AIX troubleshooting : PART 1

IBM has announced "AIX5L".  It's essentialy AIX Version 5.  The 'L' stands
for "Linux Affinity".  A statement that AIX is going to support some of the
Linux API's and interfaces (for instance: the /proc filesystem)

Some changes to the filesystem limits, virtual IP's, dynamic dealocation
of swapspaces.

Using SMIT is probably very different from your normal way of doing
system administration, but could prove very useful in the long run. In
some areas, in particular TCP/IP, NFS, etc., you can also do things the
normal way, but it is unfortunately difficult to know exactly when the
normal way works. Again, always using SMIT is probably your best way
to go, even when you have to learn a new tool. 

What SMIT actually does is build up commands with all required options
to perform the functions requested and execute them. The commands
called and the output they produce are stored in the files smit.script
and smit.log in your home directory. Looking in smit.script may teach
you more about system administration.

How do I import an /etc/passwd or /etc/group file
                 from another box?

If the other box is non-AIX,  copy the password and group entries for
the non-system users into AIX's /etc/passwd and /etc/group files.
Then run /bin/pwdck -t ALL.  This will create the proper entries in
the shadow password file (/etc/security/users).  You should also run
usrck and grpck.

To duplicate the password and  group entries from another AIX box,
copy /etc/passwd, /etc/group, /etc/security/passwd, /etc/security/group,
/etc/security/user, /etc/security/limits, /etc/security/environ.  The
last three are optional unless you modified them.  If you modified
/etc/security/login.cfg, you should also copy that file.

How to fsck the root filesystem

You can run fsck either in maintenance mode or on mounted filesystems.
Try this:

  1. boot from diskette (AIX 3 only --- AIX 4 boot from CD or tape)
  2. select maintenance mode
  3. type /etc/continue hdisk0 exit (replace hdisk0 with boot disk if
     not hdisk0)
  4. fsck /dev/hd4

How can I unmount /usr to run fsck on it?

In order to fsck /usr, it has to be unmounted. But /usr cannot be
unmounted because /bin is symbolically linked to /usr/bin. Also
/etc/fsck is symbolically linked to /usr/sbin/fsck.

To work around this, when you boot from the boot/maintenance diskettes
and enter maintenance mode, enter "getrootfs hdisk0 sh" instead of
"getrootfs hdisk0" where hdisk0 is the name of the boot disk. Then run
"fsck /dev/hd2".


How do I see/change parameters like number of
                processes per user?

You can use SMIT as described below or simply use lsattr/chdev.
The former will list the current setting as in:

  # lsattr -E -l sys0 -a maxuproc
  maxuproc 40 Maximum # of processes allowed per user True

and you can then increase the maxuproc parameter:

  # chdev -l sys0 -a maxuproc=200
  sys0 changed

If you just type 'lsattr -E -l sys0' you will get a list of all
parameters, some of which can be changed but not others.

If you want to use smit, do as follows:

smit
  System Environments and Processes
    Change / Show Operating System Parameters
      - on this screen you can change by overtyping the following fields:
        - Maximum number of PROCESSES allowed per user
        - Maximum number of pages in block I/O BUFFER CACHE
        - Maximum Kbytes of real memory allowed for MBUFS
      - toggle fields exist for:
        - Automatically REBOOT system after a crash  (false/true)
        - Continuously maintain DISK I/O history (true/false)


How do I shrink the default paging space on hd6?


create a paging space to use temporarily
   mkps -s 20 -a rootvg
change default paging space hd6 so it is not used at next reboot
   chps -a n hd6
swapon /dev/paging00

sysdumpdev -p /dev/paging00

Update information in boot logical volume

   bosboot -ad /dev/hdisk0 

reboot

remove current hd6 and create a new one of smaller size
   rmps hd6
   mklv -y hd6 -t paging rootvg 

swapon /dev/hd6

change the dump device back to hd6:

   sysdumpdev -p /dev/hd6

Update information in boot logical volume

   bosboot -ad hdisk0


change current paging device (paging00) so it is inactive at next boot
   chps -a n /dev/paging00

shutdown, reboot, remove paging00 using the command:
    rmps paging00

You can check your paging space with `lsps -a`

The swapper seems to use enormous amounts of paging space, why?

When you run ps, you may see a line like:

USER   PID %CPU %MEM    SZ   RSS     TT STAT  TIME CMD
root     0 0.0%  14% 386528  8688      -    S 17:06 swapper

This is normal behavior, the swapper looks to ps like it has the entire
paging space plus real memory allocated.

How do I remove a committed lpp?


installp has a new option, uninstall (-u) which can be used to remove lpps.  BEWARE of pre-requisite chains.

How can I recover space after installing updates?


Note: If you are a /usr server, do not use this because the files
      mentioned below are needed by /usr clients and cannot be deleted.

Installp creates numerous files in /usr to clean up after failed/rejected installs and also for de-installing uncommitted lpps. Once you have COMMITted packages you can remove these files safely. Depending on your installation activity the numbers can be significant: hundreds-to-thousands of files, megabytes of data.

Files eligible for removal are associated with each "product" you have installed; the largest collection being due to bos. After COMMITting bos lpps, you may safely remove all files of the form:

/usr/lpp/bos/deinstl*
              /usr/lpp/bos/inst_U4*
              /usr/lpp/bosadt/deinstl*
        and   /usr/lpp/bosadt/inst_U4*


You may repeat this for all additional COMMITted products (e.g.,bostext1, bosnet, xlc) you have on your system.

This problem of lingering install files is a known defect in installp. If you have installed PTF U411711 (or any superseder of it: U412397, U413366, U413425) the deadwood in /usr will not be quite as prevalent. No single PTF currently available completely corrects this problem.

On my own 320, the following freed up 12.4M in /usr:

       # rm -R /usr/lpp/bos/deinstl*
       # rm -R /usr/lpp/bos/inst_U4*

Where are the AIX log files kept?

AIX logs messages as specified in /etc/syslog.conf.  Here's an
example

#
*.err;kern.debug;auth.notice;user.none          /dev/console
*.err;kern.debug;daemon,auth.notice;mail.crit;user.none /var/adm/messages
lpr.debug                                       /var/adm/lpd-errs

*.alert;kern.err;daemon.err;user.none           operator
*.alert;user.none                               root
*.emerg;user.none                               *

# for loghost machines, to have authentication messages (su, login, etc.)
# logged to a file, un-comment out the following line and adjust the 
# file name as appropriate.
#
# if a non-loghost machine chooses to have such messages 
# sent to the loghost machine, un-comment out the following line.
#
auth.notice                     /var/log/authlog
mail.debug                      /var/log/syslog

# following line for compatibility with old sendmails. they will send
# messages with no facility code, which will be turned into "user" messages
# by the local syslog daemon. only the "loghost" machine needs the following
# line, to cause these old sendmail log messages to be logged in the
# mail syslog file.
#
user.alert                       /var/log/syslog
#
# non-loghost machines will use the following lines to cause "user"
# log messages to be logged locally.
#
user.err                         /dev/console
user.err                         /var/adm/messages
user.alert                       `root, operator'
user.emerg                       *


How can I log information about ftp accesses to a file?


1) In /etc/syslog.conf, add the line:
   daemon.debug  /tmp/daemon.log

2) # touch  /tmp/daemon.log
   # refresh -s syslogd

3) Modify your inetd.conf so that ftpd is called with the "-l" flag.
   You may also want the "-d" flag. This can be done with 'smit inetdconf'.

All the syslog messages from various system daemons should now appear in
the file "/tmp/daemon.log".



How do I find a file name from the inode number?


ncheck -i nnnn /mntpoint

AIX troubleshooting : PART 2

How do I set the tty name associated with a physical port?

Let's say you wanted to make a tty on the s1 port and call it rs0000 and a tty on the s2 port and call it rs0001.

You could run:

mkdev -c tty -s rs232 -t tty -l rs0000 -p sa0 -w s1   # creates rs0000
and
mkdev -c tty -s rs232 -t tty -l rs0001 -p sa1 -w s2   # creates rs0001

How do I use mksysb to clone a system?


I use the following steps on the master machine to clone an AIX system:
1) Remove the password from root.
2) Remove the NIS line from the end of the /etc/group file (the last
   line with the +: )
3) Change most of the level '2' designations in /etc/inittab to level
   '3' to prevent them from being started up when the new system is
   booted (the minimum ones to change are rc.nfs and rc.tcpip)
4) Boot in service mode and change the name and ip address to a "spare"
   set to avoid address collision.
5) Clear /tmp, /usr/tmp and /usr/spool/lpd/stat.
6) Run mkszfile and edit it to be sure /usr is as small as possible;
   then mksysb from the command line.

The above changes allow me to boot in normal mode the first time, get in
as root, change the above files back and do the other things necessary
to configure the new system.

Then, of course, I go back and clean up and reboot my master machine.

Note: 1 and 2 lets you log in even if you can't get on the network. 
   It prevents the login process from trying to reach an NIS server. 
   Step 2 needed only if you use NIS.


How do I remove a non-existant physical volume?

To delete a phantom disk from the ODM use reducevg with the pvid instead of the disk name.  You are running some command such as lsvg or varyonvg and it is griping about a disk that is no longer findable right?  In that warning message, it should give you a pvid.  Try one of the following, (note: reducevg updates the VGDA but not the ODM).

reducevg -f  

ldeletepv -g VGid -p PVid
      -g Required, specify the VGid of the volume group you are
         removing the physical volume from
      -p Required, specify the PVid of the PV to be removed



How do I kill a process that ignores   kill -QUIT -KILL -STOP

If there is i/o pending in a device driver, and the driver does not catch the signal, you can't kill it - a reboot is the only way to clear it.

Furthermore, if the process stays hung for more than a few minutes, you can find out what device is wedged by doing this --

% echo trace -k $(expr  / 256) | crash | tee stack





How can I see "console" messages?

Use the swcons command to redirect the console to a file.  Or use chcons to do it permanently.

How do I merge my /etc/password and /etc/security/password for Crack?


/usr/sbin/mrgpwd.  You must have permissions to read /etc/security/password.


I lost the root password, what should I do?


Boot from boot diskettes, bootable tape, or bootable CD.  
At the Installation/Maint menu select item 4, "Start a limited function
 maintenance shell.
At the subsequent "#" prompt enter the command:
 getrootfs hdiskN
 (where "N" is replaced by the number of a disk on your system
 that is in rootvg.)
That will run for about a minute or so and you get a # prompt back.  At this
 point you are logged in as root in single user mode.
Change to /etc/security and edit the passwd file.  Delete the three lines
 under root: password, update time (or whatever it's called), and
 flags.  Save the file.
Then at the prompt, give root a new password.
Shutdown/reboot in normal mode.  Log in with new password.

SOURCE CODE CONTROL SYSTEM (SCCS)

SCCS Overview
The source code control system allows versions of a program to be stored in a special file, so that any version may be retrieved. There are a few commands involved (not all of them listed here). All source code files start with 's.'

get -r

- get a program out of source code to read only. Missing out the -r flag gets the most recent version. e.g.
$ get $SCUK/s.parser.c # extracts file parser.c from source code file $SCUK/s.parser.c as read only. See get -e for editing.

get -e 

- get a piece of code out for edit, so that the code may be modified and a new version created using 'delta'. e.g.
$ get -e $SCUK/s.parser.c # extracts file parser.c from source code file $SCUK/s.parser.c for editing. See get for read-only.

delta 

- you must be in the directory with the modified piece of code when you execute this command. This adds the latest version to the source code file. e.g.
$ delta $SCUK/s.parser.c # writes file parser.c to the source code file $SCUK/s.parser.c . See get -e for information on how to extract the file from source code.

prs 

- show comments/details on source code file.

admin -r -i

- create a new source code file with progam. -r specifies the initial revision of the program and may be missed out (default is 1.1 I think). Must be spaced correctly! admin is also used for sccs administration, but it gets to fear and loathing time pretty fast. e.g.
admin -iparser.c $SCUK/s.parser.c # creates a new source code file called $SCUK/s.parser.c from the file parser.c

unget 

- cancels a get -e

SCREEN COMMUNICATION

echo

- a command mainly used in shell scripts. Examples:
$ echo "Hello" # will print Hello on your screen
$ echo "Hello" > /dev/tty616 # will print Hello on someone elses screen (warning - can crash their screen!)
$ echo $DESTF10 # will print the value of the environment variable DESTF10
$ echo "\033Fdemo demo" # will echo demo to the status bar at the top of a wyse terminal
See also file shellscripts

read

- will read text from standard input and place it in the variable name specified. See file shellscripts

line

- waits until the user presses return before carrying on (writes what is typed to standard output). If used in a crontab/at job this instruction is ignored. See file shellscripts

talk 

- set up an interactive communication dialogue box between two users. Looks good but isn't really that useful.

write 

- writes a message to someone elses screen. Try typing 'write root' and then type a message, finishing with control-D.

banner 

- writes in huge letters across your screen! (max: 10 chars per word)

wall 

- send a message to all people on a system. Can only be executed by root (I think).

tput 

- tty type independent attribute setting (requires TERM variable and TERMCAP to be set). I only know these few bits:

• tput cnorm - turns the screen cursor on
• tput civis - turns the screen cursor off
• tput clear - clears the screen
• tput smso - turns all new text to bold
• tput rmso - turns all bold text off

tee (-a)

- command used in pipes to take a copy of the standard output. e.g.
$ ls | tee /tmp/x # would output ls normally and put a copy in /tmp/x. The option '-a' is used to append rather than replace files.

TAPES AND DISKS

Please see this page for more information on disks in AIX

dd if= of= bs= conv=sync

- direct (and I mean DIRECT) copy, normally to tape. Archaic syntax and very rarely used. flags:

• if - input filename or device
• of - output filename
• bs - block size
• conv - ??

e.g. To write a file to tape use
$ dd if=/etc/hosts of=/dev/rmt0 bs=1024 conv=sync # write hosts file to tape using dd

cpio

stands for copy in-out, and is extremely powerful if you can cope with the innumerable flags that you have to use(!)
$ cpio -iBcvumd "etc/hosts"

 # Grab /etc/hosts file from tape
$ find /etc -print | cpio -oBcv >/dev/rmt0 # Write the contents of the /etc directory to tape
$ find /etc -print | cpio -pdumv /usr2/etcbackup/ # copy directory /etc to /usr2/etcbackup and retain all permissions.
meaning of the flags:

• i - input
• o - output
• B - Block size of 5120 bytes
• c - read/write header info
• v - list file names
• u - unconditional copy - overwrites existing file.
• m - keep modification dates
• d - creates directories as needed.
• t - generate listing of what is on the tape.
• p - preserve permissions.

tapeutil -f

- A program which came with the tape library to control it's working. Called without arguments gives a menu. Is useful for doing things like moving tapes from the slot to the drive. e.g.
$ tapeutil -f /dev/smc0 move -s 10 -d 23 # which moves the tape in slot 10 to the drive (obviously, this will depend on your own individual tape library, may I suggest the manual?).

doswrite -a

- copy unixfile to rs6000's floppy disk drive in DOS format. -a option expands certain characters, for certain ascii conversions.

dosdir 

- show list of files on a dos floppy disk. Useful with option -l (long format). Like dos command 'dir'

dosread -a

- copy dos file in floppy disk drive to unix - if UNIXFILE is omitted, it outputs to the screen.

dosdel 

- delete dos file on floppy disk.

dosformat

- format dos floppy disk (High Density)

tar

- Read/Write stuff to archive.
tar cvf /dev/rmt0  # will write files to tape
tar xvf /dev/rmt0 will read files from tape
tar tvf /dev/rmt0 will give a listing of what's on the tape. If you're using an archive file then replace /dev/rmt0 in the examples above with the name of the archive file.

CONNECTIVITY

exit

- end current shell process. If you log in, then type this command, it will return you to login. ^D (control-D) and logout (in some shells) does the same.

rlogin

- login to a remote machine, e.g.
$ rlogin hollandrs # log in to machine called hollandrs
Useful with -l option to specify username - e.g.
$ rlogin cityrs -l ismsdev # log in to machine cityrs as user ismsdev For further info about trust network see .rhosts file and /etc/resolv.conf (I think).

telnet

- very similar to rlogin except that it is more flexible (just type telnet with no arguments and then '?' to see the options). Useful because you can specify a telnet to a different port.

ftp

- File Transfer Protocol - a quick and easy method for transferring files between machines. The .netrc file in your $HOME directory holds initial commands. type ftp without arguments and then '?' to see options)

rcp

- Remote copy. Copies a file from one unix box to another, as long as they trust each other (see .rhosts file or /etc/resolv.conf I think). Options

• -f (to force the copy to occur)
• -r (to recursively copy a directory)
• -p (to attempt to preserve permissions when copying)

su -

- switch user, option '-' means that the users .profile is run, without option you merely assume the id and permissions of the user, without (for example) changing PATH and DBPATH, e.g.
$ su - root # become root
$ su root # gain permissions of root but don't change the current environment variables
$ su - vlink # switch to user vlink
If you are root, you may su to any other user without being prompted for a password. su without arguments is the same as 'su root'. Note that the 'su' option is not available on all UNIX machines as it can crash some of them.

ping 

- check that is alive and well (do not expect an immediate response from a machine that is linked over an ISDN line). Firewalls often block ping packets after the Ping of Death so quite often you'll find you can't ping internet sites either. Options include:

• -q ping quietly
• -i wait no of seconds between each packet sending. The default is 1 second. If you are using ping to keep an ISDN line up then using something like $ ping -i 5 -q hollandrs is ideal.
• -f Never use this! Sends as many packets as it possibly can as fast as possible, used for network debugging and is likely to slow networks horribly when used. Known as 'flood' pinging.
• -c send no of packets before giving up

To check that your machine can ping, try pinging 127.0.0.1 - this acts as a feedback loop, checking the network card's ability to ping.

rsh 

- remote shell - e.g.
$ rsh altos more /tmp/chk # will run the command more the file /tmp/chk on the machine called altos. Useful in pipes for example. rsh on its own will execute a login. Use option '-l' to specify logon name. You can also use rcmd and remsh on other flavours of unix.

host 

- lookup the ip address in the /etc/hosts file and give its name

SYSTEM COMMANDS

kill -

- sends a signal (normally a kill) to a process. kill -9 terminates the job no questions asked, kill -15 tries to clear up as much as possible - e.g. remove semaphores and such-like. Other signals may be sent as well, see manual and /usr/include/sys/signal.h to see what signals you can send to a process.

renice 

- make a process not hog the system so much by setting its nice value.

smit

- system admin program for AIX

df

- list volume groups + usage. see also lsvg. Usually used with the -k flag so the number of blocks is displayed in 1024-blocks.

cu -l

- log on to device such as a pad or a modem. See related files /etc/uucp/* and /etc/locks and /etc/services

stty sane

- Changes terminal settings back to normal. If a tetra module for example crashes your screen so that no keys function except ^C which doesn't even do very much then typing ^Jstty sane^J should cure the problem. To fully cure the problem you also need to type stty tab3 (and stty -ixon if you're feeling a little overzealous)

stty

- allows you to change terminal settings such as the interrupt key, quit key, etc. e.g.
$ stty intr ^A # would change the interrupt key to being control-A
$ stty quit ^L #would set the quit key (normally ^\) to control-L. other key changes are:

• erase (normally ^H)
• xon (normally ^Q)
• xoff (normally ^S)
• eof (normally ^D)

To really annoy a systems administrator, change interrupt to 't' and quit to '^D' . hehehehehehe

lscfg

- show all connected devices

lsvg

- list volume groups (see related file diskhelp)

lspv

- list physical disks (and see related file diskhelp)
lspv without arguments will produce a list of all the hard-disks used. lspv  will produce a list of information about the hard disk. lspv -l  will show any logical volumes which are mapped on to that drive.

lsdev

- list devices. Options:

• -C list Configured devices
• -P list Possible devices

produces different output when you are root.

mkdev

- make devices. e.g. To make a tty:
# Script to add a tty. Options that need amending are:
# -l name of tty to be created - e.g '-l tty600' wil create
# a tty called 'tty600'
# -p RAN name
# -w Port number on RAN
# -a Attributes (e.g. to set up auto login, etc.)
mkdev -c tty -t 'tty' -s 'rs232' -l tty433 -p sa2 -w 2 -a term='wyse50' -a forcedcd='enable' -a login='enable' -a speed='19200'e.g. To create a printer (raw device):
mkdev -c printer -t 'osp' -s 'rs232' -p 'sa3' -w '10' -l label2 -a xon='yes' -a dtr='no' -a col=500
It is highly recommended that you make and change devices using smit

chdev

- change devices. See mkdev

cc

- c compiler, use with

• -o
• -O optimise
• -w or -W all warning flags.

shutdown

- shutdown the system so that it may be switched off. Rather obviously, this may only be run by root. Options:

• -f shuts the system down immediately (rather than waiting for a minute)
• -R reboot the system immediately after halt

oslevel

- show the current revision of the operating system.

GENERAL INFORMATION COMMANDS

smon

- monitor's system usage - F5 shows processes which are hogging the machine. Not available on AIX 4.1 and above sadly.

uptime

- shows how long the system has been up and how hard it is being hammered. The load average fields show how many jobs on average are waiting. <1 5="" around="" bad="" good="" is="" less="" not="" or="" pretty="" though="" unusual="" very="">10 the machine is being seriously hammered.

who

- list users who are currently logged on (useful with option 'am i' - i.e. 'who am i' or 'whoami')

w

- list users and what they are doing, including idle time. The first line is the output from uptime

id

- similar to whoami except that it does a direct check to see who you are - who only checks /etc/utmp so any su commands will be ignored.

ps

- list processes currently running, by default on the current shell. Useful with options:

• -t - show all processes running on a terminal
• -ef - show all processes
• -u - show all processes owned by a user
• -flp - show as much information as you can about a process number
• -aux - show processes in order of usage of the processors. Useful to see what processes are hogging system resources.

fuser -u

- show who is using a file.(system hogging command). Useful when trying to work out who has locked a row or table in an informix database for example.

lpstat -p

- show the current status of a printer and any jobs in the queue. lpstat without arguments prints all of them.

enable 

- enable a printer queue. You must be root or a member of the printq group to run this command.

disable 

- disable a printer queue. You must be root or a member of the printq group to run this command.

enq 

- examine spool queue for printers.

uname -a

- will show you what machine you're currently on.

ipcs

- list semaphores and shared memory.

ipcrm -s

- remove semaphore or shared memory.

crontab

- use -l to list all regular scheduled jobs. To alter them, use option -e

at 

- perform a job at a specified time. (Useful for running something at a later date). at retains the current environment. e.g.
$ at now + 5 minutes
echo "Phone Julie McNally" > /dev/tty616
^D
job compjmd.389748732 will be run at ???Will echo to tty616 the message "Phone Julie McNally" in 5 minutes. e.g.2
$ at 0331235930
echo "April fools day!" > /dev/console
^Dwill echo "April fools day!" to the console at 11:59 and 30 seconds, on the 31st of march. Format for this is: [YYYY]MMDDhhmmss. at jobs are sometimes used in the place of crontab's because if the machine is off when the crontab is meant to take place, the job never happens. at jobs automatically start when the machine is switched on if the machine was down at the time. typing at -l will show you all the at jobs you have queued, at -r  will remove an at job (only the owner or root is allowed to do this).

date

- show current date and time. This command may also be used to set the system clock (ONLY WHEN EVERYONE IS LOGGED OFF) with a root user id. A date change is never simple, even when adjusting things by an hour. The safest way to do it is to change the date then reboot the machine because otherwise the crontab daemon may start doing jobs at odd times. I believe there might be a 'go slow/fast' option to set the clock, and the clock will then run 'slower/quicker' until it catches up with the required time.

last 

- shows a list of recent logins. It looks at /var/adm/wtmp so it only shows initial logins, and not whether those users have been su'd to.

fileplace -pv

- show the physical (as in disk location) location of a file. Useful for tracing informix files, and perhaps for working out whether defragmentation copying is required.

EXTREMELY USEFUL COMMANDS

ls -l

- lists files in a directory in long format. You cannot do without this. Here's a more detailed explanation. e.g.
$ ls -l

Part 1	Part 2	Part 3	Part 4	Part 5	Part 6	Part 7
-rw-rw-rw-	1	root	staff	28	Jan 16 09:52	README
-rw-------	1	compjmd	staff	4304	Jun 24 12:21	tabledict
drwxrwxrwx	2	compjmd	staff	512	Jul 1 16:30	testdir
-rwxrwx---	1	compjmd	system	0	Jul 1 16:30	a.out

... is a sample listing.

• Part 1: Permissions - see chmod for explanation of these. If the first field is set, then the file in question is not really a file at all, but something else, key:
  • -: normal file
  • d: directory
  • l: symbolic link created by 'ln'
  • c or b: device of some sort
  You may sometimes see an 's' where the 'x' should be in the permissions - this is normally on executable files which change other files. e.g. Permissions of 'sqlexec' the file that executes all informix queries should be '-rwsr-sr-x' - this then accesses tables with permissions of '-rw-rw----'. where the table files are owned by informix (group informix). the 's' flags allows changing of the database tables on a program level, but not on a unix level. (can change contents via sqlexec but not use 'rm' command on db file).
• Part 2: Number of links to this file (directories always have 2+).
• Part 3: The owner of the file - e.g. If the owner is 'compjmd' and permissions are set to -rw------- then only the user 'compjmd' may read or write to that file. Again, if owner is "compjmd" and permissions are -r-x------ then only the user compjmd may read or execute that file. Only the owner of a file or root may chmod it.
• Part 4: The group ownership of the file - (bloody hell, this is getting complicated). On a unix system there are certain 'groups' which users can belong to, held in the file '/etc/group'. You will notice that in this file there will be a main group, e.g. 'staff' which contains every user. Which means that any user listed under staff is in that group.....right...every file has a group attached to it. Which means that if a file had permissions ----rw---- and a group reference of 'system', then only users who were part of the group system could modify that file. To see which groups the current user belongs to do id. Sorry if this wasn't comprehensible but you should never need to use this anyway(!).
• Part 5: Size of the file in bytes
• Part 6: Time of last modification
• Part 7: The name of the file

Useful options (and there are loads more). All may be combined except where specified:

• ls -a show files starting with '.' too
• ls -A show files starting with '.' but not '.' or '..'
• ls -c must be used with either option l and/or t - displays/sorts by modification time
• ls -d do not show subdirectory listings
• ls -i display the i-node number of each file
• ls -t Put the listing in time order (see options u and c)
• ls -r Put the listing in reverse order - usually used with a -t
• ls -u must be used with either options l and/or t - displays/sorts by last-access time

vi 

- love it or loathe it - the standard operating system text-file editor. See Related help file. Vi You can also use 'view' which forces Read only (-R opt). vi + enters the file at the specified line no. Also, vi +/ will enter the file and move to the first occurrence of . e.g.
$ vi +/"love it or loathe it" handycommands
Users new to vi hate it. I personally managed to get through University without using it ever (I used Joe's own editor instead). If I accidentally went into vi, I had to ^Z and kill the job. Sigh. Five years of using vi means that I'm getting a little better at it now... (I'm actually typing this now in a vi-clone for Windows).

grep 

- a phenomenally useful command which matches strings within files - e.g.
$ grep D7523 mcall_reps.out # will find all the lines in mcall_reps.out that have the string "D7523" in it. Also incredibly useful for things like pipes,e.g.
$ du | grep cred # (in /home directory will show all users that have 'cred' in their title). You may use regular expression matching - e.g.
$ grep "main.*{" x.c # would match any line containing 'main' and an open curly brackets at any point in the line afterwards. There are two variations to grep - fgrep and egrep which do virtually the same things as grep, but are either faster (having less options) or more complex (but slower). See also section on Wildcards
Options:

• -v : show all lines that do not contain pattern.
• -y : don't bother matching case
• -i : don't bother matching case
• -c : show count of matching lines rather than the lines themselves
• -l : show filename's instead of matching lines.

ksh -o vi

- The Korn Shell - pros might notice that I don't mention using the C-Shell at all - I've never used it, so that's why it doesn't appear. A Shell is a program that you run your commands in. Typing exit will end the current shell. The -o vi option of the korn shell allows vi commands to work at the shell prompt after pressing escape. For example, pressing escape and then 'k' will bring up the last command used in the shell.

awk

- this would be a damn useful command if I knew how to use it properly. see alternative page awkhelp

man 

- look at the manual, e.g.
$ man ps # will list the manual page for the command ps

OTHER FILE HANDLING COMMANDS

type 

- show where the source of a command is: e.g.
$ type sendmail
sendmail is /usr/sbin/sendmail
This command is merely an alias for 'whence -v'

whence 

- show where the source of a command is: shell builtin command. See type
Use option: -v for verbose mode

which 

- show where the source of a command is held. Almost the same as type and whence

chmod 

- change file permissions. e.g.
$ chmod 666 handycommands
changes the permissions (seen by ls -l) of the file handycommands to -rw-rw-rw-
r = 4, w = 2, x = 1. In the above example if we wanted read and write permission for a particular file then we would use r + w = 6. If we then wanted to have the file have read-write permissions for User, Group and All, then we would have permissions of 666. Therefore the command to change is that above.
$ chmod 711 a.out
Changes permissions to: -rwx--x--x
Additional explanation of file permissions and user/group/all meaning are given in the description of ls -l
You may specify chmod differently - by expressing it in terms of + and - variables. For example
$ chmod u+s /usr/bin/su
will modify the "sticky bit" on su, which allows it to gain the same access on the file as the owner of it. What it means is "add s permission to user". So a file that started off with permissions of "-rwxr-xr-x" will change to "rwsr-xr-x" when the above command is executed. You may use "u" for owner permissions, "g" for group permissions and "a" for all.

chown 

- Change ownership of a file. Must be done as root. e.g.
chown informix *.dat # change all files ending .dat to be owned by informix

chgrp 

- Change group ownership of a file. Must be done as root. e.g.
chgrp sys /.netrc # change file /.netrc to be owned by the group sys

mvdir 

- move a directory - can only be done within a volume group. To move a directory between volume groups you need to use mv -r
or find -print | cpio -pdumv ; rm -r

cpdir 

- copy a directory. See mvdir

rmdir 

- this is crap - use rm -r instead

mkdir 

- Creates a directory. e.g.
$ mkdir /tmp/jon/ # create directory called /tmp/jon/ 

find  -name "searchkey" -print

- search for files - e.g.
$ find . -name "system.log" -print # will find all files (with full path names) called system.log - Wildcards are allowed, e.g.
$ find /tmp -name "sl.*" -atime +0 -print # will print out all files in /tmp/ that start sl. and which haven't been accessed for a day. Helpful for finding lost files, or finding stuff in enormous directories. Other useful options include:

• -atime + - finds files that haven't been accessed for 1+days also, ctime (creation time) and mtime (modify time)
• -prune - stay in current directory - don't look in dirs off the directory specified in path names - e.g.
  $ find /tmp -user "compgnc" -prune -print # will find all files in /tmp which user compgnc owns and will not search lower directories (e.g. /tmp/usr)
• -size + - finds files that are bigger than
• -exec rm {} \; - remove all files found...dangerous command - e.g.
  $ find /tmp -name "sl.*" -atime +0 -prune -print -exec rm {} \; # will remove all files in /tmp starting 'sl.' that haven't been accessed for a day. Spacing of this command is important! Most exec commands are possible:
  $ find /usr2/calltest -name "*.4gl" -print -exec grep "CHECK" {} \; | pg
• -ok - like exec only it prompts for confirmation after each occurence. e.g.
  $ find /tmp/disk7 -name "*" -print -ok doswrite -a {} {} \; # Please note that you MUST end any exec or ok option with an escaped semicolon (\;).
• -user - finds all files owned by
• -group - finds all files with a group of

ln -s

- create a symbolic link to a different directory from current directory: e.g.
$ ln -s /usr/uniplex/compgnc /u/compgnc/uni # would create a link called 'uni' in the directory /u/compgnc. From then on, typing cd uni would cd to /usr/uniplex/compgnc. You can also give two files the same name. e.g.
$ ln make.e_enquiry makefile # would link the two files so that they are identical, and when you change one, you change the other. You may also create a symbolic link to a host(!). Instead of typing 'rlogin hpserver' every time, by typing
$ ln -s /usr/bin/rsh hpserver # will create a link so that whenever you type 'hpserver' it will execute a remote shell on the machine.
Option -f forces the link to occur

head -

- prints out the first few line of a file to screen. Specify number to indicate how many lines (default is 10). e.g. If you sent something to a labels printer and it wasn't lined up, then you could print the first few labels again using:
$ head -45 label1.out | lp -dlocal1

tail -

- prints out the end of a file. Very similar to head but with a very useful option '-f' which allows you to follow the end of a file as it is being created.e.g.
$ tail -f vlink.log # follow end of vlink.log file as it is created.

wc -

- Word Count (wc) program. Counts the number of chars, words, and lines in a file or in a pipe. Options:

• -l (lines)
• -c (chars)
• -w (words)

To find out how many files there are in a directory do ls | wc -l

split -

- Splits a file into several files.e.g.
$ split -5000 CALLS1 # will split file CALLS1 into smaller files of 5000 lines each called xaa, xab, xac, etc.

tr 

- translates characters. e.g.
$ cat handycommands | tr "\t" " " # will take the file handycommands and translate all tabs into spaces. Useful when messing about with awk or you need to convert some input (e.g. that from tty) to a unique filename that does not contain special characters. e.g.
$ tty | tr "/" "." # produces for example .dev.pts.7

od 

- od converts nasty (binary save) files into character representations. Useful when back-compiling, examining raw .dat files,etc. Use with option '-c' for character display (recommended).

script

- starts recording everything in the shell to a file by default 'typescript'. Press ^D to finish the script. Provides a log of everything used. Has almost the same effect as $ ksh | tee typescript
Used for debugging shells, seeing error messages which flash off the screen too quickly, etc.

cut

- cut's the file or pipe into various fields. e.g.
$ cut -d "|" -f1,2,3 active.unl # will take the file active.unl which is delimited by pipe symbols and print the first 3 fields options:

• -d
• -f

Not too useful as you can't specify the delimiter as merely white space (defaults to tab) Alternatively, you can 'cut' up files by character positioning (useful with a fixed width file). e.g.
$ cut -c8-28 "barcode.txt" # would cut columns 8 to 28 out of the barcode.txt file.

paste

- paste will join two files together horizontally rather than just tacking one on to the end of the other. e.g. If you had one file with two lines:
Name:
Employee Number:and another file with the lines:
Fred Bloggs
E666then by doing:
$ paste file1 file2 > file3 # this would then produce (in file3).
Name: Fred Bloggs
Employee Number: E666Note that paste puts horizontal tabs between the files, so you may need a sed 's/   //g' command to get rid of these.

sort 

- sorts the information from the file and displays the result on standard output (stdout). e.g.
$ sort /tmp/list_of_names # will sort the file into alphabetical order, and display it to the screen. Useful with option '-u' to filter out duplicates.

uniq 

- filters out all duplicate lines from a file or input stream (file or stream must be sorted!). Useful with option -c which merely produces a count of unique lines.

ex 

- ex is an old line editor, and almost never used now (similar to DOS edlin if you remember that - me, I've repressed it). You are most likely to come across ex within the vi editor - all commands beginning with a colon (:) are ex commands

INPUTS, OUTPUTS AND WILDCARDS

Unix commands generally get their information from the screen, and output to it. There are three main 'streams' which unix uses to get/place it's information on. These streams are called:

• stdin (Standard Input) - normally, what you type into the screen
• stdout (Standard Output) - normally, what is output to the screen
• stderr (Standard Error) - normally, error messages which go to the screen

any of these may be redirected by the following symbols:

• <  take input from rather than the screen. e.g.
  $ ksh < x # will read all commands from the file x and execute them using the Korn shell.
• >  take output from the command and place it in . e.g.
  $ ls > x will place the output of the command 'ls' in the file x
• >>  take output from the command and append it to . e.g.
  $ ls /tmp >> x will place the output of the command 'ls' and append it to the file x
• 2>  take any error messages from the command and put it in . e.g.
  $ ls /tmp 2>/dev/null would throw away any error messages that are produced by ls (sorry, /dev/null is a file that, if written to, the information disappears never to be seen again).
• command1 | command2 Pipe - Takes the standard output of the first command, and turns it into the standard input of the second command. The output of the second command will then be put on the standard output (which, again, may be a pipe) e.g.
  $ ls | more will send the output of 'ls' into the command 'more', thus producing a directory listing which stops after every page. This method is called piping.

command1 & - the ampersand (&) forces command1 to run in the background. so that you may continue to type other commands in the shell, while command1 executes. It is not advisable to run a command in the background if it outputs to the screen, or takes it's input from the screen

See also tee which allows splitting of the input stream and output to several different places at once.

Wildcards

B Bib Baby Fox Fib

There are various wildcards which you may use. One is '*' which means 0 or more characters. e.g. 'B*' will match 'B,Bib and Baby' from the list above, another wildcard is '?' which matches 1 character, e.g. '?ib' will match 'Bib and Fib'. Wildcards differ depending on the program in use: awk derivatives (awk,sed,grep,ex,vi,expr and others) have the following special characters:

• ^ beginning of the line
• $ end of the line
• . any character
• * one or more of the preceding character
• .* any number of characters
• \n Carriage return
• \t Tab character
• \ Treat as is (so, \$ would try to match a '$')

Given the following four lines:

Chargeable calls in bundle: $47.50
Chargeable calls out of bundle: $20.50
Other bundle charges: $0.00
Total Charge: $20.50

$ grep "^Charg.*bundle.*\$.*"
would match the first two lines.
In english - match all lines which start with 'Charg', then have any number of characters and then the word 'bundle', then have any number of characters, and then a dollar symbol, and then have any number of characters following to the end of the line

BASIC FILE HANDLING

ls

- list files in directory; use with options

• -l (long format)
• -a (list . files too)
• -r (reverse order)
• -t (newest appears first)
• -d (do not go beyond current directory)
• -i (show inodes)

pg

- used to control input by pages - like the dos /p argument. pg performs the same function as the more command but has different control, as it is based on ex
Helpful keys for pg:

• 1 (go to top)
• $ (go to bottom)
• h (help)
• / (Search)
• ? (Search back)
• q (quit)
• -1 (back a page)

pwd

- show present working directory. e.g.
$ pwd
/usr/live/data/epx/vss2
To change the current working directory use cd

cd

- change directory (without arguments, this is the same as $ cd $HOME or $ cd ~)

cp

- copies a file from one location to another. e.g.
$ cp /etc/hosts /etc/hosts.backup # make a backup of the hosts file
$ cp /etc/motd /tmp/jon/ # Copy file /etc/motd to directory /tmp/jon/
Options

• -f (to force the copy to occur)
• -r (to recursively copy a directory)
• -p (to attempt to preserve permissions when copying)

synonym: copy

mv

- move a file from one location to another. e.g.
$ mv /tmp/jon/handycommands.txt . # move handycommands in /tmp/jon to current directory
$ mv -f vihelp vihelp.txt # Move file vihelp to vihelp.txt (forced) 
Options

• -f (to force the move to occur)
• -r (to recursively move a directory)
• -p (to attempt to preserve permissions when moving)

synonym: move
.

rm

- removes a file. e.g.
$ rm /tmp/jon/*.unl # remove all *.unl files in /tmp/jon
$ rm -r /tmp/jon/usr # remove all files recursively Options

• -f (to force the removal of the file)
• -r (to recursively remove a directory)

du

Recursively lists directories and their sizes. e.g.
$ du /etc # list recursively all directories off /etc
712 /etc/objrepos
64 /etc/security/audit
536 /etc/security
104 /etc/uucp
8 /etc/vg
232 /etc/lpp/diagnostics/data
240 /etc/lpp/diagnostics
248 /etc/lpp
16 /etc/aliasesDB
16 /etc/acct
8 /etc/ncs
8 /etc/sm
8 /etc/sm.bak
4384 /etc The sizes displayed are in 512K blocks. To view this in 1024K blocks use the option -k

lp -d

send file to printer. e.g. $ lp -dhplas14 /etc/motd # send file /etc/motd to printer hplas14
$ lp /etc/motd # send file /etc/motd to default printer

cat

- print a file to stdout (screen). e.g.
$ cat /etc/motd # display file /etc/motd to screen
*******************************************************************************
* *
* *
* Welcome to AIX Version 4.1! *
* *
* *
* Please see the README file in /usr/lpp/bos for information pertinent to *
* this release of the AIX Operating System. *
* *
* *
******************************************************************************* cat is also useful for concatenating several files. e.g.
$ cat fontfile IN* > newfile # appends fontfile and all files beginning with IN to newfileThough this might seem an essentially useless command, because most unix commands always take a filename argument, it does in fact come in extremely useful at more advanced levels. Awards are given out occasionally for the most useless usage of cat. If an option of '-' is specified, cat will take its input from stdin.