterns for mutt
color normal white black # normal text
color indicator black yellow # actual message
color tree brightmagenta default # thread arrows
color status brightyellow default # status line
color error brightred default # errors
color message magenta default # info messages
color signature magenta default # signature
color attachment brightyellow red # MIME attachments
color search brightyellow red # search matches
color tilde brightmagenta default # ~ at bottom of msg
color markers red default # + at beginning of wrapped lines
color hdrdefault cyan default # default header lines
color bold red default # hiliting bold patterns in body
color underline green default # hiliting underlined patterns in body
color quoted cyan default # quoted text
color quoted1 magenta default
color quoted2 red default
color quoted3 green default
color quoted4 magenta default
color quoted5 cyan default
color quoted6 magenta default
color quoted7 red default
color quoted8 green default
color quoted9 cyan default
color body cyan default "((ftp|http|https)://|news:)[^ >)\"\t]+"
color body cyan default "[-a-z_0-9.+]+@[-a-z_0-9.]+"
color body red default "(^| )\\*[-a-z0-9*]+\\*[,.?]?[ \n]"
color body green default "(^| )_[-a-z0-9_]+_[,.?]?[\n]"
color body red default "(^| )\\*[-a-z0-9*]+\\*[,.?]?[ \n]"
color body green default "(^| )_[-a-z0-9_]+_[,.?]?[ \n]"
color index cyan default ~F # Flagged
color index red default ~N # New
color index magenta default ~T # Tagged
color index cyan default ~D # Deleted

Also see (TIP 190)

TIP 89:

ps command in detail

Here are the possible codes when using state "$ ps -e -o state,cmd"

D uninterruptible sleep (usually IO)
R runnable (on run queue)
S sleeping
T traced or stopped
Z a defunct ("zombie") process

< high-priority (not nice to other users)
N low-priority (nice to other users)
L has pages locked into memory (for real-time and custom IO)
s is a session leader
l is multi-threaded (using CLONE_THREAD, like NPTL pthreads do)
+ is in the foreground process group

For instance:

Note that the -o is for user defined, and -e is for select
all process.

$ ps -e -o pid,state,start,time,etime,cmd

9946 S 15:40:45 00:00:00 02:23:29 /bin/bash -i
9985 T 15:41:24 00:00:01 02:22:50 emacs mout2
10003 T 15:43:59 00:00:00 02:20:15 emacs NOTES
10320 T 17:38:42 00:00:00 25:32 emacs stuff.c

You may want to command below, without the -e, which will give the
process only under the current terminal.

$ ps -o pid,state,start,time,etime,cmd

Want to find what 's impacting your load?

$ ps -e -o %cpu,pid,state,start,time,etime,%cpu,%mem,cmd|sort -rn|less

$ ps aux

root 1 0.0 0.0 1380 480 ? S Aug04 0:00 init [3]
root 2 0.0 0.0 0 0 ? SWN Aug04 0:00 [ksoftirqd/0]
root 3 0.0 0.0 0 0 ? SW< Aug04 0:00 [events/0]
root 4 0.0 0.0 0 0 ? SW< Aug04 0:00 [khelper]

Or, if you want to see the environment add the -e option

$ ps aeux

chirico 2735 0.0 0.1 4400 1492 pts/0 S Aug04 0:00 -bash USER=chirico LOGNAME=chirico HOME=/home/chirico PATH=/usr/
chirico 2771 0.0 0.0 4328 924 pts/0 S Aug04 0:00 screen -e^Pa -D -R HOSTNAME=third-fl-71.localdomain TERM=xterm S
chirico 2772 0.0 0.6 9476 6352 ? S Aug04 0:54 SCREEN -e^Pa -D -R HOSTNAME=third-fl-71.localdomain TERM=xterm S
chirico 2773 0.0 0.1 4432 1548 pts/1 S Aug04 0:10 /bin/bash STY=2772.pts-0.third-fl-71 TERM=screen TERMCAP=SC|scre
chirico 2797 0.0 0.1 4416 1496 pts/2 S Aug04 0:00 /bin/bash STY=2772.pts-0.third-fl-71 TERM=screen TERMCAP=SC|scre
root 2821 0.0 0.0 4100 952 pts/2 S Aug04 0:00 su -
root 2822 0.0 0.1 4384 1480 pts/2 S Aug04 0:00 -bash
chirico 2862 0.0 0.1 4428 1524 pts/3 S Aug04 0:00 /bin/bash STY=2772.pts-0.third-fl-71 TERM=screen TERMCAP=SC|scre
sporkey 2946 0.0 0.2 6836 2960 ? S Aug04 0:15 fetchmail
chirico 2952 0.0 0.1 4436 1552 pts/5 S Aug04 0:00 /bin/bash STY=2772.pts-0.third-fl-71 TERM=screen TERMCAP=SC|scre
chirico 3880 0.0 0.1 4416 1496 pts/6 S Aug05 0:00 /bin/bash STY=2772.pts-0.third-fl-71 TERM=screen TERMCAP=SC|scre
root 3904 0.0 0.0 4100 956 pts/6 S Aug05 0:00 su - donkey
donkey 3905 0.0 0.1 4336 1452 pts/6 S Aug05 0:00 -bash
donkey 3938 0.0 0.2 6732 2856 ? S Aug05 0:14 fetchmail
chirico 3944 0.0 0.1 4416 1496 pts/7 S Aug05 0:00 /bin/bash STY=2772.pts-0.third-fl-71 TERM=screen TERMCAP=SC|scre

There is also a -f "forrest" option. Also note below " -bash" is the start of a login shell.

$ ps aeuxwwf

The ww option above gives a wide format with all variables. Use the above command if you plan
to parse through a Perl script. Otherwise, it may be easier to do a quick read using the command
below, without "ww".

$ ps aeuxf

root 2339 0.0 0.1 3512 1444 ? S Dec01 0:00 /usr/sbin/sshd
root 25651 0.0 0.1 6764 1980 ? S Dec23 0:00 \_ /usr/sbin/sshd
chirico 25653 0.0 0.2 6840 2236 ? S Dec23 0:14 \_ /usr/sbin/sshd
chirico 25654 0.0 0.1 4364 1440 pts/4 S Dec23 0:00 \_ -bash USER=chirico LOGNAME=chirico HOME=/home/chirico
chirico 25690 0.0 0.0 4328 920 pts/4 S Dec23 0:00 \_ screen -e^Pa -D -R HOSTNAME=third-fl-71.localdomain TERM=xterm
root 2355 0.0 0.0 2068 904 ? S Dec01 0:00 xinetd -stayalive -pidfile /var/run/

It is also possible to list the process by command line. For example, the following command will only list the emacs

$ ps -fC emacs
chirico 5049 5020 0 May11 pts/13 00:00:00 emacs -nw Notes
chirico 12368 5104 0 May12 pts/18 00:00:00 emacs -nw dnotify.c
chirico 19792 18028 0 May13 pts/20 00:00:00 emacs -nw hello.c
chirico 14034 27367 0 18:52 pts/8 00:00:00 emacs -nw How_to_Linux_and_Open_Source.txt

You may also want to consider using top in batch mode. Here the "-n 1" means refresh once,
and the "b" is for batch. The "fmt -s" is to put it in a more readable format.

$ top -n 1 b |fmt -s >>statfile

TIP 90:

Learning Assembly.

Once you have written the source, assuming the file is "exit.s", it can be compiled as follows:

$ as exit.s -o exit.o
$ ld exit.o -o exit

Here is the program:

#INPUT: none
#OUTPUT: returns a status code. This can be viewed
# by typing
# echo $?
# after running the program
# %eax holds the system call number
# (this is always the case)
# %ebx holds the return status
.section .data
.section .text

.globl _start
movl $1, %eax # this is the linux kernel command
# number (system call) for exiting
# a program
movl $0, %ebx # this is the status number we will
# return to the operating system.
# Change this around and it will
# return different things to
# echo $?
int $0x80 # this wakes up the kernel to run
# the exit command

After running this program, you can get the exit code.

$ exit $?

That is about all it does; but, get the book for more details. The
book is free.

TIP 91:

Creating a sandbox for reiserfstune,debugreiserfs and ACL. Also see TIP 4.

Assume you have a reisers files system created from a disk file, which
means you have done something like the following:

# dd if=/dev/zero of=disk-rfs count=102400
# losetup /dev/loop4 ./disk-rfs
# mkfs -t reiserfs /dev/loop4
# mkdir /fs2
# mount -o loop,acl ./disk-rfs /fs2

Now, you can run reiserfstune. But, first you will need to umount fs2

# umount /fs2
# reiserfstune ./disk-rfs

Or you can run the debug command

# debugreiserfs -J ./disk-rfs

Now, suppose you run through a lot of the debug options on and you destroy this file.

You can recreate the file and delete the loop device.

# dd if=/dev/zero of=disk-rfs count=102400
# losetup -d /dev/loop4
# mount -o loop,acl ./disk-rfs /fs2

Now, try working with some of the ACL options - you can only do this
with the latest kernel and tools -- Fedora Core 2 will work.

Assume you have 3 users, donkey, chirico and bozo2. You can give
everyone rights to this file system as follows:

# setfacl -R -m d:u:donkey:rwx,d:u:chirico:rwx,d:u:bozo2:rwx /fs2

TIP 92:

SpamAssassin - Setup.

Step 1.

Installing the SpamAssassin CPAN utility. You will need to do this
as root.

$ su -

Once you have root privileges invoke cpan.

# perl -MCPAN -e shell


Now install with prerequisites policy set to ask.

cpan> o conf prerequisites_policy ask

cpan> install Mail::SpamAssassin

You will get lots of output as the necessary modules are downloaded and
compiled and installed.

Step 2.


Edit the following "/etc/mail/spamassassin/"

Here is a look at my file

$ cat /etc/mail/spamassassin/

# This is the right place to customize your installation of SpamAssassin.
# See 'perldoc Mail::SpamAssassin::Conf' for details of what can be
# tweaked.
# rewrite_subject 0
# report_safe 1
# trusted_networks 212.17.35.

# Below added from book
# You may want to set this to 5, then, work your way down.
# Currently I have this 3
required_hits 3

# This determines how spam is reported. Currently safe email is reported
# in the message.
report_safe 1

# The will rewrite the tag of the spam message.
rewrite_subject 1

# By default, SpamAssassin will run RBL checks. If your ISP already
# does this, set this to 1.
skip_rbl_checks 0

Step 3.

Update .procmail.

You should update the .procmail file as follows. Here is my /home/chirico/.procmail file.

$ cat /home/chirico/.procmailrc

# Must have folder MailTRASH

# Will get everything from this mail
* ^From:.*

# Spamassassin
* <300000


TIP 93:

Make Graphs: using dot and neato.

$ dot -Tpng dotfile -o myout.png

To see the output reference the following:

Where "dotfile" is the following:

$ cat dotfile

digraph g
node [shape = record];

node0 [ label =" stuff | J | "];
node1 [ label =" | E | "];
node4 [ label =" | C | "];
node6 [ label =" | I | "];
node2 [ label =" | U | "];
node5 [ label =" | N | "];
node9 [ label =" | Y | "];
node8 [ label =" | W | "];
node10 [ label =" | Z | "];
node7 [ label =" | A | "];
node3 [ label =" | G | "];

"node0":f0 -> "node1":f1;
"node0":f2 -> "node2":f1;

"node1":f0 -> "node4":f1;
"node1":f2 -> "node6":f1;
"node4":f0 -> "node7":f1;
"node4":f2 -> "node3":f1;

"node2":f0 -> "node5":f1;
"node2":f2 -> "node9":f1;

"node9":f0 -> "node8":f1;
"node9":f2 -> "node10":f1;

Checkout the following article:

To download this software

TIP 94:

Makefile: working with conditions

First note that all the indentations of the file must be
a single tab. There cannot be any spaces, or make will
not run.

$ cat Makefile

# Compiler flags
sqliteLIB := $(shell ls /usr/local/lib/
sqlite3LIB := $(shell ls /usr/local/lib/
# all assumes sqlite and sqlite3 are installed

ifeq ("$(sqlite3LIB)","/usr/local/lib/")
@echo -e "True -- we found the file"
@echo "False -- we did not find the file"

So, if I run make I will get the following output.

$ make
True -- we found the file

This is because I have a file /usr/local/lib/ on my system.
Note how the assignment is made, with the shell command

sqlite3LIB := $(shell ls /usr/local/lib/

TIP 95:

Bash: Conditional Expressions

if [ -e /etc/ntp.conf ]
echo "You have the ntp config file"
echo "You do not have the ntp config file"

Now using an AND condition inside the [ ]. By the way, above, you
can put the "then" on the same line as the if "if [ -e /etc/ntp.conf ]; then"
as long as you use the ";".

if [ \( -e /etc/ntp.conf \) -a \( -e /etc/ntp/ntpservers \) ]
echo "You have ntp config and ntpservers"
elif [ -e /etc/ntp.conf ]; then
echo " You just have ntp.conf "
elif [ -e /etc/ntp/ntpservers ]; then
echo " You just have ntpservers "
echo " you have neither ntp.conf or ntpservers"

A few things to note above. Else if statement is written as "elif", and when
dealing with "(" you will need to insert "\(". By the way "-o" can replace "-a"
and the "-o" is for OR condition. AND can be done as follows too.

if [ -e /etc/ntp.conf ] && [ -e /etc/ntp/ntpservers ]
echo "You have ntp config and ntpservers"
elif [ -e /etc/ntp.conf ]; then
echo " You just have ntp.conf "
elif [ -e /etc/ntp/ntpservers ]; then
echo " You just have ntpservers "
echo " you have neither ntp.conf or ntpservers"

Conditional Expressions (files).

-b file True if file exists and is a block file
-c file True if file exists and is a character device file
-d file True if file exists and is a directory
-e file True if file exists
-f file True if file exists and is a regular file
-g file True if file exists and is set goup id
-G file True if owned by the effective group ID

-k file True if "sticky" bit is set and file exists
-L file True if file exists and is a symbolic link
-n string True if string is non-null

-O file Ture if file exists and is owned by the effective user ID

-p file True if file is a named pipe (FIFO)
-r file True if file is readable
-s file True if file has size > 0
-S file True if file exists and is a socket

-t file True if file is open and refers to a terminal.
-u file True if setuid bit is set
-w file True if file exists and is writeable
-x file True if file executable
-x dir True if directory can be searched

file1 -nt file2 True if file1 modification date newer than file2
file1 -ot file2 True if file1 modification date older than file2
file1 -ef file2 True if file1 and file2 have same inode

Conditional Expressions (Integers).

-lt Less than
-le Less than or equal
-eq Equal
-ge Greater than or equal
-gt Greater than
-ne Not equal

Example usage.

while read num value; do
if [ $num -gt 2 ]; then
echo $value
} < somefile

Conditional Expressions (Strings).

str1 = str2 str1 matches str2
str1 != str2 str1 does not matches str2
str1 < str2 str1 is less than str2
str1 > str2 str1 is greater than str2
-n str1 str1 is not null (length greater than 0)
-z str1 str1 is null (las length 0)

TIP 96:

CVS: Working with cvs


To create a repository, and this is normally done by the system admin. This
is NOT creating a project to checkout, but the location where everything
will be stored! The initial repository!

cvs -d repository_root_directory init

Or here is a specific example:

cvs -d /work/cvsREPOSITORY/ init

Creating a directory tree from scratch. For a new project, the easiest thing to
do is probably to create an empty directory structure, like this:

$ mkdir sqlite_examples
$ mkdir sqlite_examples/man
$ mkdir sqlite_examples/testing

After that, you use the import command to create the
corresponding (empty) directory structure inside the repository:

$ cd
$ cvs -d repository_root_directory import -m "Created directory structure" yoyodyne/dir yoyo start

Or, here is a specific example.

$ cd sqlite_examples
$ cvs -d /work/cvsREPOSITORY/ import -m 'test SQlite' sqlite_examples sqlite_examples start

Now, you can delete the directory sqlite_examples, or go to another directory and type
the following:

$ cvs -d /work/cvsREPOSITORY/ co sqlite_examples


1. cvsps
2. cvsreport

cvsps which you can find at

$ cvsps -f README_sqlite_tutorial.html

TIP 97:

Common vi and vim commands

Command mode ESC

dd delete
u undelete
y yank (copy to buffer)
p/P p before cursor/P after cursor

Ctl-g show current line number
shft-G end of file
n shft-G move to line n

/stuff/ search
n repeat in same direction
N repeat in opposite direction
/return repeat seach forward
?return repeat seach backward

"dyy Yank current line to buffer d
"a7yy Yank next 7 lines to buffer a
:1,7ya a Yank [ya] lines 1,7 to buffer a
:1,7ya b Yank [ya] lines 1,7 to buffer b

:5 pu b Put [pu] buffer b after line 5

"dP Put the content of buffer d before cursor
"ap Put the contents of buffer a after cursor

:1,4 w! file2 Write lines 1,4 to file2

:set nu Display line numbers
:set nonum Turns off display

:e Edit a file in a new buffer

:split new

ctl-w To move between windows
ctl-w- To change size
ctl+wv Split windows vertically
ctl-wq Close window

:only To view only 1 window

vim dictionary - put the following command in ~/.vimrc

set dictionary+=/usr/share/dict/words
set thesaurus+=/usr/share/dict/words

Now, after you type a word and to
go back in the listing


TIP 98:

Using apt-get

$ apt-get update
$ apt-get -s install <---- if everything is ok, then, remove the s

Note you may want to use dpkg to purge if you have to do a reinstall.

$ dpkg --purge exim4-base
$ dpkg --purge exim4-config
$ apt-get install exim4

$ dpkg-reconfigure exim4-config

TIP 99:

Mounting a cdrom on openbsd and installing packages

$ mkdir -p /cdrom
$ mount /dev/cd0a /cdrom
$ cd /cdrom

To add packages

$ pkg_add -v

Mounting a cdrom on linux to a user's home sub-directory:

$ mkdir -p /home/chirico/cdrom
$ mount /dev/cdrom /home/chirico/cdrom

TIP 100:

Creating a boot floppy for knoppix cd:

$ dd if=/mnt/cdrom/KNOPPIX/boot.img of=/dev/fd0 bs=1440k


For a lot of the knoppix how-to's

TIP 101:

Diction and Style Tools for Linux

$ diction mytext|less

Or, this can be done interactively

$ diction
This is more text to read and you can do with it
what you want.
(stdin):1: This is more text to read and you [can -> (do not confuse with "may")] do with it what you want.

Diction finds all sentences in a document, that contain phrases from a
database of frequently misused, bad or wordy diction. It further
checks for double words. If no files are given, the document is read
from standard input. Each found phrase is enclosed in [ ] (brackets).
Suggestions and advice, if any, are printed headed by a right arrow ->.
A sentence is a sequence of words, that starts with a capitalised word
and ends with a full stop, double colon, question mark or exclaimation
mark. A single letter followed by a dot is considered an abbreviation,
so it does not terminate a sentence. Various multi-letter abbrevia-
tions are recognized, they do not terminate a sentence as well.

TIP 102:

Using a mail alias.

Suppose all root mail on your system to go to one root account

In the following file:


Add this line


Next, run newaliases [/usr/bin/newaliases] as follows:

$ newaliases

Special note: It's possible to send mail to more than one address. Suppose you want
mail going to above, plus you want it going to user donkey
on the local system.

root: donkey

TIP 103:

Chrony - this service is similiar to ntp. It keeps accurate time
on your computer against a very accurate clock in across
a network with various time delays.


In the file "/etc/chrony/chrony.conf" add/replace the following


Next start the chrony service

$ /etc/init.d/chrony restart

Next verify that this is working. It may take 20 or 30 minutes to update
the clock.

Shell command:
# chronyc
chronyc> sourcestats
210 Number of sources = 3
Name/IP Address NP NR Span Frequency Freq Skew Std Dev
======================================================================== 2 0 64 0.000 2000.000 4000ms 2 0 66 0.000 2000.000 4000ms
FS3.ECE.CMU.EDU 2 0 64 0.000 2000.000 4000ms

It is probably best to let chrony do its work. However, if you want to
set both the hardware and software clock, the following will work:

Sets the hardware clock
# hwclock --set --date="12/10/04 10:18:05"
Sync the hardware clock to software
# hwclock --hctosys

Normally the system keep accurate time with the software clock.

TIP 104:

NFS mount


Make sure nfs is running on the server

$ /etc/init.d/nfs restart

At the server the contents of /etc/exports for
allowing 2 computers ( and
to access the home directory of this server. Note that
read write (rw) access is allowed.

$ cat /etc/exports

Or, if you have a lot of clients on 192.168.1.* then consider
the following:


Next, still at the server, run the exportfs command

$ exportfs -rv

IPTABLES (lokkit). If you're using fedora with default lokkit firewall
then you can put the following under "Other ports".

Other ports nfs:tcp nfs:udp

If the above does not work or you are not using lokkit
IPTABLES (values in /etc/sysconfig/iptables on SERVER )

# NFS Need to accept fragmented packets and may not have header
# so you will not know where they are coming from
-A INPUT -p tcp -m tcp -s -m multiport --dports 111,683,686,685,1026,2049,2219 -j ACCEPT
-A INPUT -p tcp -s -d 0/0 --dport 32765:32768 -j ACCEPT
-A INPUT -p udp -m udp -s -m multiport --dports 111,683,686,685,1026,2049,2219 -j ACCEPT
-A INPUT -p udp -s -d 0/0 --dport 32765:32768 -j ACCEPT

-A INPUT -p tcp -m tcp -s -m multiport --dports 111,683,686,685,1026,2049,2219 -j ACCEPT
-A INPUT -p tcp -s -d 0/0 --dport 32765:32768 -j ACCEPT
-A INPUT -p udp -m udp -s -m multiport --dports 111,683,686,685,1026,2049,2219 -j ACCEPT
-A INPUT -p udp -s -d 0/0 --dport 32765:32768 -j ACCEPT



$ mkdir -p /home2

$ cat /etc/fstab /home2 nfs rw 0 0

$ mount -a -t nfs

Or to do a one time mounting by hand

$ mount -t nfs /home2

Now /home2 on the client will be /home on the server



To monitor the client:

$ nfsstat -c

Also note you can "cat /proc/net/rpc/nfs" as well.

To monitor the server (note the -s instead of the -c).

$ nfsstat -s

Also note you can "cat /proc/net/rpc/nfsd" as well.

The following "cat" command is done on the NFS server, and shows which
clients are mounting. This does not go with examples above. By the way,
"root_squash" is the default, and means that root access on the clients is
denied. So, how does the client root get access to these filesystems? You have
to "su - ".

$ cat /proc/fs/nfs/exports
# Version 1.1
# Path Client(Flags) # IPs

(Reference: )

TIP 105:

Ports used for Microsoft products

To find out common port mappings, take a look at "/etc/services"

TIP 106:

Man pages: If man pages are formatting incorrectly with PuTTY, try editing
the "/etc/man.config" file with the following changes:

NROFF /usr/bin/groff -Tlatin1 -mandoc
NEQN /usr/bin/geqn -Tlatin1

(Reference TIP 7 for using man)

TIP 107:

Valgrind: check for memory leaks in your programs. (

This is how you can run it on the program "a.out" for valgrind version 2.2.0

$ valgrind --logfile=valgrind.output --tool=memcheck ./a.out

This is how you write the logfile "--log-file" for valgrind-3.0.1

$ valgrind --log-file=valgrind --leak-check=yes --tool=memcheck ./a.out

With C++ programs with gcc 3.4 and later that use STL, export GLIBCXX_FORCE_NEW
only when testing to disable memory caching. Remember to enable for production
as this will have a performance penalty. Reference

TIP 108:

Runlevel Configuring.

The program ntsysv, run as root, gives you a ncurses GUI to what will
run on your system on boot. The chkconfig program (man chkconfig) has
the ability to list which programs are set to start on the chosen
run level.

# ntsysv

# chkconfig

If at this moment you want to see what services are currently running,
then, run the following command:

# /sbin/service --status-all

Note, you can also set these manually. For example, normally you will
have files in "/etc/init.d/" that will take parameters like "start","stop"

Take a look at "/etc/init.d/mysql" this file will start and stop the
mysql daemon. So, how does know which run levels, and the order it gets
loaded in the run level to other programs? By the K and S

$ ls /etc/rc3.d/*mysql


So here on my system the start value is 85. Looking in /etc/rc3.d, which is
run level 3, any program with a lower number S84something will get loaded
before mysql.

I manually set the run level as follows for mysql.

# cd /etc/rc3.d
# ln -s ../init.d/mysql S85mysql
# ln -s ../init.d/mysql K85mysql

# cd /etc/rc5.d
# ln -s ../init.d/mysql S85mysql
# ln -s ../init.d/mysql K85mysql

Note that I could have chose other numbers as well. "ntsysv" gives
you a graphical interface.

This is a way of doing this with "chkconfig" at the command prompt.

# chkconfig --list mysqld
mysqld 0:off 1:off 2:off 3:on 4:off 5:on 6:off

Above you can see it's on. Here's how we would have turned this on with chkconfig.

# chkconfig --level 35 mysqld on


TIP 109:

File Alteration Monitor - Gamin a FAM replacement

Working with fam - file alteration monitor. Mail uses this to signify
a change in a file's status.

Below is the sample C program ftest.c which can be compiled as

$ gcc -o ftest ftest.c -lfam

You will need to work with this as root

# ./ftest


TIP 110:

glibc - this is the main library used by C, and the following
link below gives you examples on everything from sockets,math,
date and time functions, user environment, and much more.

How do you know which version of glibc you are running?

int main (void)
puts (gnu_get_libc_version ());
return 0;

TIP 111:

nslookup and dig - query Internet name servers interactively.

$ nslookup


The nslookup command will query the dns server is "/etc/resolve.conf"
However, you can force a certain dns with "- server". For example the
command below goes to the server named dilbert

$ nslookup - dilbert


dig gives you more information. You should probably use dig instead
of nslookup.

Below I am forcing the lookup from DNS of the name, and
note that the query time is return too.

$ dig @ +qr

; <<>> DiG 9.2.1 <<>> @ +qr
;; global options: printcmd
;; Sending:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 55908
;; flags: rd; QUERY: 1, ANSWER: 0, AUTHORITY: 0, ADDITIONAL: 0

; IN A

;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 55908
;; flags: qr rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 2, ADDITIONAL: 2

; IN A




;; Query time: 155 msec
;; WHEN: Thu Dec 23 07:48:23 2004
;; MSG SIZE rcvd: 127

So what if you wanted to know what name the IP address
resolves to, when using dns

$ dig @ -x
210.0/ 3600 IN PTR

Above you can see it resolved to ""

Reference ( )
Also see TIP 223.

TIP 112:

Using GNU Autotools - so you can produce the familiar "./configure" "make" and "make install"
commands. There is also a "make dist".

The program and the rest of this code can be found at

A "" is required:

bin_PROGRAMS = sprog
sprog_SOURCES =

In addition, a "" file is required. Note, AC_CHECK_LIB will
check the "" file for the "sqlite3_open" file. Note that
"sqlite3", is a shortcut for "libsqlite3" by convention. If this file
is not found, AC_CHECK_FILE looks for "/usr/local/lib/libsqlite3.a". If
this is found, then, "-lsqlite3" is added to the LIBS environment variable.
Also, "-I/usr/local/include" and "-L/usr/local/lib" will be added on the
command line. This is common when some one does not have the library in
the path. (See TIP 49)

dnl Process this file with autoconf to produce a configure script.
AM_INIT_AUTOMAKE(sqliteprog, 1.0)
CXXFLAGS='-Wall -W -O2 -s -pipe'
if test "$found" = "no"; then
AC_CHECK_FILE(/usr/local/lib/libsqlite3.a, found=yes)
if test "$found" = "yes"; then
LIBS="$LIBS -lsqlite3"
INCLUDES="$INCLUDES -I/usr/local/include"
echo "Are you SURE sqlite3 is installed?"

To build the configure file, just run the following:

$ aclocal
$ autoconf
$ automake --add-missing

Now if you want to make a tar.gz file "sqliteprog-1.0.tar.gz", then
all you have to run is the following:

$ make dist

Note: did you ever want to save all the output from a ./configure? Well, it
is automatically saved in the "config.log" file. In fact, this file may
contain a lot more than what you saw on the screen.

Also, you may need to rerun ./configure. But before you do, delete
the "config.cache" file to get a clean build.

TIP 113:

EMACS - common emacs commands.

M is the ESC
C or c is the Ctl

Shell - when working in a shell. "M-x rename-uniquely" is good for split screen editing.

M-x rename-uniquely Use this for multiple shells (renames buffer so it's not the same shell)
C-c C-z Send job in background (when working in a shell)
C-c C-o commit-kill-output (gets rid of a lot of shell output)
C-c C-r reposition at beginning of output
C-c C-e reposition at end of output
M-x send-invisible Hide passwords - use this before typing a password

Note: if the shell prompt does not show up correctly, then, you may want to creat a ".emacs_bash"
file with the following contents:

PS1="emacs:\W \$ "

Directories (C-x d) give you a directory listing. You know all those annoying "~" and "#"
file that you get? You can easily delete these when in "dired" mode by hitting
"~", then "d" to flag it for delete. Then, hit "x" to and confirm deletion.

These are other command that work on highlighted files in "dired" mode.

R rename
v view
Z compress the file
+ create directory

Other common commands:

c-x l list the line you are on, and how many lines in the document.
You will get something like: Page has 4881 lines (4440 + 442),
which means you are on the 4440 line.

c-x rm bookmark make
c-x rb bookmark bounce

c-x rb notes
c-x rb emacs

c-x / (save position in register )
c-x j (jump to position in register )
c-x r SPC 1 (mark current point in register 1)
c-x r j 1 (jump to marked point in register 1)
c-x r t (insert string into register)

c-x r s 1 (save marked region in register 1)
c-x r i 1 (insert marked region)

c-x c-o (delete all blank lines, except one)

c-x z (repeat the last command ... stop with an a)
c-x zz (repeat the last command twice)

goto the next region
then, C-x r r "name of register"

to insert the register
C-x r i "name of register"

c-x ( start macro
c-x ) end macro
c-x e execute macro

c-x m mail
c-c c-s send

C-x C-e
(insert "\n\nExtra Line of text")

;; chirico functions in .emacs
;; This creates an html template
(defun my-html ()
(insert "


Backspace issues when using "emacs -nw"? They putting the following in your "~/.emacs" file

(global-set-key "\C-d" 'backward-delete-char)
(global-set-key "\C-h" 'backward-delete-char)
(global-set-key (kbd "DEL") 'delete-char)

TIP 114:

ncftpget - an intelligent ftp client ( Also
check your fedora or debian install. This package allows
you to easily download packages from ftp sites.

This is an example of connect to an ftp site, with a subdirectory, and
downloading all in one command.

$ ncftpget

Of if you want to get the fedora core 3 installs

$ ncftpget*

TIP 115:

expr - evaluate expressions. You can use this on the command line

$ expr 6 + 4

Note the spaces. Without spaces, you get the following:

$ expr 6+4

If you're using "*", you'll need a "\" before it

$ expr 10 \* 10

This also works for variables

$ var1=34
$ expr $var1 + 3


$ var1=2
$ var1=`expr $var1 \* 2`
$ echo $var1

see (TIP 25) you can get the cosine(.23)

$ var1=`echo "c(.23)"|bc -l`
$ echo $var1

You can also do substrings:

$ expr substr "BigBear" 4 4

And length of strings

$ mstr="12345"
$ expr length $mstr

Regular expressions

$ expr "a3" : [a-z][1-9]

Or you can get a bit fancy

$ myexpr="[a-z][1-9]"
$ echo $myexpr

$ expr "a3" : $myexpr

This may not be the best way to find out if it is Friday, but
it seems to work. It's more of an exercise in xargs.

$ date
Fri Dec 31 16:44:47 EST 2004
$ date|xargs -i expr {} : "[Fri]"

TIP 116:


$ mypipe="|"
$ eval ls $mypipe wc
6 6 129

Did you catch that? The above statement is the same as

$ ls | wc

Where "|" is put into the variable $mypipe

(also see TIP 118)

TIP 117:

lxr, glimpse, patchset - tools for reading the kernel source

This example puts some of the files in /home/src since my home
partition is the largest. Plus, you do not want to over write
the source in /usr/src/ If you want to put your files elsewhere
just substitute /home/src for your desired directory.

patchset -- download and setup

$ export SRCDIR=/home/src
$ cd $SRCDIR
$ wget
$ export PATH=$PATH:$SRCDIR/patchset-0.5/bin

Now edit "/home/src/patchset-0.5/etc/patchset.conf" and set WWW_USER to
whatever your website runs as

export WWW_USER=nobody

Getting kernel source. The last step builds and asks a lot of questions. Enter
yes to things that interest you, since this is what you will see in the source
code. It is not going to build for booting. The "downlaod -p" is for downloading
a patch.

$ download 2.6.10
$ createset 2.6.10
$ make-kernel -b 2.6.10

glimpse -- download and setup

$ mkdir -p /home/src/glimpse
$ cd /home/src/glimpse
$ wget
$ tar -xzf glimpse-latest.tar.gz
$ cd glimpse-4.18.0
$ ./configure; make
$ make install

lxr -- download and setup

$ make -p /home/src/lxr
$ cd /home/src/lxr
$ wget
$ cd lxr-0.3

Edit "Makefile" and set PERLBIN to "/usr/bin/perl" or the where perl is
on your system. Also set INSTALLPREFIX to "/var/www/lxr". Then, as root
do the following:

$ make install

Apache changes

Next edit the apache httpd.conf. On my system it is
"/usr/local/apache2/conf/httpd.conf", but if you did a fedora install
I think this file is located at "/etc/httpd/conf/httpd.conf".

Alias /lxr/ "/var/www/lxr/"

Options ExecCGI Indexes Includes FollowSymLinks MultiViews
AllowOverride all
Order allow,deny
Allow from all

SetHandler cgi-script

lxr - continued "/var/www/lxr/http/lxr.conf" changes. The following contains
my lxr.conf with changes made to almost every variable. Make sure you use
your website in place of

# Configuration file.

# Define typed variable "v", read valueset from file.
variable: v, Version, [/var/www/lxr/source/versions], [/var/www/lxr/source/defversion]

# Define typed variable "a". First value is default.
variable: a, Architecture, (i386, alpha, m68k, mips, ppc, sparc, sparc64)

# Define the base url for the LXR files.

# These are the templates for the HTML heading, directory listing and
# footer, respectively.
htmlhead: /var/www/lxr/http/template-head
htmltail: /var/www/lxr/http/template-tail
htmldir: /var/www/lxr/http/template-dir

# The source is here.
sourceroot: /var/www/lxr/source/$v/
srcrootname: Linux

# "#include " is mapped to this directory (in the LXR source
# tree)
incprefix: /include

# The database files go here.
dbdir: /var/www/lxr/source/$v/

# Glimpse can be found here.
glimpsebin: /usr/local/bin/glimpse

# The power of regexps. This is pretty Linux-specific, but quite
# useful. Tinker with it and see what it does. (How's that for
# documentation?)
map: /include/asm[^\/]*/ /include/asm-$a/
map: /arch/[^\/]+/ /arch/$a/

Now you should be ready to run "make-lxr". Make sure the path is setup to patchset,
which is repeated here. The last step take awhile.

$ export SRCDIR=/home/src
$ cd $SRCDIR
$ export PATH=$PATH:$SRCDIR/patchset-0.5/bin

$ make-lxr 2.6.10

Now you need to index the source. Below the ./glimpse_* file will be put in
root. Checkout the -H option if you do not want them here on a temporary
bases of if you run out of room.

$ glimpseindex -o -t -w 5000 /var/www/lxr/source/2.6.10 >& .glimpse_out

Since the above put the files under /root/.glimpse_* they should be moved

$ mv /root/.glimps_* /var/www/lxr/source/2.6.10/.
$ chown -R nobody.nobody ./.glimpse_*

TIP 118:

exec - you can change standard output and input without starting a new

The exec redirect the output from ls and date to a file. Nothing
is show on the terminal until "exec > /dev/tty" is performed

$ exec > mfile
$ ls
$ date
$ exec > /dev/tty

This is an example of assigning file descriptor 3 to file "output3" for
output, then, redirecting "ls" to this descriptor. Finally, file descriptor
3 is used for input, and the contents are read into the cat command.

$ exec 3>output3
$ ls >& 3
$ exec 3 $ cat <&3

Could you redirect the output to 3 files and stderr?

$ exec 3>output3
$ exec 4>output4
$ exec 5>output5

$ ls >& 3 >& 4 >& 5 >& 2 // Nope, can't do this.
output3 output4 output5

Instead, you should do the following:

$ ls | tee output3 | tee output4 |tee output5

Closing the "output" file descriptor

$ >&3-

Closing the "input" file descriptor

$ 3<&-

See what is still open on 0-10

$ lsof -a -p $$ -d 0-10

Recursion - the following counts to 5, then, quits.

sleep 1
declare -x n
let n=${n:=0}+1
[ $n -le 5 ] && echo "$n" && exec $0

There are some real-life applications for this technique, as follows:

declare -x N
declare -x n
N=${N:=$(od -vAn -N1 -tu4 < /dev/urandom)}
let n=${n:=0}+1
[ $(($n%2)) -eq 0 ] && echo "She Loves Me!" || echo "She Loves Me NOT!"
[ $n -lt $N ] && exec $0

TIP 119:

runlevel - need to know the current runlevel?

$ who -r
run-level 3 Dec 31 19:02 last=S

Need to know the architecture?

$ arch

TIP 120:

at - executes commands at a specified time.

A few examples here. The 1970 program will run
next Auguest 2 even though the year 1970 has long past.

$ at 6:30am Jan 12 < program
$ at noon tomorrow < program
$ at 1970 pm August 2 < program

This is an interactive way to use the command:

$ at now + 6 minutes
warning: commands will be executed using (in order) a) $SHELL b) login shell c) /bin/sh
at> ls
at> date > /tmp/5min
at> ^D
job 3 at 2005-01-01 08:50

What jobs are in the queue?

$ atq


$ at -l

TIP 121:

Creating a Manpage

As root you can copy the following to /usr/local/man/man1/soup.1 which will
give you a manpage for soup.

.\" Manpage for souptonuts.
.\" Contact to correct errors or omissions.
.TH man 1 "04 January 2005" "1.0" "souptonuts man page"
soup \- man page for souptonuts
souptonuts is a collection of linux and open
source tips.
off for golf.
The souptonuts does not take any options.
doughnut(1), golf(8)
No known bugs at this time.
Mike Chirico (

So, to view this man page

$ man soup

It's also possible to compress

$ gzip /usr/local/man/man1/soup.1

For plenty of examples look at the other man pages. Also the following
is helpful. The last one is a tutorial "man 7 mdoc"

$ man manpath
$ man groff
$ man 7 mdoc

TIP 122:

dmesg - print out boot messages, or what is in the kernel ring buffer.

If you missed the messages on boot-up, you can use dmesg to print them.

$ dmesg > boot.msg

Or to print, then, clear the ring

# dmesg -c > boot.msg

(also see TIP 20)

TIP 123:

gnus - emacs email nntp news reader (comcast as example with NO TLS or SSL)

First check that you can connect to the news group:

$ telnet 119
Connected to
Escape character is '^]'.
200 News.GigaNews.Com

If you want to check for TLS or SSL see (TIP 54).

Here is a very simple configuration example without encryption. It
appears that comcast does not support ssl or TLS.

In the "~/.emacs" file you would add the following to get comcast
news groups

(setq gnus-select-method '(nntp ""))

Then, create an "~/.authinfo" file with the following settings using
you own username and password.

machine login password borkeypass0rd

Next create a "~/.newsrc" with your groups

comp.lang.c++.moderated! 1-500
comp.unix.programmer! 1-500! 1-500
gnu.emacs.gnus! 1-500

Finally, create a "~/.gnus" with the following email settings for you

(setq user-mail-address "")

(defun my-message-mode-setup ()
(setq fill-column 72)
(add-hook 'message-mode-hook 'my-message-mode-setup)

To get into gnus

E-x gnus

The following are common gnus commands

RET view the article under the cursor

A A (shift-a, shift a): List all newsgroups known
to the server.

l (lower-case L) : List only subscribed groups
with unread articles.

L : List all newsgroups in .newsrc file.

g : See if new articles have arrived.

Some commands for reading

n next unread article

p previous article

SPC scroll down moves to next unread
when at the bottom of the article

del scroll up

F follow-up to group on the article you are
reading now.

f follow-up to group without citing the article

R reply by mail and cite the article

r reply by mail without citing the article

m new mail

a new posting

c Catchup

C-u / t Show only young headers
/ t without C-u limits the summary
to old headers

T T toggle threading

C-u g Display raw article
hit g to return to normal view

t Show all headers it's a toggle

W w Wordwrap the current article

W r Decode ROT13 a toggle

^ fetch parent of article

L create a scorefile-entry based
on the current article (low score)
? gives you information what each char means

I like L but high score

Commands to send email

C-c C-c send message

C-c C-d save message as draft

C-c C-k kill message

C-c C-m f attach file

M-q reformat paragraph

TIP 124:

Sending Email from telnet

Note, if you are on the computer you can sometime use the local loopback.
In fact, sometimes you can only use the local loop back in
place of ""

1 [mchirico@soup Notes]$ telnet 25
2 Trying
3 Connected to
4 Escape character is '^]'.
5 220 ESMTP Postfix (Postfix-20010228-pl03) (Mandrake Linux)
7 HELO // server echo
8 250
10 MAIL FROM: // server echo
11 250 Ok
13 RCPT TO: // server echo
14 250 Ok
16 DATA // echo
17 354 Enter mail, end with "." on a line by itself
18 This is a test message
19 This is a test message
20 to send
21 to send
22 .
23 250 2.0.0 j0B0uH3L018469 Message accepted for delivery

Above on line 6 you can type in any domain name. Line 7 is an echo. All
echos are listed in the comment field.

TIP 125:

IP forwarding, IP Masquerade

# echo 1 > /proc/sys/net/ipv4/ip_forward
# ipchains -F forward
# ipchains -P forward DENY
# ipchains -A forward -s -j MASQ
# ipchains -A forward -i eth1 -j MASQ

This assumes that your internal network is on eth1, and the
internet is connected to eth0.

(Also See TIP 182)

TIP 126:

Setting KDE as the default desktop manager

Edit "/etc/sysconfig/desktop" to include the two lines:


TIP 127:

Have a file and you do not know whay type it is (tar, gz, ASCII, binary) ?
Use the file command. Below it is used on the file "mftp"

$ file mftp
mftp: Bourne-Again shell script text executable

TIP 128:

Software RAID: Two good references

Note, you must setup grub for each RAID 1 device. Suppose you have
2 SCSI drives (sda and sdb). By default grub is setup on sda; but, you
need to enable it for sdb (/dev/hdb for ide) as follows:

grub>device (hd0) /dev/sdb
grub>root (hd0,0)
grub>setup (hd0)

Checking if "/boot/grub/stage1" exists... no
Checking if "/grub/stage1" exists... yes
Checking if "/grub/stage2" exists... yes
Checking if "/grub/e2fs_stage1_5" exists.. yes
Running "embed /grub/e2fs_stage1_5 (hd0)"... 16 sectors are embedded.
Running "install /grub/stage1 (hd0) (hd0)1+16 p (hd0,0)/grub/stage2 /grub/grub
.conf"... succeeded.


Checking to see if everything is working:

$ cat /proc/mdstat

Checking the drives

$ sfdisk -d /dev/sdb
$ sfdisk -d /dev/sda

$ fdisk -l /dev/sda "This will give general information"
$ fdisk -l "General information for all drives"

Adding raid (assume you want to add the first drive "sda1", or if it is the second
drive then substitute "sda2" below )

$ raidhotadd /dev/md0 /dev/sda1
$ raidhotadd /dev/md1 /dev/sda2
$ raidhotadd /dev/md2 /dev/sda3

This is an example of an cat /proc/mdstat that is working. Note that
there is a listing for both sda1[0] and sdb1[1]

$ cat /proc/mdstat

Personalities : [raid1]
read_ahead 1024 sectors
Event: 12
md0 : active raid1 sda1[0] sdb1[1]
104320 blocks [2/2] [UU]

md1 : active raid1 sda2[0] sdb2[1]
1044160 blocks [2/2] [UU]

md2 : active raid1 sda3[0] sdb3[1]
34411136 blocks [2/2] [UU]

unused devices:

Compare that to this where md2 is missing sdb3

$ cat /proc/mdstat

Personalities : [raid1]
read_ahead 1024 sectors
Event: 9
md0 : active raid1 sda1[0] sdb1[1]
104320 blocks [2/2] [UU]

md1 : active raid1 sda2[0] sdb2[1]
1044160 blocks [2/2] [UU]

md2 : active raid1 sdb3[1] <---- HERE
34411136 blocks [2/1] [_U]

unused devices:

If you are rebuilding an array, you can watch it by doing the following:

$ watch -n1 cat /proc/mdstat

Need to know the raid setup?

$ cat /etc/raidtab

TIP 129:

Resetting Redhat Linux Passwords using GRUB

1. Press 'e'
2. Press 'e' again
3. Append 'single' to the kernel version listing


TIP 130:

mtr - matt's traceroute. This is an advanced traceroute that keeps
$ mtr

Matt's traceroute [v0.52]
third-fl-71.localdomain Thu Jan 20 11:05:57 2005
Keys: D - Display mode R - Restart statistics Q - Quit
Packets Pings
Hostname %Loss Rcv Snt Last Best Avg Worst
1. 0% 3 3 0 0 0 1
2. ???
3. 0% 3 3 8 7 7 8
4. 0% 2 2 8 8 8 8
5. 0% 2 2 8 8 8 8
6. 0% 2 2 12 12 12 13
7. 0% 2 2 12 12 13 13
8. 0% 2 2 13 13 13 13
9. 0% 2 2 12 12 13 14
10. so-1-0-0.gar4.NewYork1.Level3.n 0% 2 2 14 14 37 61
11. 0% 2 2 13 12 13 13
12. ge-0-3-0.bbr2.Washington1.Level 0% 2 2 19 19 19 19
13. ge-1-1-51.car1.Washington1.Leve 0% 2 2 18 18 19 20
14. 0% 2 2 21 19 20 21
15. 0% 2 2 21 20 20 21
16. 0% 2 2 23 21 22 23

TIP 131:

chfn - change finger information

$ chfn

Next you are asked for a password and user information.

TIP 132:

chsh - change login shell

First, you may want to get a listing of all the possible

$ chsh -l


TIP 133:

bash - working with binary, hex and base 3.

For the variable must be declare as an integer. Then
specify the #. The example below is 22 in
base 3.

$ declare -i n
$ n=3#22
$ echo $n

Base 16 (hex)

$ declare -i n2
$ n2=16#a
$ echo $n2

Base 8 (octal)

$ declare -i n3
$ n3=8#11
$ echo $n3
9 Note 8+1=9

TIP 134:

monitoring IP traffic. Try iptraf

TIP 135:

enscript - convert text files to PostScript

TIP 136:

dd and tar - blocking factor. How to determine the blocking factor, block size
so that tar and dd can work together.

Step 1: Create a large file on local disk, in a directory "1" that will eventually
be written to tape. This will be created with dd as follows:

$ mkdir 1
$ dd if=/dev/zero of=disk-image count=40960
40960+0 records in
40960+0 records out

$ cd ..

Step 2: tar the directory and contents to tape. First rewind the tape. These examples
use /dev/nst0 as the location of the tape. Make sure to substitute your values
if needed.

$ mt -f /dev/nst0 rewind
$ tar --label="Test 1" --create --blocking-factor=128 --file=/dev/nst0 1

Step 3: Read data from the tape using a block size of 128k. If you get an I/O error, which
could happend if you used a different blocking factor above, then, you may need
to increase the bs to 256, or 512 etc. as needed.

$ mt -f /dev/nst0 rewind
$ dd if=/dev/nst0 bs=128k of=testblocksz count=1
0+1 records in
0+1 records out

$ ls -l testblocksz
-rw-r--r-- 1 root root 65536 Feb 9 10:41 testblocksz

$ ls -lh testblocksz
-rw-r--r-- 1 root root 64k Feb 9 10:41 testblocksz

Note above that the size 65536 is equal to 64k. That "h" switch in "ls" is for
human readable.

Step 4: tar uses a multiplier of 512*blocking-factor to get block size. Again

512 * blocking-factor = block size used in dd command.

Putting in the values, we see that

512 * 128 = 65536

Step 5: So what does this tell you? You can now use these numbers to "dd" files
to tape. But, first tar will be used to create the file locally.

$ tar --label="Test 1" --create --blocking-factor=128 --file=test.tar 1

Step 6: Send this to tape with the dd command. Remember 64k is equal to 65536.

$ mt -f /dev/nst0 rewind
$ dd if=test.tar bs=64k of=/dev/nst0

Step 7: Now test that it can be read with tar command using blocking-factor=128.
Note the "t" command in tar is for tell. It will not write data.

$ mt -f /dev/nst0 rewind
$ tar -tvf /dev/nst0 --blocking-factor=128
V--------- 0/0 0 2005-02-09 10:38:20 Test 1--Volume Header--
drwxr-xr-x root/root 0 2005-02-09 10:34:10 1/
-rw-r--r-- root/root 20971520 2005-02-09 10:34:11 1/disk-image

Step 8: Reading tape data with dd. Most of the time a high "ibs" input block size

$ mt -f /dev/nst0 rewind
$ dd if=/dev/nst0 of=outfromdd.tar ibs=64k
321+0 records in
41088+0 records out

Step 9: Verify that outfromdd.tar can be read by tar with blocking-factor=128

$ tar -tvf outfromdd.tar --blocking-factor=128
V--------- 0/0 0 2005-02-09 10:38:20 Test 1--Volume Header--
drwxr-xr-x root/root 0 2005-02-09 10:34:10 1/
-rw-r--r-- root/root 20971520 2005-02-09 10:34:11 1/disk-image

PULLING FILES: The dd command can be used to pull files.

ssh target_address dd if=remotefile | dd of=localfile

Or, a specific example of getting a file from a computer called hamlet.

$ ssh root@hamlet dd if=/home/cvs/test | dd of=/home/storage/test


Go to end of data
$ mt -f /dev/nst0 eod

Previous record
$ mt -f /dev/nst0 bsfm 1

Forward record
$ mt -f /dev/nst0 fsf 1

$ mt -f /dev/nst0 rewind

$ mt -f /dev/nst0 tell

(Reference TIP 151 - for how to get around firewalls)

Below is a script that I use to backup computers via ssh. The
tape drive is on "nis" and the extra space is on "hamlet".

# Program to backup server remotely
# Assume remote server is nis, you are on squeezel
# Recover from tape
# dd if=/dev/nst0 of=test.tar.gz bs=64k
filename="support1.$(date "+%m%d%y%H%M").tar.gz"
#tar cvzf - $DIRTOBACKUP | ssh root@nis '(mt -f /dev/nst0 rewind; dd of=/dev/nst0 bs=64k )'
tar cvzf - $DIRTOBACKUP | ssh support1@hamlet "dd of=/home/support1/backups/${filename} "

Another example program, below, pushes the last ".tar.gz" file to tape:

# Program to push files to tape
# Notes on recovering from tape
# dd if=/dev/nst0 of=test.tar.gz ibs=64k
# or
# $ ssh root@tapeserver "mt -f /dev/nst0 rewind"
# $ ssh root@tapeserver "dd if=/dev/nst0 ibs=64k"|dd of=cvs1.tar.gz
# First rewind tape
ssh root@tapeserver 'mt -f /dev/nst0 rewind'
# Grab only the last file
file=$(find /home/cvs -iname 'cvs*.tar.gz'|sort|tail -n 1)
dd if=${file}|ssh root@tapeserver 'dd of=/dev/nst0 bs=64k'

TIP 137:

Apache - redirecting pages. All changes are in httpd.conf

RedirectMatch (.*)\.gif$$1.jpg

Redirect /service

If more than one DNS record points to the server, then, it's
possible to redirect based upon which DNS entry was used in
the web query.

For example, a single web server has the following three
DNS entries mapped to its single IP address.

It's possible to redirect or rewrite the page delivered to
the client with the following changes in httpd.conf

RewriteCond %{HTTP_HOST} ^$
RewriteRule ^/$ [L]

RewriteCond %{HTTP_HOST} ^$
RewriteRule ^/$ [L]

TIP 138:

samba mounts via ssh - mounting a samba share through an ssh tunnel, going
through an intermediate computer, that accepts ssh. We'll call this
intermediate computer middle [], and we want to get to
destination []. The user will be mchirico.


$ mkdir -p /samba/share


This has to be done as root, since we are using a lower port.

$ ssh -N -L 139: mchirico@


umount /samba/sales
/bin/mount -t smbfs -o username=donkey,workgroup=donkeydomain,
netbiosname=homecpu //localhost/share /samba/share

TIP 139:

Music on Fedora Core -- How to play music on with "xmms".

The following command will show the sound driver:

$ lspci|grep -i audio


Unmute amixer with the following command:

$ amixer set Master 100% unmute
$ amixer set PCM 100% unmute

Note you can also get a graphical interface with "alsamixer"

$ alsamixer

h,F1 -- for help
Esc -- exit
Tab -- move to selections


Test a sound file "*.au" with aplay. To quickly find files on your system use
the "locate *.au" command.

$ aplay /usr/lib/python2.3/test/


Install "xmms-mp3-1.2.10-9.2.1.fc3.rf.i386.rpm" which does not come with Fedora because
of GPL license restrictions. The latest version of this package can be found
at the following url:

$ rpm -ivh xmms-mp3-1.2.10-9.2.1.fc3.rf.i386.rpm


Go to magnatun "", select genre and make sure xmms
is the default player.

TIP 140:

Routing -- getting access to a network 1 hop away. You are currently on the 192 network
and you want access to the network that has a computer straddling
the two, with /proc/sys/net/ipv4/ip_forward set to 1.

$ route add -net netmask gw

To undo:

$ route del -net netmask gw

Now you can ping

Does not work?

Go on to and execute the following commands:

$ echo 1 > /proc/sys/net/ipv4/ip_forward
$ cat /proc/sys/net/ipv4/ip_forward

To Look at the the gateway, execute the following command.

$ netstat -r


TIP 141:

RAM disk -- creating a filesystem in RAM.

$ mkfs -t ext3 -q /dev/ram1 4096
$ mkdir -p /fsram
$ mount /dev/ram1 /fsram -o defaults,rw

TIP 142:

Create a Live Linux CDROM using BusyBox and OpenSSH.

These steps are rather long. A complete tutorial is given at
the following link:

TIP 143:

SystemImager ( SystemImager is software that automates Linux installs,
software distribution, and production deployment.

TIP 144:

Mounted a filesystem in rescue mode, yet, you cannot read and write? Remount.

$ mount -o remount /

TIP 145:

Nmap commands to check for Microsoft VPN connection.

$ nmap -sO -p 47
$ nmap -sS -p T:1723

By the way, with nmap you can specify multiple ports. Below
is an example of multiple ports; but, use the commands above
for Microsoft VPN services.

$ nmap -sS -p T:1723-3000

TIP 146:

Perl and ssh - monitoring systems. The output from ssh can be parsed. Below is
a simple procedure to just to read the ssh ouput into perl.

$pid = open $readme, "ssh root\@hamlet df -lh|" or die "Could not ssh\n";
while(<$readme>) {
print $_
close $readme

But note, you probably want to do something more complex. Below is a more robust
example that bypassed all the fortune, heading junk that you may encounter when
logging in.

$pid = open $readme, "ssh root\@hamlet df -lh 2>/dev/null|" or die "Could not ssh\n";
while(<$readme>) {
print $_
close $readme

NO! you CANNOT do bidirectional communication with the open statement. Note the "|" before
and after below, which cannot be done.

# Cannot do this!
$pid = open $readme, "|ssh root\@hamlet df -lh 2>/dev/null|" or die "Could not ssh\n";

Below is a simple Perl example working with arrays:

@ArrayOfArray = (
[ "ant", "bee" ],
[ "mouse", "mole", "rat" ],
[ "duck", "goose", "flamingo" ],
[ "rose","carnation","sunflower"],

for $i ( 0 .. $#ArrayOfArray ) {
for $j ( 0 .. $#{$ArrayOfArray[$i]} ) {
print "Element $i $j is $ArrayOfArray[$i][$j]\n";

# Or this is another way to list elements
foreach( @ArrayOfArray ) {
foreach $i (0..$#$_) {
print "$_->[$i] "
print "\n";

Below is an example of working with Hash of Arrays:

# ./program < /etc/passwd
next unless s/^(.*?):\s*//;
$HoA{$1} = [ split(/:/) ];
for $i (keys %HoA ) {
print "$i: @{ $HoA{$i} } \n";

Example of regular expression. This is my most used regular expression - I like
this sample. See the "" link at the end of this tip.

"hot cross buns" =~ /cross/;
print "Matched: <$`> $& <$'>\n"; # Matched: cross <>
print "Left: <$`>\n"; # Left:
print "Match: <$&>\n"; # Match:
print "Right: <$'>\n"; # Right: <>

If you're looking for Perl information, type "man perl", which will show you how
to get even more information. Or better yet, take a look at the following

For a quick example on using Perl with SQLite, see the following links:

Standard input for files. This example will read from stdin, or open a file if given as
an argument, and convert all "<" to "<" and ">" to ">", which can be handy when
converting text files to html files. Note the "while(<>)" will take multiple file names
on the command line.

while(<>) {
s/ s/>/>/g;

Perl Debugger is very useful for testing commands and works like an interpreter, just
like python. So to get into the Perl Debugger execute the command below, "q" to quit.

$ perl -de 0

Reference TIP 170

TIP 147:


# shutdown 8:00 -- Shutdown at 8:00

# shutdown +13 -- Shutdown after 13min

# shutdown -r now -- Shutdown now and restart

# shutdown -k +2 -- "The system is going DOWN to maintenance mode in 2 minutes!"
The above is only a warning.

# shutdown -h now -- Shutdown now and halt

# shutdown -c -- Cancel shutdown

TIP 148:

ac - print statistics about users connect time

$ ac -p -- print hour usage by user (individual)
$ ac -dy -- print daily usage

Options can also be combined

$ ac -dyp

TIP 149:

Smart Monitoring Tools:
Disk failing? Or want to know the temperature of your hard-drive?

For a good, quick tutorial, see the Linux Journal article

Below are some common commands:

$ smartctl -i /dev/hda

$ smartctl -Hc /dev/hda

$ smartctl -A /dev/hda

TIP 150:

Monitor dhcp trafic - dhcpdump and tcpdump.

Download dhcpdump

$ wget
$ ./configure
$ make && make install

Once it's installed, you can monitor all dhcp traffic as follows, if done with root.

$ tcpdump -lenx -i eth0 -s 1500 port bootps or port bootpc| dhcpdump

The above assumes you are using eth0 (ethernet port 0).

TIP 151:

Breaking Firewalls with ssh

A sample .ssh/config file (note this must have chmod 600 rights)

## Server1 ##
LocalForward 20000
LocalForward 22000

With the above "~/.ssh/config" file, after sshing into it
is then possible to ssh into nearby computers directly.

$ ssh -l mchirico
$ scp -P 22000 authorized_keys* mchirico@localhost:.
$ ssh -l mchirico localhost -p 22000

For the complete article reference the following link:

TIP 152:

Renaming files - suppose you want to rename all the ".htm" files to ".html"

$ rename .htm .html *.htm

Or, suppose you files file1, file2, file3 ...

$ touch file1 file2 file3 file4 file5 file6
$ rename file file. file*

The above command will give you "file.1", "file.2" ... "file.6"

TIP 153:

Renaming files with Perl - this is taken from "Programming Perl 3rd Edition"

# rename - change filenames
$op = shift;
for (@ARGV) {
$was = $_;
eval $op;
die if $@;
# next line calls built-in function, not the script
rename($was,$_) unless $was eq $_;

The above Perl program can be used as follows:

$ rename 's/\.orig$//' *.orig
$ rename 'y/A-Z/a-z/ unless /^Make/' *

Also reference:

TIP 154:

R project (

To start R, just type "R" at the command prompt and "q()" to quit. Below
2 is raised to powers 0 through 6 and thrown into an array.

$ R
> N <- 2^(0:6)
> N
[1] 1 2 4 8 16 32 64

There is a summary summary() command.

> summary(N)
Min. 1st Qu. Median Mean 3rd Qu. Max.
1.00 3.00 8.00 18.14 24.00 64.00

Note that the array begins as 1 and not 0

> N[1:3]
[1] 1 2 4

TIP 155:

ls - listing files by size, with the biggest file listed last

$ ls --sort=size -lhr

The above command sorts files by size, listing the contents in
"h" human readable format in reverse order.

Note the options: --sort={none,time,size,extension}

TIP 156:

Perl - program to clean up old versions of files

# Copyright (c) GPL 2005 Mike Chirico
# This program deletes old files from several directories
# and within each directory there must be x number of copies
# each y number of bytes

sub delete_old_ones {
# Don't change setting here of '-lt'
$pid = open $readme, "ls -lt $directory_and_file|" or die "Could not execute\n";
while(<$readme>) {
my @fields = split;
# Make sure we have $save_count good ones with data
if ($fields[4] > $bytes_in_file && $save_count > 0) {
print "Kept files: $fields[4] $fields[8]\n";
# delete the old ones
if ($save_count <= 0 )
print "Deleted files: $fields[4] $fields[8]\n";
unlink $fields[8];
close $readme;

@AofA = (
[ "/home/cvs/backups/*.gz", "6",196621 ],
[ "/home/mail/backups/*.gz","5",34 ],
[ "/home/snort/backups/*.gz","2",34 ],
[ "/home/server1/backups/*.gz","2",34 ],
[ "/home/actserver/backups/*.gz","2",34 ],
[ "/home/server2/backups/*.gz","2",34 ],

foreach( @AofA ) {

Reference TIP 170 and the following link:

TIP 157:

Graphics and Visualization Software that runs on Linux

TIP 158:

Keeping files in sync going both ways. Unlike rsync, this is not a one way mirror

You will need ocaml installed first.

$ wget
$ tar -xzf ocaml-3.08.3.tar.gz
$ cd ocaml-3.08.3

$ ./configure
$ make world
$ make opt
$ make install

Next, get unison and put it in a different directory.

$ wget
$ tar -xzf unison-2.10.2.tar.gz
$ cd unison-2.10.2
$ make UISTYLE=text
$ su
# cp unison /usr/local/bin/.

Note, you have to copy the file manually.

See the following article []

TIP 159:

Dump ext2/ext3 filesystem information with "dumpe2fs". Perform the mount command
and query away.

$ dumpe2fs /dev/sda1

TIP 160:

sysreport - a script that generates an HTML report on the system configuration. It
gathers information about the hardware and is somewhat redhat specific. The utility
should be run as root.

$ /usr/sbin/sysreport

TIP 161:

Key Bindings Using bind. You can bind, say, ctl-t to a command.

Add the following to you "~/.inputrc" file, just as it is typed below with quotes.

"\C-t": ls -l

Next, run the command

$ bind -f .inputrc

Or, you can do everything on the command line; however, it won't be there the next time
you log in. Below is the way to do everything on the command line.

$ bind -x '"\C-t":ls -l'

To unbind use the "-r" option. Single quotes are not needed.

$ bind -r "\C-t"

Getting a list of all bindings can be done as follows, and not this can be redirected
to the ".inputrc" file for further editing.

$ bind -p > .inputrc

TIP 162:

awk - common awk commands.

Find device names "sd" or with major number 4 and device name "tty". Print the
record number NR, plus the major number and minor number.

$ awk '$2 == "sd"||$1 == 4 && $2 == "tty" { print NR,$1,$2}' /proc/devices

Find device name equal to "sound".

$ awk '/sound/{print NR,$1,$2}' /proc/devices

Print the 5th record, first field, in file test

$ awk 'NR==5{print $1}' test

Print a record, skip 4 records, print a record etc from file1

$ awk '(NR-1) % 4 == 0 {print $1}' file1

Print all records except the last one from file1

$ tac file1|awk 'NR > 1 {print $0}'|tac

Print A,B,C ..Z on each line, cycling back to A if greater than 26 lines

$ awk '{ print substr("ABCDEFGHIJKLMNOPQRSTUVWXYZ",(NR-1)%26+1,1),$0}' file1

Number of bytes in a directory.

$ ls -l|awk 'BEGIN{ c=0}{ c+=$5} END{ print c}'

Remove duplicate, nonconsecutive line. As an advantage over "sort|uniq"
you can eliminate duplicate lines in an unsorted file.

$ awk '! a[$0]++' file1

Or the more efficient script

$ awk '!($0 in a) {a[$0];print}' file1

Print only the lines in file1 that have 80 characters or more

$ awk 'length < 80' file1

Print line number 25 on an extremely large file -- note it has
to be efficient and exit after printing line number 25.

$ awk 'NR==25 {print; exit}' verybigfile

TIP 163:

Configuring Remote Logging. If you have several servers on, you can setup remote logging
as follows.


Firewall - allow UDP port 514 on the main server that will receive the logs.

$ iptables -A INPUT -p udp -s --dport 514 -j ACCEPT

Edit "/etc/sysconfig/syslog" and add the "-r" option to SYSLOGD_OPTIONS as shown below.


Note, the "-r" is to allow remote logging and "-m 0" specifies that that the syslog process should
not write regular timestamps. I prefer to only write timestamps for the clients.

Next, restart the logging process

$ service syslog restart


Edit "/etc/syslog.conf" and add the ip address of the log server, or put in the hostname.

*.* @

Next, restart the logging process

$ service syslog restart

TIP 164:

kudzu - hardware on your system. To probe the hardware on your system without doing
anything, issue the following command.

$ kudzu -p

But wait, a lot of this information is already recorded in the following file


You can also use lspci to list all PCI devices.

$ lspci

Also, take a look at the script /etc/sbin/sysreport, since this script has a lot of
info gathering commands. You can pick and choose what you want, or run the complete

If you just want information on the NIC

$ ip link show eth0
2: eth0: mtu 1500 qdisc pfifo_fast qlen 1000
link/ether 00:11:11:8a:be:3f brd ff:ff:ff:ff:ff:ff

TIP 165:

cfengine - a very power agent for monitoring and administrating both a single computer
and or multiple computers. [ ]

The following is a quick example on downloading and installing cfengine.

$ ncftpget
$ md5sum cfengine-2.1.15.tar.gz
f03de82709f84c3d6d916b6e557321f9 cfengine-2.1.15.tar.gz

$ tar -xzf cfengine-2.1.15.tar.gz

You need to have a current version of BerkeleyDB (
Note that BerkeleyDB has a funny install. You cd to the "build_unix" directory, then,

Installing BerkeleyDB if needed:
$ wget
$ tar -xzf db-4.3.28.tar.gz
$ cd db-4.3.28/build_unix/
$ ../dist/configure
make install

You also need a current version of OpenSSL. For instructions on how to install OpenSSL see

See (TIP 49) on putting "/usr/local/BerkeleyDB.4.3/lib" in the "/etc/" file. Or
once BerkeleyDB is installed, you can put the location on the command line as follows:

Configuring cfengine with direct reference to BerkeleyDB.4.3. First cd to the cfengine source.

$ ./configure --with-berkeleydb=/usr/local/BerkeleyDB.4.3/lib
$ make
$ make install

Next create the following directories:

$ mkdir -p /var/cfengine/bin
$ mkdir -p /var/cfengine/inputs

Copy needed files (cfagent, cfdoc, cfenvd, cfenvgraph, cfexecd, cfkey, cfrun, cfservd, cfshow):

$ cp /usr/local/sbin/cf* /var/cfengine/bin

You'll also need to generate keys. As root, execute the following:

$ cfkey

The command above will write the public and private keys in

You probably want (cfexecd, cfservd, and cfenvd) running on all servers. If you
add the following to "/etc/rc.local" these daemons will start on reboot.

# Lines in /etc/rc.local

Also, make sure you run each command now as follows:

$ /usr/local/sbin/cfexecd
$ /usr/local/sbin/cfservd
$ /usr/local/sbin/cfenvd

Firewall settings must be adjusted to allows 5308 for tcp/udp. My local network
is, so I'm opening it up for all my computers.

$ iptables -A INPUT -p udp -s --dport 5308 -j ACCEPT
$ iptables -A INPUT -p tcp -s --dport 5308 -j ACCEPT

A set of keys needs to be on the server and hosts. For example, my key on ""
should be copied over to the server "" as follows:

This is done from

$ scp /var/cfengine/ppkeys/
$ scp /var/cfengine/ppkeys/

Also, "/var/cfengine/inputs/cfrun.hosts" on the server "" must contain
all the computers that will get updated. This is "cfrun.hosts" on ""

Once I'm done, from "" I can run the following test:

$ cfrun -v

TIP 166:

cfengine - a quick example. This example will be run as root. You create the file "cfagent.conf" in
"/var/cfengine/inputs/". The example below will checksum all the files in /home/chirico/deleteme/tripwire,
it will also comment out the line "finger" in any file located in /tmp/testdir/stuff, also appending
the command in this file " Edit change with cfengine".

# /var/cfengine/inputs/cfagent.conf
# You run this with the following:
# cfagent -vK

actionsequence = ( files tidy editfiles )
ChecksumDatabase = ( /var/cfengine/cache.db )
# Below, true to update md5
ChecksumUpdates = ( true )

/home/chirico/deleteme/tripwire checksum=md5 recurse=inf
/home/chirico/deleteme/tripwire/moredata checksum=md5 recurse=inf
#/home/chirico/deleteme/tripwire/compress recurse=inf include=*.txt acti on=compress
# If the database isn't secure, nothing is secure...
/var/cfengine/cache.db mode=600 owner=root action=fixall

/home/chirico/deleteme/tripwire pattern=*~ recurse=inf age=0
# You must put an age. 0 runs now.


{ /tmp/testdir/stuff

HashCommentLinesContaining "finger"
AppendIfNoSuchLine "# Edit Change with cfengine "

A few further notes on the above. The command "actionsequence = ( files tidy editfiles) tells the order
of what to execute. The heading "tidy:" deletes files, and of course, "editfiles" does the editing of files.

To run the example, execute the following command. The "-K" causes the lock file to be ignored.

$ cfagent -vK

TIP 167:

Implementing Disk Quotas - a quick example that can easily be done on a live system for testing. There
is no need to reboot, since you'll be creating a virtual filesystem.

Do the following as root. First create a mount point.

# mkdir -p /quota

Next, create 20M file. Since I have many of these files, I created a special directory "/usr/disk-img"

# mkdir -p /usr/disk-img
# dd if=/dev/zero of=/usr/disk-img/disk-quota.ext3 count=40960

The dd command above create a 20 MB file because, by default, dd uses a block size of 512 bytes. That makes
the size: 40960*512=20971520.

Next, format this as an ext3 filesystem

# /sbin/mkfs -t ext3 -q /usr/disk-img/disk-quota.ext3 -F

Add the following line to "/etc/fstab"

/usr/disk-img/disk-quota.ext3 /quota ext3 rw,loop,usrquota,grpquota 0 0

Now, mount this filesystem

# mount /quota

Take a look at it:

# ls -l /quota

Now, run "quotacheck"

# quotacheck -vug /quota

You'll get errors the first time this is run, because you have no quota files.
But, run it a second time and you'll see something similiar to the following:

# quotacheck -vug /quota
quotacheck: Scanning /dev/loop2 [/quota] done
quotacheck: Checked 3 directories and 4 files

Now take a look at the files:

# ls -l /quota
total 26
-rw------- 1 root root 6144 Jun 14 12:23
-rw------- 1 root root 6144 Jun 14 12:23 aquota.user
drwx------ 2 root root 12288 Jun 14 12:18 lost+found

Next use "edquota" to grant the user "chirico" a certain quota

# edquota -f /quota chirico

This will bring up a menu, and here I have edited so that user "chirico"
has a soft limit of 120*512=61K, and a soft limit of 2 inodes and a hard limit of 5.

Disk quotas for user chirico (uid 500):
Filesystem blocks soft hard inodes soft hard
/dev/loop2 2 120 150 1 2 3

Next, turn quotas on with the following command:

$ quotaon /quota

If you need to turn off quotas, the command is "quotaoff -a" for all filesystems. You'll run into
errors if you try to run quotacheck, say "quotacheck -avug" because this tries to unmount and mount
the filesystem. You need to turn off quotas first "quotaoff /quota". Note you only need to run
quotacheck once, or when doing maintenance after a system crash.

To get a report on the quote, runn "repquota" as follows:

$ repquota /quota
*** Report for user quotas on device /dev/loop0
Block grace time: 7days; Inode grace time: 7days
Block limits File limits
User used soft hard grace used soft hard grace
root -- 1189 0 0 2 0 0
chirico -+ 93 0 0 4 2 5 6days

Note above that user "chirico" has used 4 on the file limits. This user has a hard
limit of 5. So when this user tries to create 2 more files (bring this over the limit of 5)
then he will get the following error as demonstrated below.

[chirico@squeezel chirico]$ touch one
[chirico@squeezel chirico]$ touch two
loop0: write failed, user file limit reached.
touch: cannot touch `two': Disk quota exceeded

Now, if repquota (run by root) is executed it shows the following:

$ repquota /quota
*** Report for user quotas on device /dev/loop0
Block grace time: 7days; Inode grace time: 7days
Block limits File limits
User used soft hard grace used soft hard grace
root -- 1189 0 0 2 0 0
chirico -+ 94 0 0 5 2 5 6days

Note the "+" sign above. User "chirico" is above the File soft limits, and in this case
above the hard limits.

To warn user by sending email to them, run "warnquota", but you need check that
"/etc/warnquota.conf" is setup correctly. For the example above, this file should
look as follows:

$ cat /etc/quotatab
# This is sample quotatab (/etc/quotatab)
# Here you can specify description of each device for user
# Comments begin with hash in the beginning of the line

# Example of description
/dev/loop0: This is loopback device

Just run the following as root:

$ warnquota

By the way, if you want to change the grace period, it can only be done on a filesystem
basis. Not per user.

$ edquota -t

Users can run "quota" to see their usage as follows:

[chirico@squeezel ~]$ quota
Disk quotas for user chirico (uid 500):
Filesystem blocks quota limit grace files quota limit grace
/dev/loop0 94 0 0 5 10 50

As you can see from above, I changed my inode limit to 50.

What about running this on the whole filesystem? Yes, below is an example where I'm running
this on FC3, on the root of the filesystem "/". This assumes that you have installed the
quota package. Try doing "rpm -q quota" to see if this package is installed.

Step 1:

Check to make sure the quota software is installed. You can either do a "whereis quota",
or check for the rpm package.

$ whereis quota
whereis quota
quota: /usr/bin/quota /usr/share/man/man1/quota.1.gz

Checking for the rpm package.

$ rpm -q quota

Step 2:

Edit /etc/fstab and add usrquota and grpquota options for "/dev/VolGroup00/LogVol00",
which is shown on the first line below:

/dev/VolGroup00/LogVol00 / ext3 defaults,usrquota,grpquota 1 1
LABEL=/boot /boot ext3 defaults 1 2
none /dev/pts devpts gid=5,mode=620 0 0
none /dev/shm tmpfs defaults 0 0
none /proc proc defaults 0 0
none /sys sysfs defaults 0 0
/dev/VolGroup00/LogVol01 swap swap defaults 0 0

Step 3:

Remount the filesystem as follows:

$ mount -o remount /

Step 4:

Run quotacheck with the "-m" option. Like the above statement, this will have to be run with
root priviliges. This creates the quota database files, and it can take a long time if it is
a large full filesystem.

$ quotacheck -cugm /

Step 5:

This step is optional, but it's good to know if you need to recalculate quotas because of a
system crash. It's demonstrated here, because at this point quota's have not been turned on.
Again, note the "m" option below.

$ quotacheck -avumg

Step 6:

Set limits for specific users or groups using the "edquota" command. Shown below is the command
to setup quotas for user "chirico". Shown below this user has used 161560 blocks, he has a soft
limit of 1161560 and a hard limit of 900000. He has used 3085 inodes and has a soft limit of 10000
and a hard limit of 12000.

$ edquota -f / chirico

Disk quotas for user chirico (uid 500):
Filesystem blocks soft hard inodes soft hard
/dev/mapper/VolGroup00-LogVol00 161560 1161560 900000 3085 10000 12000

You can put quotas on groups as well. The following is done as root. See (TIP 186 and TIP 6) for creating
groups and adding users to groups.

$ edquota -g share

If you create a sharable directory for anyone in the group "share" (TIP 6), quota restrictions against
group "share" will only apply to files added in the "/home/share" directory. When user "chirico" creates
files in "/home/share" they also go against this user quota as well. However, when files are created in
his home directory they do not go against the "share" group.

Note - if you get errors when trying to run "edquota -g share", turn quotas off "quotaoff /" and
run "quotacheck -avugm". Then, turn the quotas back on "quotaon /".

You can see the status of the group quota with the following command:

$ quota -g share

Step 7:

Turn on quotas with the "qutoaon" command. This command needs to be done with root privileges.

$ quotaon /

Step 8:

Check "/etc/quotatab" file for the correct entries. Note that when you do the "mount" command
the filesystem returned needs to match what is in the "quotatab" file. I have noticed that this
is not the case by default.

$ mount
/dev/mapper/VolGroup00-LogVol00 on / type ext3 (rw,usrquota,grpquota)

So the "/etc/quotatab" must contain the following line.

/dev/mapper/VolGroup00-LogVol00: This is the Volume group

Step 9:

Run "warnquota" as a check that the "/etc/quotatab" files is setup correctly.

$ warnquota

Step 10:

Setup a daily cron job for running "warnquota". The following should be placed
in "/etc/cron.daily"

# Place this file in /etc/cron.daily
# with rights 0755
if [ $EXITVALUE != 0 ]; then
/usr/bin/logger -t warnquota "ALERT exited abnormally with [$EXITVALUE]"
exit 0

(TIP 6, TIP 186, and TIP 205)

TIP 168:

rdist - remote file distribution client program. You can use this program in combination with
ssh. This program does more than just copy files. Once a file has been copied, you can dictate
other actions to be performed. Or you can hold off copying all together if the destination is
running low on inodes, or disk space.

For the purpose of this example, all commands will been run on "", and the
computers that will be updated are "" and "". Obviously, you
would substitute your computer names.

It helps to setup ssh keys on each computer first. Reference []
and (TIP 12).

Step 1: Create the Configuration file myDistfile

Below is my sample "myDistfile". This file will access hosts "" using username chirico
and "" with the username running this command, and copy the
files "/home/chirico/file1" and "/home/chirico/file2" to the these two servers creating the
directory ~/tmpdir if it doesn't exist. Once these files are updated, a mail check ("sendmail -bv")
will be performed, and mail will be sent to "chirico@squeezel". This happens twice, once for each file.

Note, the line "/home/chirico/file2 ->" which moves the file "file2" to renaming the file to "tapedest" in the directory "/home/chirico". Once this file
is copied, the rights are modified to "chmod +r". Likewise, "/home/chirico/file2 ->"
copies the file file2, which is renamed as closetdest.

# Contents of myDistfile
HOSTS = ( )

FILES = ( /home/chirico/file1 /home/chirico/file2 )

${FILES} -> ${HOSTS}
# Directory tmpdir will be created if it doesn't exist
install tmpdir ;
special /home/chirico/file1 "/usr/sbin/sendmail -bv";
notify chirico@squeezel;

/home/chirico/file2 ->
install /home/chirico/tapedest;
special /home/chirico/tapedest "chmod +r /home/chirico/tapedest";

/home/chirico/file2 ->
install /home/chirico/closetdest;

Step 2: Command from to run myDistfile above

Below is the command that will execute the contents in "myDistfile". This command is run from the
computer "". All output will go in the file "cmd1rdist.log".

$ rdist -P /usr/local/bin/ssh -f ./myDistfile -l file=./cmd1rdist.log=all

Obviously you want a secure copy (using scp), so the -P option uses ssh as your secure
transport mechanism.

TIP 169:

Restricting root logins (/etc/securetty). ctl-alt-F4 will give you a prompt for tty3. Note
that it is one number less. Take a look at the contents of "/etc/securetty". To prevent
root from logging in on this device, take out tty3 from this listing. Note, you can always
login as another user, then, su to root. Below is an example of the default
"/etc/securetty" that allows root to login to everything.

[root@squeezel ~]# cat /etc/securetty

TIP 170:

Perl map function. Try the following to get a quick take on this function,
which increments each value in the array a;

@a = (1,2,3);
map {$_++} @a;
map { print "$_\n" } @a;


@a = (1,2,3);
map { print "$_\n"} map {++$_} @a;

And you can easily make modifications, like reversing the order

@a = (1,2,3);
map { print "$_\n"} reverse map {++$_} @a;

Plus there is a grep() function that works on each element as well

@a = (1,2,3);
map { print "$_\n"} reverse grep{ $_ > 3} map {++$_} @a;

To get only odd numbers in reverse order:

@a = (1,2,3);
map { print "$_\n"} reverse grep{ !($_ % 2)} map {++$_} @a;


TIP 171:

Perl - subroutine call and shifting through variables. A simple and useful

sub test {
local $mval;
while( $mval = shift ) {
print " $mval\n";


TIP 172:

Tcp wrappers - First "/etc/hosts.allow" is check, and if there is an entry in this file, no more
checking it done. If are no matches in "/etc/hosts.allow", the "/etc/hosts.deny" file is checked
and if a match is found, that service is blocked for that host.

Example "/etc/hosts.deny" file:


The above file blocks access to computer It's also possible to run commands when
someone from this computer tries to ssh in. This example sends mail.

sshd: spawn (echo -e "%d %h %H %u"| /bin/mail -s 'hosts.deny entry' root)

Of course, you can also run commands in the "/etc/hosts.allow" if you wanted mail sent for a successful

TIP 173:

pgrep, pkill - look up or signal process based on name and other attributes.

To quick find all instances of ssh running, for user root, execute the following

$ pgrep -u root -l ssh

To kill a process, or send a signal use the "pkill" option. For example, to
make syslog reread its configuration file:

$ pkill -HUP syslogd

Another command command is "pidof" that can tell you how many processes are running.
This can be useful for detecting DOS attacks.

$ pidof sshd
4783 4781 30008 30006 29888 29886 2246

Above there are 7 sshd's running. Reference "Tcpdump, Raw Socket and Libpap Tutorial"
at [].

TIP 174:

Password Cracking - tools to check your users passwords:

John The Ripper



TIP 175:

Password Aging - setting the number of days a password is valid.

$ chage -M 90

TIP 176:

Kernel Performance Tuning - /Documentation/sysctl/vm.txt documents kernel settings to
improve performance. Below are some examples.

overcommit_memory: 0 -- default estimates the amount of memory for malloc
1 -- kernel pretends there is always enough memory until it runs out
3 -- never overcommit

$ cat /proc/sys/vm/overcommit_memory

The Linux VM subsystem avoids excessive disk seeks by reading
multiple pages on a page fault. The number of pages it reads
is dependent on the amount of memory in your machine.

The number of pages the kernel reads in at once is equal to
2 ^ page-cluster. Values above 2 ^ 5 don't make much sense
for swap because we only cluster swap data in 32-page groups.

$ cat /proc/sys/vm/page-cluster

This is used to force the Linux VM to keep a minimum number
of kilobytes free. The VM uses this number to compute a pages_min
value for each lowmem zone in the system. Each lowmem zone gets
a number of reserved free pages based proportionally on its size.

$ cat /proc/sys/vm/min_free_kbytes

This file contains the maximum number of memory map areas a process
may have. Memory map areas are used as a side-effect of calling
malloc, directly by mmap and mprotect, and also when loading shared

While most applications need less than a thousand maps, certain
programs, particularly malloc debuggers, may consume lots of them,
e.g., up to one or two maps per allocation.

The default value is 65536.

$ cat /proc/sys/vm/max_map_count

Also see

TIP 177:

IO Scheduler - /Documentation/block/as-iosched.txt documents kernel settings for disk

If you're not sure what partitions you have "$ cat /proc/partitions". This example
assumes hda, and you can see some of the kernel settings:

$ ls /sys/block/hda/queue/iosched
back_seek_max back_seek_penalty clear_elapsed fifo_batch_expire fifo_expire_async
fifo_expire_sync find_best_crq key_type quantum queued


TIP 178:

iozone -- getting data on disk performance ( This is a very
comprehensive package.

$ wget
$ tar -xf iozone3_242.tar
$ cd iozone3_242/src/current
$ make linux

At this point you should read the documentation. There is no "make install". You
copy it to each filesystem you want to run this program on. Below are some quick
start commands.

Good comprehensive test.

$ iozone -a

I prefer this for small filesystems. It limits the record size to 10000 and does
the output in operations per second (higher numbers mean faster drive).

$ ./iozone -a -s 10000 -O

TIP 179:

history - bash command to get a history of all commands typed. But, here is a way
that you can get date and time listed as well.

$ HISTTIMEFORMAT="%y/%m/%d %T "

Defining the environment variable above give you the date/time info when you
execute history:

$ history
175 05/06/30 12:51:46 grep '141.162.' mout > mout2
176 05/06/30 12:51:48 e mout2
177 05/06/30 12:56:59 ls
178 05/06/30 12:57:02 ls
179 05/06/30 12:57:39 ls
180 05/06/30 12:57:49 ls -l
181 05/06/30 13:01:10 history
182 05/06/30 13:01:20 HISTTIMEFORMAT="%y/%m/%d %T "
183 05/06/30 13:01:23 history

TIP 180:

.config - Fedora Core getting the .config to rebuild the kernel. You can find
this file, the ".config" file at the following location:

$ ls "/lib/modules/$(uname -r)/build/.config"

Or, to see the contents

$ cat "/lib/modules/$(uname -r)/build/.config"

This can be important, if you're planning to build your own kernel.

TIP 181:

Listing control key settings.

$ stty -a
speed 38400 baud; rows 0; columns 0; line = 0;
intr = ^C; quit = ^\; erase = ; kill = ; eof = ^D; eol = ; eol2 = ; start = ^Q;
stop = ^S; susp = ^Z; rprnt = ^R; werase = ^W; lnext = ^V; flush = ^O; min = 1; time = 0;
-parenb -parodd cs8 -hupcl -cstopb cread -clocal -crtscts
-ignbrk -brkint -ignpar -parmrk -inpck -istrip -inlcr -igncr icrnl ixon -ixoff -iuclc -ixany -imaxbel
opost -olcuc -ocrnl -onlcr -onocr -onlret -ofill -ofdel nl0 cr0 tab0 bs0 vt0 ff0
isig icanon iexten -echo echoe echok -echonl -noflsh -xcase -tostop -echoprt echoctl echoke

TIP 182:

iptables DNAT and SNAT. You have a webserver on When people query this webserver, you want them
to goto, with no indication that they are going to another web server. In fact, they always make
their web hits to

The following is the iptables commands:

$ echo 1 > /proc/sys/net/ipv4/ip_forward
$ iptables -t nat -A PREROUTING -d -p tcp --dport 80 -j DNAT --to
$ iptables -t nat -A POSTROUTING -d -s -p tcp --dport 80 -j SNAT --to

Change to whatever source you expect the web browser to come in on. Below is the tcpdump showing
all traffic is relayed via

[root@closet iptables]# tcpdump -nN port 80

17:34:58.790398 IP > S 3620106373:3620106373(0) win 16384
17:34:58.790465 IP > S 3620106373:3620106373(0) win 16384
17:34:58.790703 IP > S 1973665156:1973665156(0) ack 3620106374 win 5840
17:34:58.790720 IP > S 1973665156:1973665156(0) ack 3620106374 win 5840
17:34:58.790951 IP > . ack 1 win 17520
17:34:58.790965 IP > . ack 1 win 17520
17:34:58.791451 IP > P 1:327(326) ack 1 win 17520
17:34:58.791472 IP > P 1:327(326) ack 1 win 17520
17:34:58.791973 IP > . ack 327 win 6432

Above the web client is on "". You can see that the 1st server "" then goes out to
the 2nd server "" on the second line. The third line shows the 2nd server "" responding to
the 1st server, and the forth line passes this data back to the web client "".

Note: You can save your current iptables setting with the following command:

$ iptables-save > iptables_store

The big advantage is that you can store the counters as well.

$ iptables-save -c > iptables_store_w_cnts

To restore the file, use the following:

$ iptables-restore -c < iptables_store_w_cnts

TIP 183:

mailstats - display mail statistics. This file reads data from "/var/log/mail/statistics"

[root@closet ~]# mailstats
Statistics from Sat Jun 25 15:59:52 2005
M msgsfr bytes_from msgsto bytes_to msgsrej msgsdis msgsqur Mailer
4 1 2K 0 0K 0 0 0 esmtp
9 0 0K 1 2K 0 0 0 local
T 1 2K 1 2K 0 0 0
C 1 0 0

TIP 184:

Profiling C Applications - Assume you have the following program p1.c:

/* Program p1.c */

t1(int i)
printf("t1:%d\n", i);

t2(int j)
printf("t2:%d\n", j);

int main(void)
int i, j;

for (i = 0; i < 5; ++i) {
for (j = 0; j < 2; ++j) {

Compile the program as follows:

$ gcc -pg -g -o p1 p1.c
$ ./p1

Next, to get the profile graph.

$ gprof -p -b p1
Flat profile:

Each sample counts as 0.01 seconds.
no time accumulated

% cumulative self self total
time seconds seconds calls Ts/call Ts/call name
0.00 0.00 0.00 10 0.00 0.00 t2
0.00 0.00 0.00 5 0.00 0.00 t1

Above note the 10 calls to t2 and 5 calls to t1.

TIP 185:

CDPATH - this is a bash variable like PATH that defines a search path
for the cd command.

Suppose you have the following directory structure:

|-- dirA
`-- dirB

Assume you define CDPATH as follows:


Now, no matter what directory you are in if you use the cd command below
you will automatically move to "/home/chirico/stuff/dirA".

$ cd dirA

Note you could be in "/etc" and will move directly to "/home/chirico/stuff/dirA".
This command has the same format as PATH - multiple entries are separated by a colon.
If the current directory contain a sub-directory dirA, then, it gets priority.

The following is part of my .bash_profile


TIP 186:

Groups - add groups and users to groups. The following shows how to create the group "share"
and add the user "chirico" to this group. The following should be done as root, and
assumes the account "chirico" already exits.

$ groupadd share
$ usermod -G share chirico

Note the change made to "/etc/group" below:

$ cat /etc/group|grep 'share'

If the user chirico is currently logged in, he should run the following
command to immediately have group "share" rights. Or, the next time he logs
in he will have access to this group.

$ newgrp share

Reference the following (TIP 6, TIP 167).

TIP 187:

oprofile - steps for running oprofile on Fedora.

Step 1:

Find out what version of the kernel you are running.

$ uname -a
Linux 2.6.12-1.1398_FC4 #1 Fri Jul 15 00:52:32 EDT 2005 i686 i686 i386 GNU/Linux

Step 2:

Download the source in a chosen directory. Above, I'm running 2.6.12-1, but I'm going to go for, since
it's a little later. You want the signed file as well.

$ wget
$ wget

Now, check the signature.

$ gpg --verify linux- linux-

Step 3:

Unpack the file.

$ tar -xzf linux-
$ cd cd linux-

Step 4:

Copy the ".config" used to compile your previous kernel. You should find it
in the following direcotry "/lib/modules/$(uname -r)/build/.config".

Copy it to the linux- directory.

$ cp "/lib/modules/$(uname -r)/build/.config" .

Step 5:

Run make as follows. It will ask for a few questions on "make oldconfig". The
make installs below will have to be done with root privileges.

$ make oldconfig
$ make bzImage
$ make modules
$ make modules_install
$ make install

Step 6:

Edit the "/boot/grub/grub.conf" and set default = 0 as shown below in this

title Fedora Core (
root (hd0,2)
kernel /vmlinuz- ro root=/dev/VolGroup00/LogVol00 rhgb quiet
initrd /initrd-
title Fedora Core (2.6.12-1.1398_FC4)
root (hd0,2)
kernel /vmlinuz-2.6.12-1.1398_FC4 ro root=/dev/VolGroup00/LogVol00 rhgb quiet
initrd /initrd-2.6.12-1.1398_FC4.img
title Fedora Core (2.6.11-1.1369_FC4)
root (hd0,2)
kernel /vmlinuz-2.6.11-1.1369_FC4 ro root=/dev/VolGroup00/LogVol00 rhgb quiet
initrd /initrd-2.6.11-1.1369_FC4.img
title Other
rootnoverify (hd0,1)
chainloader +1

Step 7:

Shutdown with the restart option.

$ shutdown -r now

Step 8:

Run opcontrol. The commands below are done as root. My kernel was compiled in the following
directory "/home/kernel/linux-", so I'll run opcontrol as follows:

$ opcontrol --vmlinux=/home/kernel/linux-

Now start.

$ opcontrol --start
Using 2.6+ OProfile kernel interface.
Reading module info.
Using log file /var/lib/oprofile/oprofiled.log
Daemon started.
Profiler running.

Shutdown opcontrol.

$ opcontrol --shutdown

Run report.

$ opreport

CPU: CPU with timer interrupt, speed 0 MHz (estimated)
Profiling through timer interrupt
samples| %|
156088 99.8746 vmlinux
60 0.0384
30 0.0192 oprofiled
23 0.0147
13 0.0083 bash
12 0.0077 screen
10 0.0064 sshd
9 0.0058 ssh
6 0.0038 ip_tables
6 0.0038
5 0.0032 b44
5 0.0032 ext3
5 0.0032
4 0.0026 ip_conntrack
4 0.0026 jbd
2 0.0013 grep
1 6.4e-04
1 6.4e-04

Reference the following for more documentation:

TIP 188:

cyrus-imapd with Postfix using sasldb for authentication. For this example
the server is and the user is chirico.

Step 1:

$ yum install cyrus-imapd
$ yum install cyrus-imapd-utils

You need "cyrus-imapd-utils" for cyradm.

Step 2:

Edit /etc/imapd.conf

configdirectory: /var/lib/imap
partition-default: /var/spool/imap
admins: cyrus
sievedir: /var/lib/imap/sieve
sendmail: /usr/sbin/sendmail
hashimapspool: true
# Chirico Commented the below line
# sasl_pwcheck_method: saslauthd
# Because using sasldb
sasl_pwcheck_method: auxprop
sasl_auxprop_plugin: sasldb
# Chirico end change
sasl_mech_list: PLAIN
tls_cert_file: /usr/share/ssl/certs/cyrus-imapd.pem
tls_key_file: /usr/share/ssl/certs/cyrus-imapd.pem
tls_ca_file: /usr/share/ssl/certs/ca-bundle.crt

Step 3:

Create a user and password:

$ saslpasswd2 -c -u `postconf -h myhostname` cyrus
$ saslpasswd2 -c -u `postconf -h myhostname` chirico
$ saslpasswd2 -c -u `postconf -h myhostname` allmail

This will automatically create the file /etc/sasldb2. But look
at the default rights, assuming you ran saslpasswd2 as root:

$ ls -l /etc/sasldb2
-rw-r----- 1 root root 12288 Jul 31 09:50 /etc/sasldb2

We need to correct this in step 4.

Step 4:

$ chown root.mail /etc/sasldb2
$ ls -l /etc/sasldb2
-rw-r----- 1 root mail 12288 Jul 31 09:50 /etc/sasldb2

Step 5:

Update "/etc/postfix/". Note in /etc/imapd.conf the configdirectory
points to /var/lib/imap, and if I look at this directory I see the
socket directory. However, after staring /etc/init.d/cyrus-imapd there
will be a socket file "/var/lib/imap/socket/lmtp". (See step 6).

mailbox_transport = lmtp:unix:/var/lib/imap/socket/lmtp
mailbox_transport = cyrus

Restart postfix.

/etc/init.d/postfix restart

Step 6:

Start cyrus-imapd and look for the socket file.

$ /etc/init.d/cyrus-imapd restart
Shutting down cyrus-imapd: [ OK ]
Starting cyrus-imapd: preparing databases... done. [ OK ]

Now you should see the lmtp file:

$ ls -l /var/lib/imap/socket/lmtp
srwxrwxrwx 1 root root 0 Jul 31 10:04 /var/lib/imap/socket/lmtp

Step 7:

Add users. Note, you may have to go back to step 3 to add them to /etc/sasldb2
as well.

$ su - cyrus
$ cyradm> cm user.chirico> quit

Now got back as root, and check that everything was created correctly.

$ ls /var/spool/imap/c/user/
total 8
drwx------ 2 cyrus mail 4096 Jul 31 10:21 chirico

Step 8:

Run a mail test. We'll do this as root to the chirico account.

$ mail -s 'First test' chirico
first test

Now, still as root check the maillog. Normally everything should work.

$ tail /var/log/maillog

However, I got the following error below.

Jul 31 10:29:03 tape postfix/cleanup[30124]: AE7CB1B34A4: message-id=<>
Jul 31 10:29:03 tape postfix/qmgr[30120]: AE7CB1B34A4: from=, size=315, nrcpt=1 (queue active)
Jul 31 10:29:03 tape pipe[30128]: fatal: pipe_comand: execvp /cyrus/bin/deliver: No such file or directory

If you get a similiar error, you may need to adjust the settting in /etc/postfix/

# This is the problem in /etc/postfix/
cyrus unix - n n - - pipe
user=cyrus argv=/cyrus/bin/deliver -e -r ${sender} -m ${extension} ${user}

My deliver file is the following

$ ls -l /usr/lib/cyrus-imapd/deliver
-rwxr-xr-x 1 root root 846228 Apr 4 18:59 /usr/lib/cyrus-imapd/deliver

So I need to change my /etc/postfix/ as follows:

# Fix because by deliver file is under /usr/lib/cyrus-imapd/deliver
cyrus unix - n n - - pipe
user=cyrus argv=/usr/lib/cyrus-imapd/deliver -e -r ${sender} -m ${extension} ${user}

If changes were needed, like I had to do, restart postfix

$ /etc/init.d/postfix restart

Now, if everything works, you should start to see numbers in the spool directory like "1." and

$ ls -l /var/spool/imap/c/user/chirico/
total 40
-rw------- 1 cyrus mail 545 Jul 31 10:44 1.
-rw------- 1 cyrus mail 547 Jul 31 10:45 2.
-rw------- 1 cyrus mail 1276 Jul 31 10:45 cyrus.cache
-rw------- 1 cyrus mail 153 Jul 31 10:21 cyrus.header
-rw------- 1 cyrus mail 196 Jul 31 10:45 cyrus.index

Step 9:

Local firewall.

# imap
iptables -A INPUT -p udp -s --dport 143 -j ACCEPT
iptables -A INPUT -p tcp -s --dport 143 -j ACCEPT

Step 10:

Configure cyrus-imapd to start for run-level 3 and 5.

# chkconfig --level 35 cyrus-imapd on


Something to watch out for:

Something to watch out for: If a user creates a .forward file in their shell account with the
following entry, then, mail will not get mail relayed to cyrus.

"|exec /usr/bin/procmail"

The /etc/maillog will show something like this:

to=, orig_to=, relay=local, delay=0,
status=sent (delivered to command: exec /usr/bin/procmail)

Remove the ".forward" file from their home directory and you'll get the following:

to=, relay=cyrus, delay=0,
status=sent (

mutt with IMAP? (See TIP 190)

TIP 189:

expand - convert tabs to spaces in a file.

$ expand How_to_Linux_and_Open_Source.txt > notabs

TIP 190:

mutt with imap - assume you have setup imap (see tip 188). Now how do you configure
your ".muttrc" file to automatically connect, securely to the IMAP server?

Below is an example of my ".muttrc" file. For this example, assume my password is "S0m3paSSw0r9".

$ cat .muttrc
set spoolfile = "imaps://
set imap_force_ssl=yes
set certificate_file=~/.mutt/certificates/72d31154.0

Now, you want to copy the certificate as a "file.pem" and run "c_rehash" to convert this
file to a number. See the article. See the following article on how to do this under the
fetchmail section.

This is a quick summary of creating this key.

$ openssl s_client -connect -showcerts > file.pem
$ c_rehash ~/.mutt/certificates

TIP 191:

Apache - CGI scripts. There are two ways to enable CGI scripts. The second method is the
prefered method.

First way, the easy way. Look for the "http.conf" file. On Fedora Core, this file can be
found under "/etc/httpd/conf/httpd.conf". Edit this file as follows to make
"" execute scripts.

ScriptAlias /chirico-cgi/ "/home/chirico/cgi-bin/"

Second way, the better way. Instead of doing the above, make the following change in

Options +ExecCGI
SetHandler chirico-cgi

Running a test script. Now copy the following test script into the directory "/home/chirico/cgi-bin"
and change the rights to execute for the user running this.

# Save as test.cgi
# chown apache.apache test.cgi
# chmod 700 test.cgi
echo "Content-Type: text/html"
echo "Hello world from user `whoami`! "

TIP 192:

Bash - using getopts for your bash scripts.

while getopts "ab:cd:" Option
# b and d take arguments
case $Option in
a) echo -e "a = $OPTIND";;
b) echo -e "b = $OPTIND $OPTARG";;
c) echo -e "c = $OPTIND";;
d) echo -e "d = $OPTIND $OPTARG";;
shift $(($OPTIND - 1))

TIP 193:

Sieve - creating sieve recipes with "sieveshell"

The following sieve script put all mail into the
folder jefferson. This assumes that I have already created the IMP
directory, or mail box (INBOX.jefferson), which can be done in mutt
with the "C" command. Below is an example of finding ""
anywhere in the header.

# This is a file named jefferson.siv
require ["fileinto"];
if header :contains "Received" "from" {
fileinto "INBOX.jefferson";

Now, from the command propt execute "sieveshell" with the hostname of the
imap server. My server is, so I would execute the

$ sieveshell
connecting to
Please enter your password:****
> put jefferson.siv
> activate jefferson.siv
> list
jefferson.siv <- active script
> quit

Note the put brings in the script. And you need to activiate it.

You can activate a sieve script for any user on your system if you are
root. This is an example of activating a script for user chirico. Assume
below the root prompt is "#".

# sieveshell -a chirico -u chirico

You can also automate everything from a bash script. But note after
the -e the commands, and not a file with the commands, follows within
quotes. This is the script I use for my home system.

sieveshell -a chirico -u chirico -e 'deactivate
delete chirico.siv
put chirico.siv
activate chirico.siv


TIP 194:

emacs - editing files remotely with tramp. Tramp comes with the latest version of emacs.
That means if you're using Fedora core 4, with emacs, you have tramp. This is
ideal for editing files on remote computers that do not use emacs.

Edit the ".emacs" file and add the following line:

(require 'tramp)
(setq tramp-default-method "scp")

Now, to edit a file on computer (C-x, C-f) and
enter the following in Find file:

Find file:/


TIP 195:

trusted X11 forwarding - running gnome and KDE both on one screen, at the same
time securely. The following assumes gnome is running on the current
computer and "" has KDE

$ ssh -Y
$ startkde

Or assume you want to run gnome on ""

$ ssh -Y
$ gnome-session

By default Fedora Core allows ForwardX11 over ssh. Note you want to use
the -Y option above and NOT -X.

Suppose you want a remote "gnome-session" on ctl-alt-F12. Below is an
example of getting the remote computer, and you
can still have the above configuration.

First you must allow magic cookies for each server connection.

$ MCOOKIE=$(mcookie)
$ xauth add $(hostname)/unix:1 MIT-MAGIC-COOKIE-1 $MCOOKIE
$ xauth add localhost/unix:1 MIT-MAGIC-COOKIE-1 $MCOOKIE

Again, note that you have to add this for EACH connection. So if you wanted 2 as well

$ MCOOKIE=$(mcookie)
$ xauth add $(hostname)/unix:2 MIT-MAGIC-COOKIE-1 $MCOOKIE
$ xauth add localhost/unix:2 MIT-MAGIC-COOKIE-1 $MCOOKIE

On create a new xterm. If :1 is take below
try :2. The vt12 is for switching to ctl-alt-F12.

$ xinit -- :1 vt12

Note, if you do not add the above cookies, you will get the follow error:

Xlib: connection to ":1.0" refused by server
Xlib: No protocol specified

The screen may be hard to read. At this point ssh -Y to the remote computer.

$ ssh -Y
$ gnome-session

Yes, you will get errors about sound and some custom drivers is the remove
computer has different hardware. After is loads, you can switch back and
forth between session with (ctl-alt-F12) and (ctl-alt-F7)

TIP 196:

Suspend ssh session - you have just sshed into a computer "ssh -l user", and you
want to get back to the terminal prompt of the computer you started with. Escapte, by
default with ssh is "~", so enter "~" followed by "ctl-z" to suspend.

TIP 197:

Quick way to send a text file

$ sendmail -f < /etc/fstab

Or you can use mutt and send a binary file

$ mutt -s "Pictures of the Kids" -a kids.jpg < text.txt

TIP 198:

size - determining the size of the text segment, data segment, and "bss" or uninitialized data segment.

$ size /bin/sh /bin/bash
text data bss dec hex filename
586946 22444 18784 628174 995ce /bin/sh
586946 22444 18784 628174 995ce /bin/bash

Note above that "/bin/sh" and "/bin/bash" have equal text,data and bss numbers. It's
highly likely that these are the same programs.

$ ls -l /bin/sh
lrwxrwxrwx 1 root root 4 Jan 14 2005 /bin/sh -> bash

Yep, it's the same program. Here's a further definition of each segment.

Text segment: The machine instructions that the CPU executes. This is usually
read only and sharable.

Data segment: Contains initialized variables in a program. You also know these
as declarations and definitions.

int max = 200;

Uninitialized data segment: Think of this as a declaration only, or data that
is only initialized by the kernel to arithmetic 0 or null pointers
before program execution.

char s[10];

TIP 199:

Using the at command.

Below is a simple example if running the ls command at 11:42am that
will send mail -m to the user that executed it.

We'll execute job1 defined as follows and set to be executable.

$ cat ./job1
date >> /tmp/job1

The at command is listed below. For queue "-q" names you can only
specify one letter. Here we're using x. The letter determines the
priority with "a" the highest.

$ at -q x -f ./job1 -m 11:54am
job 3 at 2005-10-04 11:54

Now, if you execute the atq command, you'll get the following.

$ atq
3 2005-10-04 11:54 x chirico

It's also possible to execute jobs at the command line entering
a ctl-d at the end of the input.

$ at -q x -m 12:08pm
at> ls -l
at> who
at> date
at> ^D

Or for a job to execute 1 minute from now.

$ at -q x -m `date -d '1 minute' +"%H:%M"`
at> ls -l
at> date

Important points: The atd daemon must be running. To check if
it's running do the following:

$ /etc/init.d/atd status

Also, if there is an /etc/at.allow file, then only users in that
file will be allowed to execute at.

If /etc/at.deny exists but is empty and there is no /etc/at.allow,
then, everyone can execute the at command.

TIP 200:

lsusb - command will display all USB buses and all devices connected.

$ lsusb
Bus 005 Device 003: ID 413c:2010 Dell Computer Corp.
Bus 005 Device 002: ID 413c:1003 Dell Computer Corp.
Bus 005 Device 001: ID 0000:0000
Bus 004 Device 001: ID 0000:0000
Bus 003 Device 003: ID 0fc5:1227 Delcom Engineering
Bus 003 Device 002: ID 046d:c016 Logitech, Inc. Optical Mouse
Bus 003 Device 001: ID 0000:0000
Bus 002 Device 001: ID 0000:0000
Bus 001 Device 001: ID 0000:0000

TIP 201:

Memory fragmentation - if you suspect workload memory fragmentation issues
and you want to monitor the current state of you system, then, consider
looking at the output from /proc/buddyinfo on recent kernels.

$ cat /proc/buddyinfo
Node 0, zone DMA 541 218 42 2 0 0 0 1 1 1 0
Node 0, zone Normal 2508 2614 52 1 5 5 0 1 1 1 0
Node 0, zone HighMem 0 1 3 0 1 0 0 0 0 0 0

The following definition is taken from ./Documentation/filesystems/proc.txt in the
Linux kernel source.

Each column represents the number of pages of a certain order which are
available. In this case, there are 0 chunks of 2^0*PAGE_SIZE available in
ZONE_DMA, 4 chunks of 2^1*PAGE_SIZE in ZONE_DMA, 101 chunks of 2^4*PAGE_SIZE
available in ZONE_NORMAL, etc...

TIP 202:

arp - Linux ARP kernel moduel. This command implements the Address Resolution Protocol.

This is an example of the command.

$ arp
Address HWtype HWaddress Flags Mask Iface ether 00:50:DA:60:5B:AD C eth0 ether 00:11:11:8A:BE:3F C eth0 ether 00:0F:66:47:15:73 C eth0

TIP 203:

dbench - performance monitoring.

So, how does your system react when the load average is above 600. Have you even seen a
computer with a load average of 600? Well, this could be your chance.


The following gives a load average of 10 on my system.

$ dbench 34

If you want a higher load, just increase the number.

TIP 204:

/etc guide - a listing of common files in the /etc directory.

/etc/exports: this file is used to configure NFS.

/etc/ftpusers: the users on your system who are restricted from FTP login.

/etc/motd: message of the day, which users see after login.

/etc/named.conf: DNS config file.

/etc/profile: common user information.

/etc/inittab: this file contains runlevel start information.

/etc/services: the services and their respective ports.

/etc/shells: this contains the names of all shells installed on the system.

/etc/passwd: this file contains user information.

/etc/group: security group rights.

TIP 205:

logger - is a bash command utility for writing to /var/log/messages or the
other files defined in /etc/syslog.conf.

$ logger -t TEST more of a test here

This is what shows up in /var/log/messages

Oct 28 07:15:50 squeezel TEST: more of a test here

TIP 206:

accton, lastcomm - accouting on and last command. This is
a way to monitor users on your system. As root, you
would implement this as follows:

$ accton -h
Usage: accton [-hV] [file]
[--help] [--version]

The system's default process accounting file is /var/account/pacct.

Note the default file location is /var/account/pacct so we'll turn
it on system wide with the following command.

$ accton /var/account/pacct

Now take a look at this file. It will grow. To see command that
are executed, use the lastcomm command.

$ lastcomm

The above command gives output for all users. To get the data
for user "chirico" execute the following command:

$ lastcomm --user chirico

You can also get a summary of commands with sa.

[chirico@big ~]$ sa
30 5.23re 0.00cp 10185k
11 4.83re 0.00cp 8961k ***other
8 0.13re 0.00cp 19744k nagios*
4 0.00re 0.00cp 2542k automount*
3 0.00re 0.00cp 680k sa
2 0.13re 0.00cp 17424k check_ping
2 0.13re 0.00cp 978k ping

To turn off accounting, execute accton without a filename.

$ accton

TIP 207:

CPU Temperature on a laptop. The following is the temperature
of my Dell laptop.

$ cat /proc/acpi/thermal_zone/THM/temperature
temperature: 58 C

TIP 208:

script -f with mkfifo to allow another user to view what you type
in real-time.

Step 1. Create a fifo (first in first out) file that the other
user can view. For this example create the file /tmp/scriptout

[chirico@laptop ~]$ mkfifo /tmp/scriptout

Step 2. Have the second user, voyeur user, cat this file. Output will block
for them until you complete step 3. The other user, voyer,
is executing the command below.

[voyeur@laptop ~]$ cat /tmp/scriptout

Step 3. The original user runs the following command.

[chirico@laptop ~]$ script -f /tmp/scriptout
Script started, file is /tmp/scriptout

Now anything typed, including a vi session, will be displayed to the
voyeur user in step 2.

See TIP 46.

TIP 209:

fsck forced on next reboot. To do this, as root issue the following commands.

$ cd /
$ touch forcefsck

Now reboot the system, and when it comes up fsck will be forced on the system.

$ shutdown -r now

TIP 210:

/dev/random and /dev/urandom differ in their random generating properties. /dev/random
only returns bytes when enough noise has been generated from the entropy pool. In
contrast /dev/urandom will always return bytes.

Reference: (rand.c)

TIP 211:

Want to find out the speed of your NIC? (Full Duplex or Half), then use ethtool.

[root@squeezel ~]# ethtool eth0
Settings for eth0:
Supported ports: [ MII ]
Supported link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Half 1000baseT/Full
Supports auto-negotiation: Yes
Advertised link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Half 1000baseT/Full
Advertised auto-negotiation: Yes
Speed: 100Mb/s
Duplex: Full
Port: Twisted Pair
Transceiver: internal
Auto-negotiation: on
Supports Wake-on: g
Wake-on: d
Current message level: 0x000000ff (255)
Link detected: yes

TIP 212:

rpm install hang? You might need to delete the lock state information.

$ nl /etc/rc.d/rc.sysinit | grep rpm
720 rm -f /var/lib/rpm/__db* &> /dev/null

Note the command

$ rm -f /var/lib/rpm/__db*

Because sometimes you will run "rpm -ivh somerpm" and it will just sit

TIP 213:

Apache - limit access to certain directories based on IP address in the
httpd.conf file.

You can do this completely from /etc/httpd/conf/httpd.conf which
are shown below for multiple IP addresses. Note that all 3 setting
are the same.

However, the following is different only allows to

Some complete settings in /etc/httpd/conf/httpd.conf

Order allow,deny
Allow from # All 10.
Allow from # All 192.168
Allow from 127 # All 127.

Here's an example that only allows access to .html files
and nothing else for a particular directory.

Satisfy All
Order allow,deny
Deny from all

Order deny,allow
Allow from all
Satisfy Any

Don't forget to reload httpd with the following command.

$ /etc/init.d/httpd reload

TIP 214:

Open Files - determining how many files are currently open.

$ cat /proc/sys/fs/file-nr
2030 263 104851
| | \- maximum open file descriptors
| |
| \- total free allocated file descriptors
(Total allocated file descriptors since boot)

Note the maximum number can be set or changed.

$ cat /proc/sys/fs/file-max

To change this

$ echo "804854" > /proc/sys/fs/file-max

Note lsof | wc -l will report higher numbers because this includes
open files that are not using file descriptors such as directories,
memory mapped files, and executable text files.

and also see the man page for this: man 5 proc )

TIP 215:

Ctrl-Alt-Del will cause an immediate reboot, without syncing dirty buffers by
setting the value > 0 in /proc/sys/kernel/ctrl-alt-del.

$ echo 1 > /proc/sys/kernel/ctrl-alt-del

(Reference: man 5 proc)

TIP 216:

Redefining keys in X using xev and xmodmap. The program xev, used in an X window
terminal screen will display information on mouse movements, keys pressed and

$ xev

Now type shift-4 and you'll notice the event details below:

KeyPress event, serial 29, synthetic NO, window 0x3800001,
root 0x60, subw 0x0, time 55307049, (418,242), root:(428,339),
state 0x1, keycode 13 (keysym 0x24, dollar), same_screen YES,
XLookupString gives 1 bytes: (24) "$"
XmbLookupString gives 1 bytes: (24) "$"
XFilterEvent returns: False

KeyRelease event, serial 29, synthetic NO, window 0x3800001,
root 0x60, subw 0x0, time 55307184, (418,242), root:(428,339),
state 0x1, keycode 13 (keysym 0x24, dollar), same_screen YES,
XLookupString gives 1 bytes: (24) "$"

So, if you want to redefine this key to say copyright, see (/usr/X11R6/include/X11/keysymdef.h)
you would type the following.

$ xmodmap -e 'keycode 13 = 4 copyright'

To get the key back to the dollar, issue the following command.

$ xmodmap -e 'keycode 13 = 4 dollar'

By the way it's possible to define multiple key codes for a sigle key. You'll need
to have a key defined as the Mode_switch. Perhaps you'd like to use the Windows key,
or the key with the Microsoft logo on it, since you're using Linux. This key is
keycode 115

$ xmodmap -e 'keycode 115 = Mode_switch'

Now you could define 3 values to the shift-4. For this example use ld, Yen and dollar.

$ xmodmap -e 'keycode 13 = 4 dollar sterling yen'

So pressing the keys gives you the following:

shift-$ (dollar sign)
Windows-$ (lb sign)
Windows-shift-$ (Yen sign)

You could go crazy and redefine all you keys.

(Thanks to hisham for this tip).

TIP 217:

Threads - which version of threads are you using?

NPTL 2.3.90

For a history on threads used with gcc reference the following:

TIP 218:

Screenshots using ImageMagick.

If you want the entire screen, execute the following:

$ import -window root screen.png

Or to crosshair select the region with your mouse, execute
the following instead.

$ import screen.png

KDE has the ability to take screenshots with the command below.

$ ksnapshot

GNOME likewise has a command too.

$ gnome-panel-screenshot --delay 6

Visting ImageMagick again, the xwininfo command give window information and the id can be
used to capture images with the import command.

$ xwininfo

xwininfo: Please select the window about which you
would like information by clicking the
mouse in that window.

xwininfo: Window id: 0x1e00007 "chirico@squeezel:/work/svn/souptonuts - Shell - Konsole"

Absolute upper-left X: 4
Absolute upper-left Y: 21
Relative upper-left X: 0
Relative upper-left Y: 0
Width: 880
Height: 510
Depth: 24
Visual Class: TrueColor
Border width: 0
Class: InputOutput
Colormap: 0x20 (installed)
Bit Gravity State: NorthWestGravity
Window Gravity State: NorthWestGravity
Backing Store State: NotUseful
Save Under State: no
Map State: IsViewable
Override Redirect State: no
Corners: +4+21 -396+21 -396-493 +4-493
-geometry 880x510+0+0

Now use the import command with the Window id. My example is shown below.

$ import -window 0x1e00007 id.miff

And to quickly display this image that you just saved, use the display command.

$ display id.miff

TIP 219:

File Access over SSH using FUSE (Filesystem in USErspace). This is a very good way to
mount a remote filesystem locally. It's like a secure NFS mount, but you don't require
admin privileges on the remote computer. You do need to have fuse-sshfs installed on
the local computer that will perform the filesystem mount.

The following works with Fedora Core 5. Only the users added to the fuse group can mout
external drives. Below the user chirico is being added to the group fuse.

$ yum install fuse-sshfs
$ usermod -a -G fuse chirico

You'll need to reboot.

$ shutdown -r now

Next I'm going to mount the remote filesystem This is done as user chirico
on the local computer. I'm using root on the remote computer because I
want to mount the complete drive.

$ mkdir v0
$ sshfs v0
$ cd v0
$ ls -l
bin dev home lost+found media mnt opt q sbin srv tmp var
boot etc lib master_backup misc net proc root selinux sys usr

Now to unmount the filesystem

$ fusermount -u /home/chirico/v0

Yes, you can mount the filesystem on boot. Below shows an example entry for /etc/fstab, but
this only allows user on the current system to view what is is /mnt/v0. /mnt/v0 fuse defaults 0 0


TIP 220:

OpenVPN - A full-featured SSL VPN solution. The following demonstrates
a very simple OpenVPN setup between two Fedora Core 5 computers and

As root install the package on both computers.

$ yum -y install openvpn

Setup on

$ iptables -A INPUT -p udp -s --dport 1194 -j ACCEPT
$ iptables -A INPUT -i tun+ -j ACCEPT
$ iptables -A INPUT -i tap+ -j ACCEPT
$ iptables -A INPUT -i tap+ -j ACCEPT
$ iptables -A FORWARD -i tap+ -j ACCEPT

Note - make sure you have commented out the following line
in /etc/sysconfig/iptables

# -A RH-Firewall-1-INPUT -j REJECT --reject-with icmp-host-prohibited

Now from continuting with the commands that need to be executed on do one of the following

$ openvpn --remote --dev tun1 --ifconfig --verb 9

The above statement gives lots of errors. Once it's working you may want
the following statement without the --verb 9 option.

$ openvpn --remote --dev tun1 --ifconfig

After you finish the setup commands for immediately below, you'll be
able to access as

Setup on

$ iptables -A INPUT -p udp -s --dport 1194 -j ACCEPT
$ iptables -A INPUT -i tun+ -j ACCEPT
$ iptables -A INPUT -i tap+ -j ACCEPT
$ iptables -A INPUT -i tap+ -j ACCEPT
$ iptables -A FORWARD -i tap+ -j ACCEPT

Note - again, make sure you have commented out the following line
in /etc/sysconfig/iptables

# -A RH-Firewall-1-INPUT -j REJECT --reject-with icmp-host-prohibited

The openvpn commands are are reversed from what is shown

$ openvpn --remote --dev tun1 --ifconfig --verb 9

$ openvpn --remote --dev tun1 --ifconfig

Now you can access all services and ports from on for
such services as MySQL, secure Web, imap, etc. A quick test is nmap as follows:

$ nmap -A -T4

Starting Nmap 4.03 ( ) at 2006-05-20 13:54 EDT
Interesting ports on
(The 1671 ports scanned but not shown below are in state: closed)
22/tcp open ssh OpenSSH 4.3 (protocol 2.0)
111/tcp open rpcbind 2 (rpc #100000)
3306/tcp open mysql MySQL (unauthorized)

Nmap finished: 1 IP address (1 host up) scanned in 7.116 seconds

TIP 221:

openssl - Some common commands.

Finding the openssldir (Directory for OpenSSL files).

$ openssl version -a|grep OPENSSLDIR
OPENSSLDIR: "/etc/pki/tls"

Connect to a secure SMTP server with STARTTLS, assuming the server name is

$ openssl s_client -connect -starttls

Reference (

TIP 222:

Bash functions. This is easy, and I find it very useful to create bash functions
for repeated commands. For example, suppose you want to create a quick bash function
to cd to /var/log, tail messages and tail secure. You can create this function as

[root@v5 log]# m()
> { cd /var/log
{ cd /var/log
> tail messages
tail messages
> tail secure
tail secure
> }

Above I'm typing m() then hitting return. Note the echo on the next line followed
by the prompt >. I then enter {.

TIP 223:

Stats on DNS Server. You can get stats on your DNS server.

The following works for BIND 9:

$ rndc stats

On my system I see the output in "/var/named/chroot/var/named/data/named_stats.txt", which
if an FC4 system. By the way, if you're using BIND 8, the command is "ndc stats", but that
has a completely different format.

Format of the output

+++ Statistics Dump +++ (1153791199)
success 297621
referral 32
nxrrset 21953
nxdomain 33742
recursion 28243
failure 54
--- Statistics Dump --- (1153791199)

The number (1153791199) can be converted with the date command.

$ date -d '1970-01-01 1153791199 sec'
Tue Jul 25 02:33:19 EDT 2006

That's 1153791199 seconds since 1970-01-01 UCT. Which is 4 hours fast,
from EDT.

TIP 224:

snmp - simple network monitoring protocol. The following steps setup snmp on Fedora Core 5.

$ yum install net-snmp*

Next add the following line in "/etc/snmp/snmpd.conf" at the bottom.

rocommunity pA33worD

Start the snmp service.

$ /etc/init.d/snmpd restart

Once started, from the command prompt, it's possible to get stats on the computer.

$ snmpwalk -v 1 -c pA33worD localhost system
$ snmpwalk -v 1 -c pA33worD localhost interface

$ snmpgetnext -v 1 -c pA33worD localhost sysUpTime
DISMAN-EVENT-MIB::sysUpTimeInstance = Timeticks: (26452) 0:04:24.52

Note the Timeticks is in 100th of a second. So the computer above has been running
for 264.52 seconds.

Reference( TIP 225 shows how to use MRTG for gathering snmp stats).

TIP 225:

MRTG - Multi Router Traffic Grapher.

$ cfgmaker --output=/etc/mrtg/ \
ifref=ip --global "workdir:/var/www/html/mrtg/stats"\


TIP 226:

Back Trace - This is a method of getting a back trace for all processes on the system.
it assumes the following: a. Kernel was build with CONFIG_MAGIC_SYS-REQ
enabled (which Fedora 5 kernels are) b. You can get direct access to the

Step 1.

Ctl-Alt-F1 (This brings you to the text console)

Step 2.


Note above that's Alt-ScrollLock followed by Ctl-ScrollLock. You should see
a lot of text on the screen. To fast to read, but don't worry the text will
be in /var/log/messages at the end.

On my system the ScrollLock key is next to the NumLock key.

TIP 227:

Ext3 Tuning - One advantage of Ext3 over Ext2 is directory indexing, which imporves file
access in directories containing large files or when the directory contains
many files. Directory indexing improves performance by using hashed binary

There are two ways to enable dir_index. First, find the device using the mount

$ mount

/dev/mapper/VolGroup00-LogVol00 on / type ext3 (rw)
proc on /proc type proc (rw)
sysfs on /sys type sysfs (rw)
devpts on /dev/pts type devpts (rw,gid=5,mode=620)
/dev/sda1 on /boot type ext3 (rw) <--- This is the one you want
tmpfs on /dev/shm type tmpfs (rw)
none on /proc/sys/fs/binfmt_misc type binfmt_misc (rw)
sunrpc on /var/lib/nfs/rpc_pipefs type rpc_pipefs (rw)
automount(pid2001) on /net type autofs (rw,fd=4,pgrp=2001,minproto=2,maxproto=4)

From the above command, the device used is /dev/sda1. Using the tune2fs command,
directory indexing will only apply to directories created after running the
command below.

$ tune2fs -O dir_index /dev/sda1

However, if you want it to apply to all directories, use the e2fsck command as
shown below:

$ e2fsck -D -f /dev/sda1

You'll need to bypass the warning message.

Reference: "Tuning Journaling File Systems: A small amount of effort an dtime can yield big
results",by Steve Best. Linux Magazine, September 10, 2006. This author as has
a very good book titled: "Linux Debugging and Performance Tuning."

TIP 228:

NIC bonding - binding two or more NICs to one IP address to improve performance. The following
instructions were done on Fedora Core 5.

Step 1.

Create the file ifcfg-bond0 with the IP address, netmask and gateway. Shown
below is my file.

$ cat /etc/sysconfig/network-scripts/ifcfg-bond0


Step 2.

Modify eth0, eth1 and eth2. Shown below are each one of my files. Note that
you must comment out, or remove the ip address, netmask, gateway and hardware
address from each one of these files, since settings should only come from
the ifcfg-bond0 file above. I've chosen to comment out the lines, instead of
removing, should I decide to unbond my NICS sometime in the future.

$ cat /etc/sysconfig/network-scripts/ifcfg-eth0

# Linksys Gigabit Network Adapter
# Settings for Bond

$ cat /etc/sysconfig/network-scripts/ifcfg-eth1

# Linksys Gigabit Network Adapter
# Settings for bonding

$ cat /etc/sysconfig/network-scripts/ifcfg-eth2

# Linksys Gigabit Network Adapter

Step 3.

Set the load parameters for bond0 bonding kernel module. Append the
following lines to /etc/modprobe.conf

# bonding commands
alias bond0 bonding
options bond0 mode=balance-alb miimon=100

Step 4.

Load the bond driver module from the command prompt.

$ modprobe bonding

Step 5.

Restart the network, or restart the computer. Note I restarted to computer,
since my NICs above had MAC assignments.

$ service network restart # Or restart computer

Take a look at the proc settings.

$ cat /proc/net/bonding/bond0
Ethernet Channel Bonding Driver: v3.0.3 (March 23, 2006)

Bonding Mode: adaptive load balancing
Primary Slave: None
Currently Active Slave: eth2
MII Status: up
MII Polling Interval (ms): 100
Up Delay (ms): 0
Down Delay (ms): 0

Slave Interface: eth2
MII Status: up
Link Failure Count: 0
Permanent HW addr: 00:13:72:80:62:f0

Good, well written article describing the steps above.
Documentation for bonding that can also be found in the kernel

TIP 229:

/etc/nsswitch.conf - System Databases and Name Service Switch configuration file.

This file determines lookup order of services. For example, to match a name
to an IP address, an entry can be put into the /etc/hosts file. Or a DNS query
can be made. What's the order? Normally, it's the entry in the /etc/hosts file.
because /etc/nsswitch.conf contains the following setting

hosts: files dns

See man nsswitch.conf for more settings.

TIP 230:

Finding DST settings on the live system. In 2007 Daylight Saving Time was extended in the United
States, Canada, and Bermuda. Before this change we adjusted the clocks on the last Sunday in
October - Not anymore. We now change it on the first Sunday in November.

$ zdump -v EST5EDT |grep '2007'

EST5EDT Sun Mar 11 06:59:59 2007 UTC = Sun Mar 11 01:59:59 2007 EST isdst=0 gmtoff=-18000
EST5EDT Sun Mar 11 07:00:00 2007 UTC = Sun Mar 11 03:00:00 2007 EDT isdst=1 gmtoff=-14400
EST5EDT Sun Nov 4 05:59:59 2007 UTC = Sun Nov 4 01:59:59 2007 EDT isdst=1 gmtoff=-14400
EST5EDT Sun Nov 4 06:00:00 2007 UTC = Sun Nov 4 01:00:00 2007 EST isdst=0 gmtoff=-18000

Correct settings for EDT are shown above. Note, the months Mar and Nov.

You can also run the same command by location.

$ zdump -v /usr/share/zoneinfo/America/New_York|grep '2007'

Note: This time conversion file can be created manually. For instructions on how to perform
this task, execute the following command.

$ man zic

zic is the time zone compiler.


TIP 231:

Qt - Compiling Qt 4 programs statically to run on remote systems that do
have Qt 4 libraries installed. You actually download the Qt 4 source

Step 1 - Download Qt 4.

You will download a separate version of Qt 4. Yes, even if you have
Qt 4 installed on your system, you'll want to download another
version to statically compile your programs. I performed the
following steps on my computer:

$ mkdir -p /home/src/qt
$ wget
$ cd /home/src/qt
$ tar -xzf qt-x11-opensource-src-4.2.2.tar.gz

Note, make sure you get the latest version of Qt. When I'm wrote this it
was 4.2.2. Check for updates.

Step 2 - Compile Qt for static mode

The text step is to compile qt for static mode.

$ cd /home/src/qt/qt-x11-opensource-src-4.2.2
$ ./configure -static -prefix /home/src/qt/qt-x11-opensource-src-4.2.2
$ make sub-src

At this point Qt 4 is installed in static mode.

Step 3 - Set PATH

Now set the PATH to reference this version.

$ PATH=/home/src/qt/qt-x11-opensource-src-4.2.2/bin:$PATH
$ export PATH

Step 4 - Compile Your Source

My program source is located in /home/chirico/widgetpaint

$ cd /home/chirico/widgetpaint
$ qmake -project
$ qmake -config release
$ make

TIP 232:

SELinux - FC6 quick fix for problems. Using system-config-securitylevel to
fix simple problem.

$ ssh -Y user@servertofix
$ system-config-securitylevel

You do not have to ssh into the computer as root. As long as X is running
"init 5", then you can run the system-config command above and it will
ask you for the root password.

TIP 233:

Mutt - tagging multiple messages and moving them to a different folder.

If you want to tag multiple messages with mutt, use the capital T, when
in mutt.

~A (To tag all messages. Note, enter the tilda "~" without quotes)
;s (After entering ;s, you'll be asked where to save the message)

From here you can create a new fold. If you're using IMAP mail boxes, then
use C to create a mailbox.

To delete messages without exiting mutt, enter "$", without the quotes.

(Reference: )

TIP 234:

Mutt - color coding message in mutt.

The following is written in the .muttrc file.

color index brightblue default Poker
color body brightyellow default Error

Note, the first line will color blue all indexes with
the word Poker. The second operates on the body of the

TIP 235:

cat - header, stdin, and footer. (Working with /dev/fd/0 or -)

If you have data from a command that you want preceded by
the contents of a header file and followed by data in
a footer file, then, the following command may help.

$ w|cat header /dev/fd/0 footer

Above the output of the "w" command follows the contents of
the header file. Note "/dev/fd/0" refers to stdin. Yes, you
could use "-" in its place in this situation. However, if
"-" is used as the first argument, it will be interpreted as
as a command line option, whereas "/dev/fd/0" would not.

No comments: