Frequently Asked Questions to AIX 3.x and IBM RS/6000 _____________________________________________________ This posting contains frequently asked questions and answers about the IBM RS/6000 series workstations and AIX version 3. All input is very welcome, I can be reached at jwarring@flAIXy.fd.amsinc.com The list is split into three articles to pass thru most mailers. I'll usually try to post them once a month to comp.unix.aix, news.answers, and comp.answers. Please let your input continue as I am most thankful for all of it. If you see a From: line it means that whatever follows is either an unabridged or slightly edited version of the input I have received, and that I may not have verified its contents. If there is no From: line, I probably know what I am talking about, and the entry is edited from various sources. All entries are numbered with major and minor subject number, e.g. 2.11. If the subject is preceded by an asterisk, that entry has been changed or added since the last posting. The comp.unix.aix group is for AIX on all IBM platforms -- RT, PS/2, 370, RS/6000 and mainframes (ESA based on OSF/1), but the traffic has evolved to discuss predominantly AIX 3.x and the RS/6000. The newsgroups comp.sys.ibm.pc.rt and comp.sys.ibm.ps2.hardware cover the RT, mostly hardware and AOS 4.3, and on PS/2 hardware respectively. Mark Whetzel posts FAQ lists about RTs in the rt and *.answers groups. There are few to non-existent discussions on AIX/370 and AIX/ESA. This article covers only AIX 3.x and the RS/6000, except when noted. If you post questions to comp.unix.aix, please be sure to indicate: - the machine type and brief configuration, e.g. Model 540, 64 MB RAM, 48 MB swap space (this is actually bad), 1.2 GB XYZ hard drive, etc. Note: There are now three classes of RS/6000s - the original, RS or POWER, RS/2 or POWER2, and PowerPC. - the exact AIX version number, i.e. AIX 3.1 is NOT sufficient, whereas AIX 3.1.5 or AIX 3.1 with the 3005 update is. With 3.2 you should mention any significant ptfs using the U4xxxxx numbers. I am doing this on my own time. Please do not ask me questions that should be asked to IBM. If you suspect you have software defect problems call (800) 237-5511. If you have hardware problems call (800) IBM-SERV. ______________________________________________________________________________ Table of contents: 1.0 General concepts 1.000 The AIX operating system - what is it? 1.001 I know neither Unix nor AIX - where do I find more information? 1.002 What is the Object Database? 1.003 How do I get rid of the verbose error messages? 1.004 Which release of AIX do I have? 1.1 SMIT, system administration, file systems 1.100 I am used to Unix systems programming, why should I learn SMIT? 1.101 How do I turn off the "running man" in smit? 1.102 How do I import an /etc/passwd or /etc/group file from another box? 1.103 Cleaning up utmp, who, and accounting problems 1.104 How to fsck the root filesystem 1.105 How can I unmount /usr to run fsck on it? 1.106* How do I shrink /usr? 1.107 How do I make a filesystem larger than 2 Gig? 1.108 How do I see/change system parameters like number of processes per user? 1.109 How do I shrink the default paging space on hd6? 1.110 The swapper seems to use enormous amounts of paging space, why? 1.111 How much paging space do I need? 1.112 How do I mount a floppy disk as a filesystem? 1.113 How do I remove a committed lpp? 1.114 How can I recover space after installing updates? 1.115 Where are the AIX log files kept? 1.116 How can I log information about ftp accesses to a file? 1.117 How do I find a file name from the inode number? 1.118 How do I set up postscript accounting? 1.119 How do I create boot diskettes for 3.2? 1.120 Where can I find tools for performance monitoring? 1.121 How can I tell what virtual printer a print queue is using? 1.122 Two srcmstr's are less useful than one? 1.123 How do I set the tty name associated with a physical port? 1.124 How do I use mksysb to clone a system? 1.125 How do I retain timestamps with mksysb? 1.126 How can I find out the machine type? 1.127 Updating to 3.2.5 1.128 AIX fix strategy 1.129 Are passwords limited to 8 char? 1.130 How do I increase the number of ptys > 64? 1.131 CERT advisories on sendmail and xterm 1.2 Backups, tape 1.200 Some info about tape backups 1.201 How do I do remote backup? 1.202 How do I backup a multi-disk volume group? 1.203 How do I put multiple backups on a single 8mm tape? 1.204 How can I make an exact duplicate of a tape over the network? 1.205 What is tape block size of 0? 1.206 Resetting a hung tape drive... 1.207 How do I read a mksysb tape with tar? 1.3 Memory and process management 1.300 Some info about the memory management system 1.301 How much should I trust the ps memory reports? 1.4 Shells, commands, man pages, InfoExplorer 1.400 How do I make an informative prompt in the shell? 1.401 How do I set up ksh for emacs mode command line editing? 1.402 Listing files with ls causes a core dump 1.403 How do I put my own text into InfoExplorer? 1.404 InfoExplorer ASCII key bindings 1.405 How can I add new man pages to the system? 1.5 Video, Graphics, X11 1.500 Which release of X11 do I have? 1.501 How to prevent ctrl-alt-backspace from killing the X session 1.502 Who has a termcap/terminfo source for aixterm or the HFT console? 1.503 How can I look at PostScript files? Why is "dpsexec" so lousy? 1.504 unix:0 vs `hostname`:0 1.505 VT100 key bindings for aixterm 1.506 Is there a screen saver that does not use excessive CPU? 1.6 Networks and communications 1.600 My named dies frequently, why? 1.601 How do I trace ethernet packets on an AIX system? 1.602 What is the authorized way of starting automount at boot time? 1.603 How do I set a tty port for both dial-in and dial-out? 1.604 How to move or copy whole directory trees across a network 1.605 How can I send mail to hosts that cannot be pinged? 1.606 How to configure dialup SLIP 1.607 Where is DCE discussed? 1.608 How do I make /var/spool/mail mountable? 1.609* getty spawning too rapidly 1.610* Does AIX support Compressed SLIP (CSLIP)? 1.7 1.8 1.9 Miscellaneous 1.900 SCSI-1 and SCSI-2 "interoperability" got you confused? 1.901 How to get your keyboard back after unplugging it from the 6000 1.902 How do I set up pcsim, the DOS emulator? 1.903 How do I transfer files between AIX and DOS disks? 2.00 C/C++ 2.01 I cannot make alloca work 2.02 How do I compile my BSD programs? 2.03 Isn't the linker different from what I am used to? 2.04 How do I link my program with a non-shared /lib/libc.a? 2.05 How do I make my own shared library? 2.06 Linking my program fails with strange error, why? 2.07 What's with malloc()? 2.08 Why does xlc complain about 'extern char *strcpy()' 2.09 Why do I get 'Parameter list cannot contain fewer ....' 2.10 Why does xlc complain about '(sometype *)somepointer = something' 2.11 Some more common errors 2.12 Can the compiler generate assembler code? 2.13 Curses 2.14 How do I speed up linking? 2.15 What is deadbeef? 2.16 How do I statically link in 3.2? 2.17 How do I make an export list from a library archive? 3.00 Fortran and other compilers 3.01 I have problems mixing Fortran and C code, why? 3.02 How do I statically bind Fortran libraries and dynamically bind C libraries? 3.03 How do I check if a number is NaN? 3.04 Some info sources on IEEE floating point 4.00 GNU and Public Domain software 4.01 How do I find PD software? 4.02* Are there any ftp sites? 4.03 General hints 4.04 GNU Emacs 4.05 gcc/gdb 4.06 GNU Ghostscript 4.07 TeX 4.08 Perl 4.09 X-Windows 4.10 bash 4.11 Elm 4.12 Oberon 2.2 4.13 Kermit 4.14 Gnu dbm 4.15 tcsh 4.16 Kyoto Common Lisp 4.17 Tcl/Tk 4.18 Expect 4.19 Public domain software on CD 4.20 Andrew Toolkit 4.21 sudo 4.22 Flexfax and other fax software 4.23 lsof 4.24* popper 5.00 Third party products 5.01 IBM List of third party products 5.02 Disk/Tape/SCSI 5.03 Memory 5.04 Others 5.05 C++ compilers 5.06 Memory leak detectors 5.07* PPP 6.00 Miscellaneous other stuff 6.01 Can I get support by e-mail? 6.02 List of useful faxes 6.03 List of 3.2 ptfs 6.04 Some RS232 hints 6.05 What publications are available for AIX and RS/6000? 6.06 Some acronyms 6.07 How do I get this by mailserver or ftp? 6.08 Hypertext version of the FAQ 6.09 Where can I send suggestions for tools? 6.10* copm.unix.aix archive availible on the WWW 7.00 Contributors ______________________________________________________________________________ 1.000: The AIX operating system - what is it? This is best answered by reading the text files in /usr/lpp/bos. The README file there contains general information and the BSD files contain useful information if you know BSD and/or System V. The last release for the RT PC is 2.2.1. The latest release for PS/2s and Intel architecture machines is AIX 1.3; for PS/2s only, 1.2.1. For the RS/6000, there are two major levels, 3.1 and 3.2, and various intermediate levels. IBM has dropped support for 3.1. 1.001: I know neither Unix nor AIX - where do I find more information? If you are new to Unix, you should look at the other newsgroups in the comp.unix hierarchy, in particular comp.unix.questions (for basic questions) or comp.unix.wizards (for difficult questions). These two groups have a common FAQ posting as well. If you need information about C programming, try comp.lang.c or comp.std.c, the latter for Standard ANSI C issues. comp.lang.c has a FAQ posting. 1.002: What is the Object Database? From: Uwe Geuder AIX stores most of the system management information in /etc/objrepos, /usr/lib/objrepos, and /usr/share/lib/objrepos. Files (also referred to as system object classes) in these directories are adminstered by the Object Database Manager, ODM, which is a set of library routines and programs providing basic object oriented database facilities. Under most circumstances, only SMIT or the commands SMIT call (see 1.100) should be used to change the contents of the system object classes. A harmless way to look at the object database is to use odmget where is one of the files in /etc/objrepos. Experienced users can use the ODM editor, odme, to navigate the database in detail. Modifying the database should only be attempted if you know exactly what you are doing. 1.003: How do I get rid of the verbose error messages? Many of the messages from the Unix commands are available in different languages. This is controlled by the LANG environment variable, the default being En_US meaning English in the US. All the default messages have a message number associated with them, e.g.: $ cat no-such-file cat: 0652-050 Cannot open no-such-file. If you prefer the terser Unix-looking error message, set your environment LANG to C, and you will get: $ cat no-such-file cat: Cannot open no-such-file. 1.004: Which release of AIX or other products do I have? New with 3.2.5: The oslevel command shows OS and component levels. Run oslevel -help to see options The command 'lslpp -h bos.obj' will show all lines referring to the BOS, Basic Operating System. E.g.: Fix Id Release Status Action Date Time User Name ------- --------------- --------- ---------- ---------- -------- --------- Path: /usr/lib/objrepos bos.obj 03.02.0000.0000 COMPLETE COMMIT 12/31/69 18:00:00 root U401864 03.02.0000.0000 COMPLETE COMMIT 11/12/92 20:09:35 root U401968 03.02.0000.0000 COMPLETE COMMIT 11/12/92 23:18:21 root U401969 03.02.0000.0000 COMPLETE COMMIT 11/12/92 23:18:20 root .......................... U418349 03.02.0000.0000 COMPLETE COMMIT 08/28/93 15:34:13 root U419950 03.02.0000.0000 COMPLETE COMMIT 08/28/93 15:34:11 root For AIX 3.2, you may come across discussions on 3.2.0, 3.2.1, 3.2.2, 3.2.3 extended, 3.2.4 and 3.2.5. There is no absolute way to tell which of these you are running since the newer releases are simply 3.2.0 with some sets of PTFs added. See above example listing. These selective fixes could mean there are literally thousands of slightly different variations of 3.2 in use. Please see section 6 to request some useful faxes to help with this number game. AIX 3.2.4 and later attempts to resolve this confusion. The OS is broken down into subsystems so that updates can be applied to an entire subsystem. lslpp also sports a new option; use 'lslpp -m bos.obj' to show what level and update the system is running. You can also use the new oslevel command. Unless you have a pressing need (such as applications unsupported on > 3.2.3), it is recommended that you update to this level. Also see 1.128. ______________________________________________________________________________ 1.100: I am used to Unix systems programming, why should I learn SMIT? 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. 1.101: How do I turn off the "running man" in smit? Use smitty, the standard curses version or add this line to your .kshrc file: alias smit="smit -C" 1.102: How do I import an /etc/passwd or /etc/group file from another box? Make sure that you run usrck, pwdck (/etc/passwd), and grpck (/etc/group) to let AIX work its password/group magic. 1.103: Cleaning up utmp, who, and accounting problems From: fritz@scipp.UCSC.EDU (Frederick Staats) This applies if you are running an X11R5 xterm on 3.2. Add this to the top of X11R5 mit/clients/xterm/main.c: #ifdef AIXV3 #define USE_SYSV_UTMP #define HAS_UTMP_UT_HOST #define WTMP_FILENAME "/var/adm/wtmp" #endif And your utmp problems should go away. If you want xterminal sessions to go into the wtmp file you need to define -DWTMP in the Imakefile and be sure the WTMP_FILENAME is set to the right place. The following program by jfh@rpp386.cactus.org (John F. Haugh) if run in the background by root will clean up the utmp file. #include #include #include main () { int fd; struct utmp utmp; while (1) { if ((fd = open ("/etc/utmp", O_RDWR)) < 0) exit (1); while (read (fd, &utmp, sizeof utmp) == sizeof utmp) { if (utmp.ut_type == USER_PROCESS && kill (utmp.ut_pid, 0) != 0) { lseek (fd, - (long) sizeof utmp, 1); utmp.ut_type = DEAD_PROCESS; write (fd, &utmp, sizeof utmp); } } close (fd); sleep (60); } } Another utmp program was posted to comp.sources.unix, volume 25, issue 96 by David W. Sanderson (dws@cs.wisc.edu) that also works on AIX 3.1. 1.104: How to fsck the root filesystem You can run fsck either in maintenance mode or on mounted filesystems. Try this: 1. boot from diskette 2. select maintenance mode 3. type /etc/continue hdisk0 exit (replace hdisk0 with boot disk if not hdisk0) 4. fsck /dev/hd4 1.105: How can I unmount /usr to run fsck on it? From: accapadi@mathew.austin.ibm.com (Matt Accapadi) [ This is for 3.2. ] 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". 1.106: How do I shrink /usr? From: mike@bria.UUCP (Michael Stefanik) and Richard Hasting FOR AIX 3.1 ----------- 1) Make a backup of /usr find /usr -print | backup -ivf /dev/rmt0 (or appropriate device) 2) shutdown to maintenance mode shutdown -Fm 3) export LANG=C 4) remove the filesystem and the logical volume ignore an error about the "dspmsg" command not found umount /usr rmfs /usr 5) make a new logical volume hd2 and place it on rootvg with desired size mklv -yhd2 -a'e' rootvg NNN where NNN is the number of 4 meg partitions 6) create a filesystem on /dev/hd2 crfs -vjfs -dhd2 -m'/usr' -Ayes -p'rw' 7) mount the new /usr filesystem and check it /etc/mount /usr df -v 8) restore from the tape; system won't reboot otherwise restore -xvf/dev/rmt0 9) Sync and reboot the system; you now have a smaller /usr filesystem FOR AIX 3.2 ----------- 0) Experiances posted to comp.unix.aix lead me to suggest that many administrators find the following piece of information useful after completing this procedure. I thought some of you might like to read it BEFORE getting yourself into this predicament. Call 1-800-IBM-4FAX and request document 2503 dated 1/26/94. Title is "How to recover if all files are owned by root after restoration from a mksysb tape". 1) Remove any unneeded files from /usr. 2) Make sure all filesystems in the root volume group are mounted. If not, they will not be included in the re-installed system. 3) Type mkszfile. This will create /.fs.size that contains a list of the active filesystems in the root volume group that will be included in the installation procedure. 4) Edit .fs.size. Change the size of /usr to what you want. Example: This .fs.size file shows /usr to be 40MB. rootvg 4 hd2 /usr 10 40 jfs The 10 is the number of physical partitions for the filesystem and the 40 is 40 MB. Most systems have a physical partition size of 4 MB. Therefore, the second number (40) will always be 4 times the previous number (10). Note, however, that a model 320 with a 120 MB drive will have a physical partition size of only 2 MB, and the total MB is twice the number of physical partitions. The first number (4) in the .fs.size file represents the PP size. If you want to reduce the size of /usr from 40 MB to 32 MB, edit the /usr entry to: rootvg 4 hd2 /usr 8 32 jfs IMPORTANT: Make sure that you DO NOT enter a value which is less than the size of the filesystem required to contain the current data. Doing so will cause the re-installation procedure to fail. 5) chdev -l rmt0 -a block=512 -T 6) Unmount all filesystems that are NOT in the root volume group. 7) Varyoff all user-defined volume groups, if any varyoffvg VGname 8) Export the user-defined volume groups, if any exportvg VGname 9) With a tape in the tape drive, type mksysb /dev/rmt0 This will do a complete system backup, which will include information (in the .fs.size file) for the installation procedure on how large the filesystems are to be created. 10) Follow the instructions in the Installation Kit under "How to Install and perform maintenance from Diskettes" (reportedly now called "BOS Installation from a System Backup") using the diskettes and tape that you created in the previous steps. DO NOT select the option "Reinstall AIX with Current System Settings". Instead use "Install AIX with Current System Settings" for the logical volume size changes to take affect. 11) When the installation is complete, you may then import any user-defined volume groups. importvg -y VGname PVname where "VGname" is the name of the volume group, and "PVname" is the name of any one of the physical volumes in the volume group. 12) Varyon your user-defined volume groups varyonvg VGname The reduction of the filesystems is now complete. 1.107: How do I make a filesystem larger than 2 Gig? The largest filesystem under AIX is 2 Gigabytes because the largest signed integer is 2**31 - 1. AIX 3.2.5+ can handle filesystems larger than 2 GB. 1.108: How do I see/change system 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) 1.109: How do I shrink the default paging space on hd6? 1) create a paging space to use temporarily mkps -s 20 -a rootvg 2) change default paging space hd6 so it is not used at next reboot chps -a n hd6 3) For AIX 3.1, edit /etc/rc.boot4 and change swapon /dev/hd6; for AIX 3.2, edit /sbin/rc.boot and change swapon /dev/hd6 swapon /dev/paging00 4) Update information in boot logical volume bosboot -a (3.1) bosboot -a -d hdisk0 (3.2) 5) shutdown and reboot 6) remove current hd6 and create a new one of smaller size rmps hd6 mklv -y hd6 -t paging rootvg 7) Re-edit /etc/rc.boot4 (3.1), /sbin/rc.boot (3.2) to swap to /dev/hd6 swapon /dev/hd6 8) Update information in boot logical volume bosboot -a (3.1) bosboot -a -d hdisk0 (3.2) 9) change current paging device (paging00) so it is inactive at next boot v chps -a n /dev/paging00 10) shutdown, reboot, remove paging00 using the command: rmps paging00 You can check your paging space with `lsps -a` 1.110: 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. 1.111: How much paging space do I need? See answer 2 in question 1.300. 1.112: How do I mount a floppy disk as a filesystem? From: op@holmes.acc.Virginia.EDU (Olaf Pors) You can build a filesystem on a floppy and mount it, however the filesystem will be read only. The reason that the filesystem will be read only is because AIX Version 3.1.5 cannot create a journal log on a diskette. The intended use is for temporary access to read only data. The diskette file system must be unmounted after use and during system backup procedures or errors could occur. To make the read only filesystem on a floppy: 1. Make a subdirectory on an existing filesystem and place all of the files that the diskette will contain into this subdirectory. 2. Enter the following command to create a prototype file containing information about the new filesystem, in the example /dir_struct is the pathname of the subdirectory created in step 1, and proto_filename is the name of the prototype file to be created. proto /dir_struct > proto_filename 3. Place a formatted floppy into the drive. 4. Edit the prototype file and replace the first line with the following: 0 0 5. Enter the following command to make the filesystem on your floppy: mkfs -p proto_filename -V jfs /dev/fd0 6. Create the directory upon which you will mount the floppy based filesystem, or you can use /mnt. Mount the filesystem: mount -r -V jfs /dev/fd0 /your_mount_point 7. To unmount the filesystem: umount /dev/fd0 Since the filesystem is read-only it may be of limited use but if you are going to use it for utility programs and other data that does not change much, it may still be useful. If you need to change the data, you can copy the directory from the floppy into another directory, make your modifications, and remake the filesystem using this procedure. 1.113: How do I remove a committed lpp? From: drb@chmeds.ac.nz (Ross Boswell) This seems to be a common problem. The following script for 3.2 only is >from IBM NZ AIX software support. As far as I know, it works -- I used it to remove unwanted fonts. I have no official permission to post it -- use it at your own risk! #!/bin/ksh # # Permanently remove a product from disk and AIX databases # if [ `whoami` != "root" ] ; then echo You must be root to run this script. exit 1 fi if [ `uname -a | awk '{print $1$4$3}'` != "AIX32" ] ; then echo This script only works on AIX 3.2. exit 1 fi TMP_FREE=`df /tmp | awk '$3 ~ /[0-9]/{print $3}'` if [ "$TMP_FREE" -lt 1000 ] ; then echo There is not enough room in your /tmp directory. echo You need 1000 KB free, and you have only $TMP_FREE KB free. echo Either remove some stuff from /tmp, or use chfs to make it bigger. exit 1 fi ODMDIRS="/etc/objrepos /usr/lib/objrepos /usr/share/lib/objrepos" ODMDIR=/usr/lib/objrepos export ODMDIR if [ $# -lt 1 ] then echo usage: $0 lppname [lppname ...] echo lppname is a string compatible with grep, ie "X11" or "PHIGS" echo typing $0 PHIGS will remove all LPPs with PHIGS in their name. exit 1 fi NAMES=$1 shift while [ $# -gt 0 ] ; do NAMES="$NAMES|$1" shift done echo "Searching for lpps with egrep \"$NAMES\"...\c" for ODMDIR in $ODMDIRS ; do if [ ! -d $ODMDIR -o ! -w $ODMDIR ] ; then echo $ODMDIR is not writeable or is not a directory. echo I hope this is because you are a /usr client or diskless. echo If you are not a diskless or a /usr client, you should stop. echo "Enter y to continue ->\c" read answer if [ "$answer" != "y" ] ; then exit 0 fi fi TMP=`odmget lpp | awk -F\" '/name/ {print $2}' | egrep "$NAMES"` LPPS=`echo $LPPS $TMP` done if [ "$LPPS" = "" ] then echo "failed.\nNo LPP with the name $NAMES detected." exit 1 fi echo ok. for ODMDIR in $ODMDIRS ; do mkdir -p /tmp/rmlpp/$ODMDIR > /dev/null 2>&1 done > /tmp/listOfFilesToRM$$ # truncate temporary file, just in case. echo echo This script is about to attempt to remove an LPP from your system. echo I say attempt, because it could fail. If it fails, you may have echo to at least reload the LPP. Use the \"lppchk\" command to make sure echo all is well with your system. echo # Loop through all the LPP names found. for LPP in $LPPS do DESCR=none answer="" # find the LPP ids. They will be different in the three SWVPD databases. for ODMDIR in $ODMDIRS ; do # get the lpp id for this ODMDIR (yes, they are different) LPPID=`odmget -q name=$LPP lpp | grep lpp_id | sed 's/.* = //'` # did we find the LPP? if [ "$DESCR" = "none" -a "$LPPID" != "" ] ; then # all the descriptions should be the same DESCR=`odmget -q name=$LPP lpp | grep description | sed 's/.* = //'` echo "Delete $LPP, $DESCR?" echo "y or (n) ->\c" read answer if [ "$answer" != "y" ] ; then # jump back up to the next LPP name continue 2 fi fi # if there is no DESCR, then we didn't find the LPP. Weird. if [ "$DESCR" = "none" -o "$LPPID" = "" ] ; then continue fi SOMETHING_DONE_FLAG=true # Optionally, save the ODM stuff we are about to remove, # in case something goes wrong. The problem is it is difficult # to determine if something really did fail, since these commands # don't return any decent error return codes. odmget -q lpp_id=$LPPID history > /tmp/rmlpp/$ODMDIR/$LPP.history odmget -q name=$LPP lpp > /tmp/rmlpp/$ODMDIR/$LPP.lpp odmget -q lpp_name=$LPP product > /tmp/rmlpp/$ODMDIR/$LPP.product # Get the list of files and links to remove later.... odmget -q lpp_id=$LPPID inventory > /tmp/rmlpp/$ODMDIR/$LPP.inventory awk -F\" '/loc/ {print $2}' /tmp/rmlpp/$ODMDIR/$LPP.inventory | \ sed 's/,/ /g' >> /tmp/listOfFilesToRM$$ odmdelete -o history -q lpp_id=$LPPID > /dev/null 2>&1 odmdelete -o lpp -q name=$LPP > /dev/null 2>&1 odmdelete -o product -q lpp_name=$LPP > /dev/null 2>&1 odmdelete -o inventory -q lpp_id=$LPPID > /dev/null 2>&1 done done if [ "$SOMETHING_DONE_FLAG" = "true" ] ; then echo ODM work is done. Now, time to delete files.... # This could be catastrophic if there is a problem. For example, # if the ODM database for an application had / as one of its files. # You be the judge. Here's your rope.... cat /tmp/listOfFilesToRM$$ | sort -r | uniq | xargs rm -rf rm -rf /tmp/listOfFilesToRM$$ echo done. fi # take this line out if you want to save your ODM saved files. rm -rf /tmp/rmlpp exit 0 1.114: How can I recover space after installing updates? From: Milt Cloud 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* 1.115: Where are the AIX log files kept? From: dirk@kimosabi.ucsc.edu (Dirk Coldewey) Such as the equivalent of the SunOS file '/var/log/syslog' and '/var/adm/messages'. If you want regular unix syslog files, you can configure syslogd in /etc/syslog.conf. Here's the one that I use: # # See the supplied /etc/syslog.conf file for copious comments. # *.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 * 1.116: How can I log information about ftp accesses to a file? From: elr@trintex.uucp (Ed Ravin) and map@hal.maths.monash.edu.au (Michael Page) 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". 1.117: How do I find a file name from the inode number? From: svpillay@berlioz.crs4.it (Kanthan Pillay) find /mntpoint -xdev -inum nnnn -print 1.118: How do I set up postscript accounting? From: taluskie@utpapa.ph.utexas.edu (Vince Taluskie) Ephraim Vider originally wrote this program to configure postscript page accounting. It acts as a backend wrapper which logs accounting information in /etc/qconfig and can be ftp'd from utpapa.ph.utexas.edu in /pub/aix/psacct.tar.Z. Compile with: cc pswrap.c -o pswrap -lqb and then make this program suid root: chown root pswrap chmod u+s pswrap If this step is not done, the printer will hang. Then start up SMIT and go to : Spooler Manage Local Printer Subsystem Local Printer Queue Devices Change / Show Characteristics of a Queue Device and change "BACKEND PROGRAM pathname" to the full pathname of pswrap since pswrap will now handle the backend interface with the queue. A stream of info will be written to a logfile in /tmp (prob lp0.log) but this file is mainly used for status info and raw pagecounts. If the accounting data is going to the qconfig-specified acctfile, then use the 'pac' command to read it. I prefer to have readable ascii data files instead, so I just comment out the '#define WANT_PAC' line and it will only go to the ACCTFILE specfied in pswrap.c. /* pswrap.c compile with: cc pswrap.c -o pswrap -lqb (for doc on the qb library see "understanding backend routines in libqb") BTW: The log_charge() function doesn't seem to do anything, but log_pages() updates the accounting info. Ephraim Vider, original author -- Feb 10, 1993 You can set pswrap up to use either the accounting file specified in /etc/qconfig (which means that you need to get your data from 'pac') or you can comment out the #define WANT_PAC line and then the accounting data will only go into the ACCTFILE. Also modified the logging to the ASCII acctfile so that it looks more readable. Vince Taluskie */ #include #include #include #include #include #include #include #include #include #include "qprocs.h" char *nextword (char *p); char *skipvalue (char *p); #define LOGDIR "/tmp" #define ACCTFILE "/usr/adm/acct/lpr/ps-acct" /* #define WANT_PAC 1 */ /* this define will also send accounting info to the acctfile specified in /etc/qconfig file. If this define is commented out then accounting info will only go to ACCTFILE */ char pcprog[] = "statusdict begin pagecount = end\n\x04"; char *keyw[] = { "idle", "busy", "waiting", "printing", "initializing", NULL }; enum { PS_IDLE, PS_BUSY, PS_WAIT, PS_PRINT, PS_INIT, PS_UNKNOWN }; void giveup(); jmp_buf jumper; char logfname[30]; FILE *logfile, *acctfile; main (argc, argv) int argc; char *argv[]; { char *devname; int pagcnt, c; int pid; int w, status; if (argc < 2) { fprintf(stderr, "Usage: psbe file\n"); exit(-1); } if (log_init(argv[1]) < 0) { fprintf(stderr, "log_init failed!\n"); exit(EXITBAD); } sprintf(logfname, "%s/%s.log", LOGDIR, get_device_name()); if ((logfile = fopen(logfname, "a")) == NULL) { fprintf(stderr, "Can't open logfile.\n"); exit(EXITBAD); } if ((acctfile = fopen(ACCTFILE, "a")) == NULL) { fprintf(stderr, "Can't open logfile.\n"); exit(EXITBAD); } setvbuf(logfile, NULL, _IOLBF, BUFSIZ); setvbuf(acctfile, NULL, _IOLBF, BUFSIZ); setvbuf(stdin, NULL, _IONBF, 0); setvbuf(stdout, NULL, _IONBF, 0); fprintf(logfile, "start Job no. %d, queued on %s\n", get_job_number(), get_qdate()); log_status(WAITING); pagcnt = getpagecnt(); log_status(RUNNING); if ((pid = fork()) < 0) { perror("fork"); exit(EXITBAD); } if (pid == 0) { argv[0] = "piobe"; execv("/usr/lpd/piobe", argv); perror("exec"); exit(EXITBAD); } while ((w = wait(&status)) != pid) if (w == -1) { perror("wait"); exit(EXITBAD); } if (WEXITSTATUS(status) != 0) exit(WEXITSTATUS(status)); log_status(WAITING); if (pagcnt > 0 && (c = getpagecnt()) > 0) { #ifdef WANT_PAC log_pages(c - pagcnt); #endif } fprintf(logfile, "end Job no. %d, queued on %s\n", get_job_number(), get_qdate()); /* the accounting file format is pages_printed user queue_printed_on time_queued */ fprintf(acctfile, "%d %35s %7s %s \n", (c - pagcnt), get_from(), get_queue_name(), get_qdate()); fclose(logfile); fclose(acctfile); exit(EXITOK); } void giveup () { longjmp(jumper, 1); } getpagecnt () { int pc = 0, pstat; char buf[81]; if (setjmp(jumper) != 0) { fprintf(logfile, "giving up on status\n"); return (0); } alarm(60 * 2); signal(SIGALRM, giveup); do { if (!gets(buf)) { sleep(5); putchar('\x14'); /* ^T returns status */ sleep(1); /* wait for answer from printer */ if (!gets(buf)) return (0); } fprintf(logfile, "%s\n", buf); if ((pstat = getstatus(buf)) == PS_WAIT) { putchar('\x04'); sleep(1); } } while (pstat != PS_IDLE); alarm(0); while (gets(buf)) fprintf(logfile, "%s\n", buf); printf("%s", pcprog); sleep(1); /* wait for answer from printer */ if (!gets(buf)) return (0); if (sscanf(buf, "%d", &pc) != 1) return (0); fprintf(logfile, "%d\n", pc); return (pc); } /* * Parser for printer status messages */ getstatus (p) char *p; { char *t; int i; if ((p = strchr(p, '%')) == NULL) return (PS_UNKNOWN); if (strncmp(p, "%%[", 3) != 0) return (PS_UNKNOWN); for (p = nextword(p + 3) ; p != NULL ; p = skipvalue(p)) { t = p; p = strchr(p, ':'); *p++ = '\0'; p = nextword(p); if (strcmp(t, "status") == 0) break; } if (p == NULL) return (PS_UNKNOWN); t = p; p = strchr(p, ' '); if (p[-1] == ';') p--; *p = '\0'; for (i = 0 ; keyw[i] != NULL ; i++) if (strcmp(t, keyw[i]) == 0) break; return (i); } char *nextword (p) char *p; { while (isspace(*p)) p++; if (strncmp(p, "]%%", 3) == 0) return (NULL); return (p); } char *skipvalue (p) char *p; { char *t; while (p != NULL) { p = strchr(p, ' '); t = p; p = nextword(p); if (t[-1] == ';') break; } return (p); } /********* qprocs.h ***********/ /* functions for communication between qdaemon and the backend */ char *get_from(); char *get_to(); char *get_qdate(); char *get_queue_name(); char *get_device_name(); char *get_title(); 1.119: How do I create boot diskettes for 3.2? You need to have four formatted diskettes. boot disk: bosboot -d /dev/fd0 -a display disk: mkdispdskt display extension disk: mkextdskt Install/Maintenance disk: mkinstdskt 1.120: Where can I find tools for performance monitoring? Free X based performance watcher: xsysstats - ftp site ftp.x.org:/contrib For 3.2 a few tools are available in /usr/lpp/bosperf. There are tools to monitor traces, I/O events, CPU, virtual memory, disk block usage, kernel extensions, etc. It even has a simulator, rmss, that allows one to try out different memory size configurations to see how it impacts performance. See 6.05 for the AIX Performance and Tuning Guide. 1.121: How can I tell what virtual printer a print queue is using? From: yoder@austin.ibm.com (Stuart R. Yoder) Use the command 'lsvirprt'. Don't use any parameters and it will run in an interactive mode that will give you a menu of all virtual printers on the system with the queue and device for each one. 1.122: Two srcmstr's are less useful than one? From: hubert@rs530.ncs.mainz.ibm.com (Bernhard Zeller) This can happen on systems that have no console, or systems with an async terminal as the console but not attached or turned off. One of the symptoms is a second srcmstr got run. But the second srcmstr is worthless as we can't use the stop/startsrc commands, refresh inetd, qdaemon won't start, etc. To resolve this, type: smit chgtty and add the keyword 'clocal' in following lines: STTY attributes for RUN TIME STTY attributes for LOGIN 1.123: How do I set the tty name associated with a physical port? From: accapadi@mathew.austin.ibm.com (Matt Accapadi) 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 1.124: How do I use mksysb to clone a system? From: joann@ariadne.SLAC.Stanford.EDU (Jo Ann Malina) 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. 1.125: How do I retain timestamps with mksysb? From: graeme@ccu1.aukuni.ac.nz ( Graeme Moffat) In /usr/lpp/bosinst/bosnet (for net installs), and bosrest (tape), the 'pax' commands all have '-pmop' options. m = "DON'T retain modification times". So, simply change all the '-pmop' to '-pop' and remake inst/maint diskettes or mksysb tapes. 1.126: How can I find out the machine type? From: umar@compsci.cas.vanderbilt.edu (Sait Umar) #!/bin/sh # # Author G. Vitillaro (peppe@ipgaix.unipg.it) # from an idea of Marc Pawliger (marc@ibmpa.awdpa.ibm.com) # # 10/12/92 GVT First (and last) implementation # # machine : will get machine type on RISC/6000 # # awk ' BEGIN { m["10"]="7013/530 or 7016/730"; m["11"]="7013/540"; m["14"]="7013/540"; m["18"]="7013/530H"; m["1C"]="7013/550"; m["20"]="7015/930"; m["2E"]="7015/950"; m["2E"]="7015/950E"; m["30"]="7013/520"; m["31"]="7012/320"; m["34"]="7013/520H"; m["35"]="7012/320H or 320E"; m["37"]="7012/340"; m["38"]="7012/350"; m["41"]="7011/220"; m["43"]="7011/M20"; m["46"]="250/PowerPC"; m["47"]="230/RSC"; m["5C"]="7013/560"; m["63"]="7015/970"; m["64"]="7015/980"; m["66"]="7015/580"; m["67"]="7015/570"; m["70"]="590/POWER2"; m["71"]="580H"; m["75"]="7012/370, 375, or 37T"; m["76"]="7012/360, 365, or 36T"; m["77"]="7012/350, 355, or 7013/550L"; m["78"]="7012/315 or 7013/510 "; m["80"]="990/POWER2"; You should request anyone with a model N40 run 'uname -m | cut -c9-10' so you can add an entry for it. if ( m[code] != "" ) print "This RISC/6000 is a " m[code]; else print "Unknown Machine"; } ' code=`uname -m | cut -c9-10` < /dev/null 1.127: Updating to 3.2.5 From: kraemerf@franvm3.VNET.IBM.COM (Frank Kraemer) (Ed. The following is useful if you have to update a large number of systems, large being > 3.) *** WARNING : Modify the scripts if you need dataless, diskless *** *** or remote /usr support. *** 1) Receive the PMP3250 tape from your AIX support center the PTF number is U493250. 2) Create a filesystem with 240 MB of space (60 PP's) and mount it as /dev/pmp3250 - /pub/pmp3250 3) Insert the tape (blocksize is 512) and install PTF U422467 # installp -BXacgq -d /dev/rmt0 bos.obj 3.2.0.0.U422467 4) Use the following script to load the tape in the new filesystem # cd /pub/pmp3250 # mktape2disk.sh 0 447 <<-- read 447 files from rmt0 =-=-=-=-=-=-=-= cut here =-=-=-=-=-=-=-=-=-=-= #!/bin/ksh # @(#) mktape2disk.sh creates files from tape on disk. # change name prefix here NAME="f" # from file #i to file #j integer i=1 integer j # test arguments if [ -z "${1}" ] then echo "\nusage: $(basename ${0}) <#files>\n" echo "\t: tape drive number (e.g. 0)" echo "\t<#files> : number of files to copy from the tape\n" exit 1 fi device=/dev/rmt${1}.1 # test arguments if [ -z "${2}" ] then echo "\nusage: $(basename ${0}) <#files>\n" echo "\t: tape drive number (e.g. 0)" echo "\t<#files> : number of files to copy from the tape\n" exit 1 fi j=${2} tctl -f ${device} rewind if [ $? -ne 0 ] then exit 1 fi # create tape while [ ${i} -le ${j} ] do echo "Copy file #${i} of #${j} from (${device}) to disk as (${NAME}${i})." dd if=${device} of="${NAME}${i}" bs=200k i=i+1 done tctl -f ${device} rewind exit 0 =-=-=-=-=-=-=-= cut here =-=-=-=-=-=-=-=-=-=-= 5) Create a new .toc file # cd /pub/pmp3250 # inutoc . # pg .toc 6) Place the following file named 'runme.sh' in the PMP directory =-=-=-=-=-=-=-= cut here =-=-=-=-=-=-=-=-=-=-= #!/bin/ksh # @(#) runme.sh for PMP3250 # INSTP="/usr/sbin/installp" LOG="/tmp/installp.log" TEE="/usr/bin/tee" PATCHDIR=$(pwd) # /usr/bin/cp /usr/lpp/info/data/ispaths /usr/lpp/info/data/ispaths.save # INFODIR="/usr/lpp/info/$LANG/aixmin" /usr/bin/mkdir ${INFODIR} 2>/dev/null >/dev/null if [ ! -w ${INFODIR} ] then print "\n\t*ERROR* Can not (write) access [${INFODIR}]." print "\tPlease unmount CD or NFS filesystems.\n" exit -1 fi # # Commit all ptf's # ${INSTP} -Xc all 2>&1 | ${TEE} ${LOG}.0 # # Install latest installp patch # ${INSTP} -BXacgq -d ${PATCHDIR} bos.obj 3.2.0.0.U422463 2>&1 | ${TEE} ${LOG}.1 # # Install latest installp patch # ${INSTP} -BXacgq -d ${PATCHDIR} bos.obj 3.2.0.0.U422467 2>&1 | ${TEE} ${LOG}.6 # # Run the ptfdir clean utility. # /usr/sbin/ptfdir_clean -y -f -v 2>&1 | ${TEE} ${LOG}.2 # # Install the PMP. Version 1 # /usr/lib/instl/sm_inst installp_cmd \ -T m -q -a -g -B \ -d ${PATCHDIR} \ -S '3250 AIX Maintenance Level U493250' \ -c -N -X \ 2>&1 | ${TEE} ${LOG}.3 # # Install the PMP. Version 2 # # /usr/sbin/update_all # # Install latest installp patch # /usr/bin/lppchk -v | ${TEE} ${LOG}.4 # # Show level of installp patch # /usr/bin/lslpp -m bos.obj | ${TEE} ${LOG}.5 # # Guess you should reboot now. # sync;sync print - "\n\n\tDone......guess you should reboot now !!\n" exit 0 =-=-=-=-=-=-=-= cut here =-=-=-=-=-=-=-=-=-=-= 7) To run the update on your server ===>> Please read the Installation Instructions first <<== # cd /pub/pmp3250 # ./runme.sh ........WAIT........... # /etc/shutdown -Fr 8) NFS export the PMP dir as READ-ONLY to all your clients. 9) To update a client system a - Mount the PMP filesystem from the server b - cd c - execute ./runme.sh d - Unmount PMP filesystem e - reboot 10) To create a copy 1:1 of the tape you received use the script 'mktape.sh' # cd /pub/pmp3250 # mktape.sh 0 =-=-=-=-=-=-=-= cut here =-=-=-=-=-=-=-=-=-=-= #!/bin/ksh # @(#) mktape.sh: creates a bootable tape from images on disk # change name prefix here NAME="f" # usage: mktape drive_no # drive_no = tape drive number # test arguments if [ -z "${1}" ] then echo "\nusage: mktape " echo " : tape drive number (e.g. 0)\n" exit 1 fi device=/dev/rmt${1}.1 tctl -f ${device} rewind if [ $? -ne 0 ] then exit 1 fi # create tape i=1 while test -r "${NAME}${i}" do echo "copy ${i}" if [ $i -le 3 ] then dd if="${NAME}${i}" of=${device} conv=sync bs=1k else dd if="${NAME}${i}" of=${device} bs=200k fi let i=${i}+1 done tctl -f ${device} rewind exit 0 1.128: AIX fix strategy From: leedp@austin.ibm.com (Dennis Lee, PMP Release Manager) First, a little history... The maintenance strategy for AIX 3.1 was cumulative updates. Every few months, we'd put all available fixes in one large package and ship it. There was no real strategy for providing a single fix. Although we'd occasionally produce an emergency patch, there was no method for tracking them; if you got a second one, it might overwrite the first. So, after a few of these patches, it's hard to track. In AIX 3.2 we introduced a "selective fix" strategy to support individual fixes. The package contained information about other fixes that were required for that fix to work correctly. For example, a Korn shell fix might require a change in libc.a, which might in turn require a fix in the kernel. This strategy allowed us to keep track of which fixes were installed to make sure we didn't overwrite one with another, and make sure they all worked together. But the initial selective fix design still had a few problems. o None of the fixes were cumulative. If you got a fix for Korn shell, you may not receive all of the fixes for Korn shell. This left the possibility of rediscovering other problems that were already fixed. o Since we chose to fix everything possible that was reported as a problem, instead of deferring them to the next release, the number of available fixes became quite large. o The number of additional fixes required by any given fix could also be quite large. Since the installation program ran once for each fix, the size and complexity of the fix packages grew, and installation time lengthens greatly. While developing the AIX 3.2.4 upgrade, we undertook a large effort to resolve the selective fix concerns, and dramatically increase the quality of AIX 3.2. The base operating system and most of the optional program products were split into subsystems. A subsystem is a group of logically related files. The division was made such that changes to a given subsystem were less likely to affect other subsystems. In total there are approximately 500 subsystems, but in practice, files have been modified in only about half of them. The advantages of the new packaging strategy are: o Each subsystem package is cumulative, containing all of the fixes and enhancements to date for that subsystem. o The cumulative subsystem package is tested as an entity. o The number of fix packages is greatly reduced because the number of subsystems is far fewer than the number of fixes and enhancements. o The number of other fixes required by any given fix is also greatly reduced because a subsystem package has requisites only on other subsystem packages. o The reduced number of fix packages greatly reduced installation time. Some customers also told us that they liked the maintenance level strategy that we used in AIX 3.1. They liked being able to install all of the known fixes, and they liked knowing what "level" of AIX they had. To meet these requirements, we produced a Preventive Maintenance Package (PMP). The PMP is simply a collection of the latest cumulative subsystem packages tied together in such a way that it can be installed by selecting a single fix. We also added flags to the lslpp command and added a new command, oslevel, to show which PMP is installed. Now we had both! The good attributes of selective fix along with the good attributes of maintenance levels. A few Q&As: Q. Why is the fix I just received 130 megabytes!@#? I already have the AIX 3.2.4 update installed! A. Your fix may be part of the AIX 3.2.5 update. AIX 3.2.5 is another PMP that contains all of the fixes to date, as well as enhancements to support the PowerPC model 250, and the new high-end RS/2 models 590 and 990, as well as support for new disk and tape drives, graphics adapters and more. Q. Why can't you just build my fix on 3.2.4? A. There really isn't such a thing as 3.2.1 or 3.2.2 or even 3.2.4. They're just collections of fixes and enhancements built on a 3.2 base. If the fix for your problem was built prior to 3.2.5, you can get the older version. But if your fix was built for the first time in a 3.2.5 subsystem, that's the only version of the fix that exists. See also 6.02. 1.129: Are passwords limited to 8 char? AIX passwords are only significant to 8 characters. You can set a passwd to more than 8 characters but anything over eight are ignored. No messages or warnings are given. Be careful if you're running NIS. You probably want to limit passwords to 8 char on all machines (6000 and others) to be compatible. Someone mentioned that DCE supports kerberos which supports passwords greater than 8 characters. 1.130: How do I increase the number of ptys > 64? From: mick@oahu.cern.ch (Mickey Coggins) SMIT only allows 64. Try this: odmget -q"attribute=num and uniquetype=pty/pty/pty" PdAt | sed "s/0-64/0-512/" | odmchange -q"attribute=num and uniquetype=pty/pty/pty" -o PdAt chdev -l pty0 -anum=256 -P reboot 1.131: CERT advisories on sendmail and xterm The following fixes are available for some CERT advisories on problems with sendmail and xterm. sendmail - fix available as U426396 xterm for X11R4 - fix available as U422575 xterm for X11R5 - fix available as U425811 1.200: Some info about tape backups From: Craig Anderson The following supplements the information on rmt devices in InfoExplorer. It is based on my own personal experience with IBM tape drives running on AIX 3.1. No warranty is expressed or implied. CONFIGURING THROUGH SMIT: BLOCK size (0=variable length) (ALL) Sets the tape block size. When reading, the block size must be set to the block size set when the tape was written. When using some commands, tapes written with ANY block size can be read if the block size is set to 0 (variable length) (see "BLOCK SIZES" below). Use DEVICE BUFFERS during writes (ALL) Set to yes, the device will buffer data internally on writes. This greatly improves performance, but under certain cases may be undesirable since the data is not written to tape before returning a good indication. Use EXTENDED file marks (8mm only) Extended file marks take up much more space than short (or non-extended) file marks. But extended file marks can be overwritten, allowing data not at the beginning of tape to be overwritten (see "FILE MARKS" below). RETENSION on tape change or reset (1/4" only) If set to "no" then the tape will not be retentioned automatically when the tape is inserted. Note that this will take effect only after the device is used. FILE MARKS: Tape devices support multiple tape files. Tape files are the result of a backup/cpio/tar/dd type command, where the device is opened, written to, and closed. Because tapes allow large quantities of data to be written on a single tape, several backups (that is, tape files), may be combined on one physical tape. Between each tape file is a "tape file mark" or simply "file mark". These file marks are used by the device driver to indicate where one tape file ends and another begins. B E <------- O O -------> T T __ ____________________________ _______________ physical | \ | | \ |physical beginning| \ | tape | \ | end of | \ | file | \ | of tape | \ | mark | \ | tape |_____\________|_______|__________\_________| Note that there is a distinction between the beginning of tape (BOT) side of a file mark and the end of tape (EOT) side of a file mark. If the head is on the BOT side of a file- mark, "tctl fsf 1" command will move only to the EOT side of the same file mark. With the 1/4" tape drive, writing can only take place sequentially, or after blank tape has been detected. You cannot write over data on the tape (except at BOT). If you wish to add data to a tape which has been written and then rewound you should space forward file mark until an error occurs. Only then can you start writing again. With an 8mm tape drive, writing can only take place before blank tape, an EXTENDED file mark, or at BOT. Thus if several backups have been made on one tape and you wish to overwrite one of the backups, position the tape to the place you wish to start writing and issue the following commands: tctl bsf 1 tctl eof 1 The first command skips back to the BOT side of the same file mark. The second command rewrites the file mark (writing is allowed before extended file marks). The erase head will erase data ahead of the write head, so that after writing the file mark the head will be positioned before blank tape. Only after this may you start writing over data in the middle of the tape. (All data beyond where you are currently writing will be lost). Note that you cannot write over short file marks. In order for this to work, the tape must have been written with extended file marks (use smit to change this). With the 9-track drive writing can take place anywhere on the tape although overwriting single blocks of data is not supported. On the 8mm drive extended filemarks use 2.2 megabytes of tape and can take up to 8.5 seconds to write. Short filemarks use 184K and take up to 1.5 seconds to write. BLOCK SIZES: When data is written to tape it is written in blocks. The blocks on a tape are separated by inter-record gaps. It is important to understand the structure of the written tape in order to understand the problems which can occur with changing block sizes. In fixed block size mode all blocks on the tape are the same size. They are the size of the block size set in the device configuration. All read()s and write()s to the tape drive must be a multiple of the fixed block size. In fixed block mode a read() will return as many blocks as needed to satisfy the read() request. If a file mark is encountered while reading the tape only the data up until the file mark will be returned. It is not possible for the tape drive to read a tape whose block size is not the same as the block size in the device configuration. (Unless the device configuration is in variable size blocks.) In variable block size (0) mode, the blocks written on the tape are the size of the read() and write() requests to the device driver. In this case, the actual block sizes on the tape can be changed using the options to the backup commands (tar -C, cpio -C, backup -C). In variable mode, read() requests greater than size of the block on the tape will return only the data from the next block on the tape. It is this feature that allows tapes written in any block size (fixed or variable) to read with the dd command (the output from the dd command may be piped to restore, tar, or cpio for example.) Note that backup, tar, and cpio cannot read all tapes by using a large block size because they assume there is an error if they get a short read(). dd ibs=128k obs=16k if=/dev/rmt0 | ... The tape head is always positioned at an inter-record gap, file mark, or blank tape after reading or writing. With the 8mm tape drive, using a fixed block size which is not a multiple of 1K is inefficient. The 8mm tape drive always writes internally in 1K blocks. It simulates the effect of variable block sizes, but, for example, using a fixed block size of 512 bytes (or using variable block size and write()ing 512 bytes at a time) wastes one half of the tape capacity and gives only one half the maximum transfer rate. EXCHANGING DATA WITH NON-UNIX AND OTHER VENDORS MACHINES: Many tape drives support both variable and fixed block sizes. Variable block mode writes block sizes the size of the write command issued (tar and backup specify this with the -b option). In fixed mode, block sizes are fixed and all writes must be a multiple of the fixed block size. Unix often internally chops larger reads and writes up into manageable pieces (often 65535, 65534, or 65532 bytes) before doing the actual reads and writes. This means reads and writes of 64K bytes are often broken up into a 65535 byte record and a 1 byte record (In fixed mode the write will fail). Block sizes >= 64K (-C128 and greater) should be avoided for this reason. AIX does not break up read and write requests, but be aware of the situation on other machines. If the tape is written in an unknown block size then set the device configuration in smit to use variable size blocks, use the "dd" command with a large input block size, and pipe it to the restore command. For example: chdev -l rmt0 -a block_size=0 dd if=/dev/rmt0 ibs=128k obs=16k | tar -tvf-