Unix Notes

You may want to take a peak at my Ignite document too.

• Boot to an ignite server after stopping the boot process:

  isl> hpux (;0)/boot/install
  

• To make a system image for ignite and place it into /tmp:

  make_sys_image -d /tmp -v -s local
  

• To make a recovery tape:

  ### Older version of Ignite
  make_recovery -vAC -d /dev/rmt/0mn
  
  ### Newer version of Ignite
  make_tape_recovery -v -I -x inc_entire=vg00 -a /dev/rmt/0mn
  
  Note:
  The tape recovery tool uses a no-rewind device file. It
  also will not preserve the mirrored disk configurations.
  You will need to recreate the mirror(s) and restore everything
  else from a backup.
  

• In order to send mail to uppercase users ("USER" for example). Add the "u" below:

  From: Mlocal,P=/usr/bin/rmail, F=lsDFMAw5:/|@qm9,  S=10/30, R=20/40,
  To:   Mlocal,P=/usr/bin/rmail, F=lsDFMAw5u:/|@qm9, S=10/30, R=20/40,
  

• How to make the root id change so when looked at from Outlook, pager, etc. instead of:

  root@batcave.gothamcity.com (userID@hostname.gothamcity.com), it now looks like
  root.batcave@gothamcity.com (userID.hostname@gothamcity.com)

  1. Uncomment the "O UserDatabaseSpec=/etc/mail/userdb.db" line in the /etc/mail/sendmail.cf file.
  2. vi /etc/mail/userdb and add:
     root:mailname root.batcave@gothamcity.com.
  3. Run this: makemap btree /etc/mail/userdb.db < /etc/mail/userdb.
  4. Restart sendmail: /sbin/init.d/sendmail stop;/sbin/init.d/sendmail start.
  5. Test by sending mail as root to somebody (like you).

• Check what version of sendmail your using:

  echo \$Z | /usr/sbin/sendmail -bt -d0

  or this

  sendmail -d0.1 < /dev/null | grep -i version

  or this

  echo '$=w' | sendmail -bt -d0.4 | grep Version

• Send email attachments:

  To send a plain text doc:

  mailx -m -s "Test attachments" root 1>/dev/null << EOF
  $(cat some_text_file | ux2dos - | uuencode stuff.txt)
  EOF
  

  To send a binary file:

  mailx -m -s "Test attachments" root 1>/dev/null << EOF
  $(cat some_binary_file | uuencode binary_file)
  EOF
  

  or

  uuencode [filename] [filename] | mailx -s "Subject" batman@batcave.com
  

• Spoof (sort of) your email address - this will show "name@bogus.com" sent the email.

  mailx -r "name@bogus.com" -s "subject" you@somewhere.com

• Build a new sendmail.cf file? Run this: /usr/newconfig/etc/mail/cf/cf/gen_cf

• Using mailx and the "at" command, send an email at a given time so you can get out of meetings. Compose an email message at 2.00pm, but want to send it to your pager at 5pm. First, make sure your user ID is in the file /var/adm/cron/at.allow, so that you are allowed to use "at". Then do something like:

  at 1700
  cat mail.txt | mailx -s "test mail" Dougs@pager.com ^D
  (The ^D is pressing Control-D to end the script).
  

• Additional sendmail info found in my Server Build document.

• 11.xx FTP configuration info found in my Server Build document.
• Find out what version your running: what `which ftp`

• Learn about NTP (time): http://www.ntp.org
• To check ntp: ntpq -p
• To do other checking: ntpdate -d
• Additional NTP configuration info found in my Server Build document.

• Use an NIS netgroup list to do stuff:

  for i in `netgrouplist net_group_list ; do something ; done

• Use the /etc/hosts.equiv:

  export -i -o rw=@net_group_list /users

• Add to /etc/passwd to let ONLY the admins in that are on this list:

  +@admin_access::-24:-24:::

Networkish stuff

• Use "lanscan" to get "NUM ID" (for 10.20) or "Crd In#" (for 11.x) to use on stuff below.

  lanadmin -x  0 = lan info (speed, half or full duplex, auto on or off)
  
  lanadmin -X 100HD   0 = turn lan0 to 100 speed, half duplex, auto on
  lanadmin -X 100FD   0 = turn lan0 to 100 speed, full duplex, auto off
  lanadmin -X AUTO_ON 0 = turn Auto Neg. on
  

• Check for the LAN driver being used. Find it's /etc/rc.config.d config file. The "btlan" driver config looks like this:

    HP_BASET_INTERFACE_NAME[0]=lan0 - the lan number
    HP_BASET_SPEED[0]=100FD         - the speed you want
  

  • Here's some more:

    driver   /etc/rc.config.d/file
    ------   ---------------------
    btlan0   hpeisabtconf
    btlan    hpbtlanconf
    btlan1   hpbasetconf
    btlan3   hpbase100conf
    btlan4   hpgsc100conf
    btlan5   hppci100conf
    btlan6   hpsppci100conf
    gelan    hpgelanconf
    

• Stop then start the cards: /sbin/init.d/net stop ; /sbin/init.d/net start
• Check loopback: linkloop -i 1 mac_address
• Verbose output: lanscan -v
• Status of network interfaces: ifconfig lan0 (or lan1, or lan2)
• IP interfaces: netstat -in
• Check the state of the FDDI card: /usr/sbin/fddistat /dev/lan0
• Check the FDDI network neighbors: /usr/sbin/fddinet /dev/lan0
• Check the ATM card: elstat -n ??? -v

  NOTE:
  "???" is the ATM card name gathered from the
  "netstat -i" command.
  

• Trouble with host lookup using DNS and NIS? Try adding this to the /etc/nsswitch.conf file:

  hosts: dns [NOTFOUND=continue UNAVAIL=continue TRYAGAIN=continue] nis [NOTFOUND=return UNAVAIL=continue] files[NOTFOUND=return UNAVAIL=continue TRYAGAIN=return]
  

• Run the following commands:

  ping XXX.XXX.XXX.255    example:(10.10.10.255)
  arp -a
  

• You will get a return that looks something like the following:

  "? (10.10.10.XX) at bi:g_:ma:c_:ad:dr:es:s_  ether"
  

• If you see an IP instead of a name, that means DNS (or the host file) can't resolve IP to name. You may have a spook on the line.

fsck -F vxfs -o full -y /dev/vg00/lvol6

• The "-B" option is used to create a bootable Physical Volume:

  pvcreate -B /dev/rdsk/c0t5d0
  

• Make sure to use the character device file when using mkboot:

  mkboot /dev/rdsk/c0t5d0
  

• Add diagnostics to lif:

  mkboot -b /usr/sbin/diag/lif/updatediaglif -p ISL -p AUTO -p HPUX  -p LABEL /dev/rdsk/c0t5d0
  

• Add boot info:

  mkboot -a "hpux -lq (;0)/stand/vmunix" /dev/rdsk/c0t5d0
  
  Note:
  The "-lq" indicates no quorum when two disks
  are used. When three or more are used you don't need "-lq."
  
  You also need to do this with the primary disk. Otherwise,
  when you boot normally to pri and the alt disk is not there
  for whatever reason, you will **NOT** be able to boot. The
  pri disk will "look" for the alt disk. No quorum????
  To bad so sad, you loose. Ack! So do this as well:
  
  mkboot -a "hpux -lq (;0)/stand/vmunix" /dev/rdsk/c0t6d0
  

• Adds the Physical Volume to the root volume group:

  vgextend /dev/vg00 /dev/dsk/c0t5d0
  

• Do this next command for each "lvol"[1-7]......

  lvextend -m 1 /dev/vg00/lvol1 /dev/dsk/c0t5d0
  

  ******* or *******

  for x in lvol1 lvol2 lvol3 lvol4 lvol5 lvol6 lvol7 swap2 Crash Logs
    do
      echo "\n******** Doing ${x} ********\n"
      lvextend -m 1 /dev/vg00/${x} /dev/dsk/c0t5d0
  done
  

• Don't forget any extra swap you added (like /dev/vg00/swap2). Do "swapinfo" and "bdf" to check for swap and any other lvols that may have been created.
• Specifies the root logical volume:

  lvlnboot -r /dev/vg00/lvol3
  

• Specifies the swap logical volume:

  lvlnboot -s /dev/vg00/lvol2
  

• Specifies the dump logical volume:

  lvlnboot -d /dev/vg00/lvol2
  

• Specifies the boot logical volume:

  lvlnboot -b /dev/vg00/lvol1
  

• Recovers any BDRA info:

  lvlnboot -R
  

• Verify boot, root, swap and dump settings:

  lvlnboot -v
  

• Display the Primary and Alternate boot paths that are currently set:

  setboot
  (You can also see what drive you booted off of: echo boot_string/S|adb /stand/vmunix /dev/mem)
  

• Change the Alternate Boot Path to the path of the Root Mirror:

  setboot -a 10/0.5.0
  
  Note:
  Use "ioscan -funC disk" to find out what the
  path should be. Run "setboot" again to confirm
  the change.
  

• Do "shutdown -r 0" (or "reboot") to test. After halting the boot process, type in "bo alt" to boot from the alternate disk. Also, you may want to *remove* the primary disk to test as well. The server should boot from the alternate disk if the primary is not found and no quorum was set.

• Adds alternate links then mirrors the logical volume:

  vgextend /dev/vg00 /dev/dsk/c?d?t?
  lvextend -m 1 /dev/vg00/lvol1 /dev/dsk/c?d?t?
  

• Remove the mirror from a logical volume:

  lvreduce -m 0 /dev/vg00/lvol1
  

• Remove the disk from the volume group:

  vgreduce /dev/vg00 /dev/dsk/c?t?d?
  

• Increase the target logical volume:
  • With OnlineJFS:

    fsadm -F hfs (or vxfs) -b new_size /mounted_filesystem
    

  • Without OnlineJFS:

    Unmount filesystem
    lvextend -L 400 /dev/vg00/lvol3
    extendfs /dev/vg00/rlvol3
    Mount filesystem
    

• Reduce the target logical volume. Works great if you have no more disk space but have extra in /tmp. Reduce /tmp and put your new found disk space into (lvextend) /var.

  NOTE:
  You may loose data on the selected Logical Volume.
  

  • With OnlineJFS:

    fsadm -F hfs (or vxfs) -b new_size /mounted_filesystem
    

  • Without OnlineJFS:

    Unmount filesystem
    lvreduce -L 100 /dev/vg00/lvol5
    newfs -F hfs /dev/vg00/rlvol5
    Mount filesystem
    

Let's say you need to reinstall the O/S on a server and retain the logical volume(s) and there volume group(s) that hold your data, or you want to move a set of disk that contain data and configured volume groups to a different server.

• Change directories to a safe place on your original server:

  cd /tmp
  

• Tar the volume group(s) "group" file.

  tar cvf group.tar /dev/vg01/group
  

• Make the volume group(s) inactive.

  vgchange -a n vg01
  

• Export the beast and make files needed to rebuild the volume group on the "new" server. When you run this command it will wipe out the /dev/vg01 directory. (Remember, your still in the /tmp directory.)

  vgexport -v -m vg01.map -f vg01.disk vg01
  

• You then need to move the group.tar and the vg01.* files to your "new" server or some other server while you rebuild this one. If you don't create the files needed for the vgimport command (see below), then the command, when run, will simply (re)name every logical volume you have starting with lvol1 through lvol### (the number depends on how many lvols you have). Also, if for some reason you don't save the group file, you'll need to recreate one. Use this next command to make a group file.

  mknod /dev/vg01/group c 64 0x010000
  

• At this point your disks have been moved to your target server or the server was rebuilt. In either case you need the group.tar and the vg01.*files on this "new" server. Untar the group.tar file then run the import command.

  vgimport -v -m vg01.map -f vg01.disk vg01
  

• Now make the volume group(s) active.

  vgchange -a y vg01
  

• Down ALL applications first, then:
  • Move out the old (bad?) lvmtab:

    mv /etc/lvmtab /etc/lvmtab.old
    

  • Check the lvmtab file:

    vgscan -v -p
    

  • Generate the lvmtab file:

    vgscan -a
    

  • Activate all volume groups:

    vgchange -a
    

  • Correct boot information on disk:

    lvlnboot -R
    

• Restore LVM volume group configuration from backup file:

  vgcfgrestore -n /dev/vg00 -o /dev/rdsk/c?t?d?
  

• Activates volume group:

  vgchange -a -y  /dev/vg00 /dev/rdsk/c?t?d?
  

• Synchronize stale logical volume mirrors in LVM volume groups:

  vgsync /dev/vg00 /dev/rdsk/c?t?d?
  

• Synchronize stale mirrors in LVM logical volumes:

  lvsync /dev/vg00/stuff
  

• Display statistics

  fsadm -F hfs (or vxfs) /dev/vg00/rlvol6
  

• To change it without having OnlineJFS (you may need to do this in "single user" mode):

  Umount the filesystem
  
  fsadm -F hfs (or vxfs) -o largefiles /dev/vg00/rlvol6
  

• To change it with OnlineJFS:

  fsadm -F hfs (or vxfs) -o largefiles /mounted_filesystem
  

• Make a snapshot in **seconds** using OnlineJFS:

  lvcreate -L 100 -n lvsnap vg00    <-- 10-20% of original (snapped) filesystem
  
  mkdir /test
  
  mount -F vxfs -o snapof=/home /dev/vg00/lvsnap /test
  
  # bdf /test
  Filesystem          kbytes    used   avail %used Mounted on
  /dev/vg00/lvsnap   1024000   56968  959480    6% /test
  
  # bdf /home
  Filesystem          kbytes    used   avail %used Mounted on
  /dev/vg00/lvol4    1024000   56968  959584    6% /home
  
  umount /test when done (backed up)
  

  And maybe do this if needed:

  lvremove -f /dev/vg00/lvsnap
  

• Defrag with OnlineJFS:

  Check a dir to see if we should defrag it:   fsadm -F vxfs -D -E /home
  

  To defrag it:  fsadm -F vxfs -d -e /home
  

• Use the "-e" option to re-install device files:

  insf -e
  

dd if=/dev/rdsk/c?t?d? of=/dev/null bs=4096k

• Run this to wipe your drives (write "nothing" to the drive):

  dd if=/dev/null of=/dev/rdsk/c?t?d? bs=1024k
  

• or better yet, write "0" to the drive:

  dd if=/dev/zero of=/dev/dsk/c?t?d? bs=1024k
  

  • If you don't have /dev/zero, you can make it (11.11 (11i) already has /dev/zero):

    mknod /dev/zero c 3 0x000003 #...on 10.20 
    mknod /dev/zero c 3 0x000004 #...on 11.x 
    chown bin:bin /dev/zero 
    chmod 666 /dev/zero 
    

• or....

  mediainit -v /dev/rdsk/c?t?d?
  

• Add "/dev/dsk/c?t?d? /SD_CDROM pfs-rrip xlat=unix 0 0" to the /etc/pfs_fstab file.

    c?t?d? = The CD device
    SD_CDROM = Mount point
  

• Next, run "nohup /usr/sbin/pfs_mountd"
• Then "nohup /usr/sbin/pfsd""

  Note:
  THE ORDER IS VERY IMPORTANT - pfs_mountd -
  MUST BE STARTED FIRST!
  

• Check to be sure pfs started all the correct processes, use the following command. You should see the output below if everything is correct:

  # ps -ef | grep pfs
  
  root 17456 17455 0 15:15:24 ttyp1 0:00 pfsd.rpc
  root 17453 17452 0 15:15:18 ttyp1 0:00 pfs_mountd.rpc
  root 17455 16772 0 15:15:23 ttyp1 0:00 pfsd
  root 17452 16772 0 15:15:18 ttyp1 0:00 pfs_mountd
  root 17475 16772 2 15:15:32 ttyp1 0:00 grep pfs
  

• Now mount the cd using the following syntax, were cXtXd0 is the CD-ROM device file on your machine.
  • If the /etc/pfs_fstab was made:

    pfs_mount /SD-CDROM
    

  • To mount the CD in Rock Ridge format

    pfs_mount -o xlat=unix /dev/dsk/c?t?d0 /SD_CDROM
    

  • If /etc/pfs_fstab was not made:

    pfs_mount -o xlat=unix /dev/rdsk/c?t?d0 /SD_CDROM
    

• Only use pfs_umount to dismount the CD-ROM:

  pfs_umount /SD_CDROM
  
  Note:
  A reboot will be required if at anytime the umount
  command is issued on the pfs mounted CD-ROM instead
  of using the pfs_umount command.
  

• On your local server with the CD drive:
  1. nohup /usr/sbin/pfs_mountd &
  2. nohup /usr/sbin/pfsd &
  3. Add the filesystem to /etc/pfs_fstab
  4. Add the filesystem to /etc/pfs_exports
  5. pfs_exportfs -a
  6. pfs_mount /cdrom

• On your target (remote) server:
1. nohup /usr/sbin/pfs_mountd &
2. nohup /usr/sbin/pfsd &
3. pfs_mount -o xlat=unix local_system:/cdrom /local_mnt_pnt

• On your local server with the CD drive:

  mount /dev/dsk/c?t?d? /cdrom
  exportfs -i -o ro /cdrom
  

• On your target (remote) server:

  mount -r hostname_of_server_above:/cdrom /cdrom
  

• To install a patch:
  • Un-shar the file:

    sh /tmp/Patch_Name
    

  • Now install it:

    swinstall -x autoreboot=false match_target=true -s /tmp/patch.depot
    

• To add depots to an 11.00 server using the GUI:

  swcopy
  

• Without the GUI (from the command line):

  swcopy -v -s /tmp/patch_dir/PH*.depot @ /tmp/patch_dir
  

• To swcopy a depot of 10.20 patches onto an 11.00 server using the GUI.

  swcopy -x layout_version=0.8 -i
  

• To register the software for network installs:

  swreg -l depot -v /PATH/of/depot
  

• Check filesets installed:

  swlist -l fileset -a state
  

  or to find anything NOT configured type:

  swlist -l fileset -a state | grep -v -E "^#|configured"
  

• To complete the installation of the software:

  swinstall \*
  

• To verify all products and dependency files:

  swverify \*
  

• To cleanup (remove) old patches:

  cleanup       (for 10.20)
  cleanup -c 1  (for 11.x)
  

• Creating a separate software depot for your patches (on another server?) is useful, when space is tight and patches have dependencies.
  • Assume you have two patches - PHNE_20000 and PHNE_20001.
  • Unshare the patches by typing the following commands.

    sh PHNE_20000 
    sh PHNE_20001 
    

  • Next, we create a new software depot in the directory /newdir/depot

    mkdir /newdir/depot 
    swpackage -s PHNE_20000.depot -d /newdir/depot 
    

  • Register your new depot.

    swreg -l depot  /newdir/depot 
    

  • Copy the next patch to the newly created software depot.

    swcopy -s `pwd`/PHNE_20001.depot PHNE_20001 @  /newdir/depot 
    

  • Install your patches.

    swinstall -s /newdir/depot 
            

• To find out what depots you have on your server:
  

  cat /var/adm/sw/host_object | grep path | awk '{print $2}'
  

• To unregister the depot, type:
  

  swreg -u -l depot /path/to/depot
  

• Make a new kernel in SAM but do not "put into place". Then do:

  mv /stand/system /stand/system.prev
  mv /stand/build/system.SAM /stand/system
  /usr/sbin/mk_kernel
  reboot
  
  NOTE:
  Ok... so I don't really know why you would want to
  do this. I've always let SAM put the kernel into
  place and rebooted.
  

• Make a kernel

  Create the files needed to build a new kernel using the current kernel parameters:

      cd /stand/build
      /usr/lbin/sysadm/system_prep -v -s system
  

  Run "kmtune" to see the current kernel parameters.

  Change a kernel parameter (let's say shmseg is currently set to 120 and we want it set to 128):

      kmtune -s shmseg=128
  
         or add a driver to the kernel:
  
      /usr/sbin/kmsystem -S /stand/build/system -c Y driver-name
  

  Build the new kernel:

      /usr/sbin/mk_kernel -s /stand/build/system
  

  Perform a kernel update (moves /stand/vmunix to /stand/vmunix.prev and puts the new kernel into place as /stand/vmunix):

      /usr/sbin/kmupdate
  

  Move the previous kernel just in case the new one is bad:

      mv /stand/vmunix.prev /stand/vmunix.bkup
  

  Reboot the box: reboot or shutdown -r

  Uh-Oh - If you've made a bad kernel, boot to a good one from the ISL prompt (see "Other Kernel stuff" below for other ways to boot to a new kernel):

      hpux (;0)/stand/vmunix.bkup
  

  Remove the bad kernel (this assumes you've booted to the "bkup" kernel):

      kmupdate -d /stand/vmunix
  
      NOTE:
      This will remove /stand/vmunix as we know it.
      Can you say Buh-Bye?
  

  Rebuild the kernel:
  You've booted to the backed-up version which was good. The box is up and running but now doesn't have a standard issue kernel to boot from. Remember the "kmupdate -d /stand/vmunix" you ran above? You now need to create a "new" kernel so go back to the top of this section and rebuild the kernel.

      NOTE:
      Remember NOT to overwrite the known good
      kernel - /stand/vmunix.bkup
  

• Boot from desired kernel while in ISL mode:
  • Boot normally from the previous kernel:

    hpux (;0)/stand/vmunix.prev
    

  • Boot in "single user" mode from the previous kernel:

    hpux -is (;0)/stand/vmunix.prev
    

  • Boot in "lvm maintenance" mode from the previous kernel:

    hpux -lm (;0)/stand/vmunix.prev
    
    
    Note:
    To boot from some other kernel, replace the kernel
    file name "vmunix.prev" with the name you gave your
    alternate kernel.
    

• The kernel files are:
  • A file containing the kernel itself:

    /stand/vmunix
    

  • A file containing the kernel configuration:

    /stand/system
    

  • A directory containing a few files and directories for kernel configuration purposes (for HP-UX 11.0 and up only)

    /stand/dklm/*
    

• To check the current kernel parameters:

  kmtune -l   or   sysdef
  

NOTE:
The old selections for DataBase server kernel changes are not
available in 11i. To have these again, download the following
files and place them in the /usr/sam/lib/kc/tuned directory.
Then start up SAM and select the one you need.

General OLTP/Database Client System      oltp_cli.tun
General OLTP/Database Monolithic System  oltp_sa.tun
General OLTP/Database Server System      oltp_ser.tun

• Use the "system_prep" command to extract the system file from the current kernel.

  cd /stand/build
  system_prep -v -s system 
  

• Edit the /stand/build/system file to perform your task or use "kmsystem" and "kmtune" to make changes. Build the new kernel:

  mk_kernel -s system 
  

• Move the old system file and kernel so if anything goes wrong, you still have a bootable kernel:

  mv /stand/system /stand/system.prev
  

• Update the kernel with the newly generated kernel. Automatically saves the old versions of vmunix and dlkm in /stand:

  kmupdate /stand/build/vmunix_test
  

• Reboot the system.

* Configurable IPC Message Parameters:
      mesg      Enable/disable IPC messages (Series 700 only)
         Minimum:    0 (Exclude System V IPC message parameters from
                        kernel)
         Maximum:    1 (Include System V IPC message parameters in kernel)
         Default:    1
      msgmap    message free-space map size
         Minimum:    3
         Maximum:    Memory limited
         Default:    msgtql+2
      msgmax    maximum message size
         Minimum:    0
         Maximum:    65535 bytes
         Default:    8192 bytes
      msgmnb    maximum bytes in message queue
         Minimum:    0
         Maximum:    65535 bytes
         Default:    16384 bytes
      msgmni    maximum message queues on system
         Minimum:    1
         Maximum:    Memory limited
         Default:    50
      msgseg     number of segments in message queue
         Minimum:    1
         Maximum:    32767
         Default:    2048
      msgssz     message segment size
         Minimum:    1
         Maximum:    Memory limited
         Default:    8 bytes
      msgtql     maximum total messages on system
         Minimum:    1
         Maximum:    Memory limited
         Default:    40

* Configurable IPC Semaphore Parameters:
      sema      Enable/disable semaphores (Series 700 only)
         Minimum:    0 (exclude System V IPC semaphore code from kernel)
         Maximum:    1 (include System V IPC semaphore code in kernel)
         Default:    1
      semaem    Semaphore value-change limit
         Minimum:    0
         Maximum:    semvmx or 32767, whichever is smaller
         Default:    16384
      semmap    Size of free-semaphore resource map
         Minimum:    4
         Maximum:    Memory limited
         Default:    semmni+2
      semmni    Maximum semaphores system-wide
         Minimum:    2
         Maximum:    Memory limited
         Default:    64
      semmns    Maximum user-accessible semaphores system-wide
         Minimum:    2
         Maximum:    Memory limited
         Default:    128
      semmnu    Maximum undos per semaphore
         Minimum:    1
         Maximum:    nproc-4
         Default:    30
      semume    Maximum semaphore undos per process
         Minimum:    1
         Maximum:    semmns
         Default:    10
      semvmx    Maximum allowed semaphore value
         Minimum:    1
         Maximum:    65535
         Default:    32767

* Configurable IPC Shared Memory Parameters:
      shmem     Enable/disable shared memory (Series 700 only)
         Minimum:    0 (exclude System V IPC shared memory code from
                        kernel)
         Maximum:    1 (include System V IPC shared memory code in kernel)
         Default:    1
      shmmax    Maximum shared memory segment size
         Minimum:    2 Kbytes
         Maximum:    1792 Mbytes
         Default:    0x04000000 (64 Mbytes)
      shmmni    Maximum segments on system
         Minimum:    3
         Maximum:    (memory limited)
         Default:    200 identifiers
      shmseg    Maximum segments per process
         Minimum:    1
         Maximum:    shmmni
         Default:    120

• What made the stinkin core in the first place:

  file core
  

• To get the libs used:

  what core
  

• To debug the stinker:

  gdb -c core
  

• To debug the stinker:

  gdb ProgramName -c core
  

• Also see HP doc - KNC071499001

• Get info: swapinfo -a
• Opens entire system to swap: swapon /

• Find 777 (rwxrwxrwx) files.

  find / \( -perm 000 \)  -exec ls -ld {} \;
  

• Find set-owner-id and set-group-id files:

  find / \( -perm -4000 -o -perm -2000 \) -type f -exec ls -ld {} \;
  

• Another way to find SUID and SGID stuff:

  for stuff in `awk '($3 ~ /^(hfs|vxfs)$/) { print $2 }' /etc/fstab`
    do
        find "$stuff" \( -perm -04000 -o -perm -02000 \) -type f -xdev -print
  done
  

• Find core files (not directories):

  find / -type f -name core -exec rm {} \;
  

• Find uppercase files, i.e.; Uppercase or UPPERcase

  find . -xdev -type f -name "[[:upper:]]*"
  

• Find files that a specific user owns.

  find / \( -user oracle \) -type f -exec ls -ld {} \;
  

• Find then change files (and possibly the current directory) from one user name to a new user ID:

  find . -user user_name | xargs chown new_user_name
  

• Grep "fsck" (or any other named item) in all files and list out the path to that file:

  find / -name "*" -exec grep -l -i fsck {} \;
  

• Find files that are 1,000,000 (1 meg) or larger.

  find /var -size +1000000c -exec ll {} \;
  

• Find files that are 100 meg or larger and sort them.

  find . -size +100000000c -exec ls -l {} \; | sort -n -k5,5
  

• Try to delete a lot of files in /var/tmp (or wherever) and getting "The parameter list is too long". Do a find to get rid of the stuff.

  find /var/tmp . -print -exec rm {} \;
  

  or do this to "rm" them:

  echo * | xargs rm
  

• Copy a directory from one location to another. Even if it has "special" files and links in it:

  cd /usr
  find . -print | cpio -dlmpuvx /usr2
    d = Do directories
    l = Do links
    m = Keep time
    p = Pass along
    u = Copy unconditionally
    v = verbose
    x = Save device special files
  

• Backup files and/or directories:

  fbackup -0 -f /dev/rmt/c?t?d?BEST -I /tmp/index -v
  
    0 = Or "1" - (0 = full, 1 = incremental)
    f = Device file to use
    I = Directory to place index
    v = Verbose
  
  frecover -f /dev/rmt/c?d?t?BEST -v -I /tmp
  
  mt -t /dev/rmt/c?t?d?BEST rew
    rew   = Rewind tape.
    offl  = Rewind tape and go off line.
    fsf 5 = Go forward five setmarks on the tape.
    bss 2 = Go backwards two setmarks on the tape.
  
  tar cvf /path/file_name (or device_file_name) /path/target_to_copy
    c = Copy
    x = Restore
    v = Verbose
    t = Index
    r = Add to the end of list
  

• To copy files to a tarball and gzip at the same time.

  tar cvf - * | gzip -9> filename.tar.gz
  
  NOTE:
  Change the "*" to whatever filename or directory you chose or
  you can use the "*" to grab everything in your current directory.
  

• Robotarm stuff:

  uma -ioctl /dev/rac/c?t?d?
  
  **** or ****
  
  /usr/sbin/mc -p /dev/rac/c?t?d? -s D1 -d S1 -r IDSM
      s = Source (the drive in the library).
      d = Tape source
      r = Print out status
  

• To get version info: omnidb -version
• Stop OmniBack (or do "/sbin/init.d/omni stop"): omnisv.sh stop
• Start OmniBack (or do "/sbin/init.d/omni start"): omnisv.sh start
• List all sessions in the database: omnidb -session
• List database space usage: omnidbutil -info
• Best info on specific database files: omnidbutil -extendinfo
• Add an "fvers.dat" file with it's maximum size of 2 gig. Total of 4 "fvers.dat" files can be made.

  omnidbutil -extend /var/opt/omni/db/cdb -maxsize 2048
  

• Add an "fnames.dat" file with it's maximum size of 2 gig. Total of 2 "fnames.dat" files can be made.

  omnidbutil -extendfilenames /var/opt/omni/db/cdb -maxsize 2048
  

• To clean up the database a bit:
  1. Delete items from the OmniBack "Monitor" GUI window
  2. Verify that no OmniBack sessions are running
  3. Strip the database using "omnidb -strip"
  4. More cleaning using "omnidbutil -clear"
  5. Do a purge using "omnidbutil -purge -detail -days ?"

     purge  = Purge stuff
     detail = Out of date file versions and names are removed
     days   = Data older then "?" days are removed
     

• Copy the OmniBack database to another filesystem:
  1. mkdir /new_omni_database_location
  2. Create the new filesystem and mount it.
  3. cd /var/opt/omni/db
  4. find . -xdev -depth -print | cpio -pxdm /new_omni_database_location
  5. ln -s /var/opt/omni/db /new_omni_database_location

• The syntax for the exclude option in /etc/opt/omni/datalists/filename_related_to_a_server is:

  /path         for Unix (/blah/de/blah)
  Drive:\path   for NT (c:\blah\de\blah)
  Volume:\path  for NetWare (sys:\blah\de\blah)
  

• For Host backup, the syntax used is:

  /path         for Unix (/blah/de/blah)
  Drive:\path   for NT (c:\blah\de\blah)
  Volume:\path  for NetWare (sys:\blah\de\blah)
  

• Stuff placed in the /etc/opt/omni/NOTIFICATION file:

  NOTE:
  The item - "backup" - shown below is an email alias.
  It can be changed to any email address or alias.
  
  NOTIFICATION "Device Error" 
  { -event "DeviceError" 
  -object "*" 
  -email "backup" 
  
  NOTIFICATION "Tape Mount Request" 
  { -event "MountRequest" 
  -object "*" 
  -email "backup" 
  } 
  
  NOTIFICATION "DB Space Low" 
  { -event "DbSpaceLow" 
  -object "*" 
  -email "backup" 
  } 
  
  NOTIFICATION "Backup Finished" 
  { -event "EndOfSession" 
  -object "*" 
  -email "backup" 
  }
  

• Check Veritas Netbackup process:

  /usr/openv/netbackup/bin/bpps -a
  

• Start Veritas GUI:

  /usr/openv/netbackup/bin/xnb
  

• Check if drives are up:

  /usr/openv/volmgr/bin/tpconfig -d 
  

• Pending requests:

  /usr/openv/volmgr/bin/vmoprcmd
  

• Bring a drive up:

  /usr/openv/volmgr/bin/vmoprcmd -up 0 (0 = drive number)
  

• Run the robot arm (like UMA):

  /usr/openv/volmgr/bin/robtest
  

• Tape GUI:

  /usr/openv/volmgr/bin/vxadm
  

• If EMS has found something bad. Your getting hammered with email. However, it's ok for the time being that the part is bad. Configure EMS not to notify you about the bad part.

  To stop the EMS email run EMS (or look at your email) and add the EMS line to this file - /var/stm/data/tools/monitor/disabled_instances

  Lines may look like this:

  /adapters/events/FC_adapter/8_0.8
  /adapters/events/FC_adapter/8_4.0
  /storage/events/disks/default/10_0.6.0

  1. Run EMS: /etc/opt/resmon/lbin/moncheck
  2. K)ill monitoring (Stop here to simply turn EMS off)
  3. In another window add the line to: /var/stm/data/tools/monitor/disabled_instances
  4. E)nable Monitoring
  5. Do a C)heck monitors just for fun

• Getting this at bootup:

  "Corrupted tracing and logging header
        Internal error (210)"
  
  The fix, run:
  
    /usr/sbin/nettl stop
    rm /usr/adm/nettl.log00
    /usr/sbin/nettl start
  

• You open an xterm and hit "Enter" a couple of times and the line doesn't move up. If you make the stinkin window smaller it will start to scroll up the way it should. Well.... run this next time:

  eval `resize`
  

• All hardware.

  echo 'selclass qualifier all;info;wait;infolog' | /usr/sbin/cstm
  

• Memory.

  echo "selclass qualifier memory;info;wait;infolog"|cstm
  

• To count CPUs (Processors), you can either use adb, ioscan, or cstm (if you have OnLineDiags installed).

  echo "processor_count/D" | adb /stand/vmunix /dev/kmem
  
  echo "ioscan -FkC processor | wc -l"
  
  echo "map\nquit\nok\n" | cstm | grep CPU | wc –l
  

• To report speed of CPU:

  echo itick_per_usec/D | adb -k /stand/vmunix /dev/mem
  

• Get type of CPU:

  pdcinfo -no_banner|grep PA|grep -v PATH|grep -v PAR|grep -v HPA 
  

• Drive info every 30 sec, for 20 times:

  vmstat -dnS 30 20
  

• Same as vmstat:

  iostat 30 20
  

• Have "sar" collect data every 30 sec, 20 times:

  sar -A -o /tmp/sar_info 30 20
  

• Read the data from a given file (normally in /var/adm/sa/sa##):

  sar -A -f /tmp/sar_info
  

• Additional SAR info can be found on my Server Build doc.

ipcs -mqsabcopt

32 or 64 bit supported server

• Check /etc/.supported_bits:

  grep `/usr/bin/model` /etc/.supported_bits
  

• See what an 11.x server is running:

  getconf KERNEL_BITS or chatr /stand/vmunix
  

• To see what version has been loaded:

  swlist -l bundle | grep Runtime
  

• Change or add files in /etc/skel. When you add a new user they get all the new/configured files.
• Copy the /etc/skel files for a new user:

  cp /etc/skel/.[A-z]* /home/new_user
  

• List all added users and there comments found in the /etc/passwd file.

  listusers
  

• Add a user.

  useradd -u 101 -g 200 -c "Doug Burton" -d /home/dburton -s /usr/bin/ksh -m -e 08/01/2020 dburton
    -u = UID
    -g = GID
    -c = Comments
    -d = Users home directory
    -s = Users shell
    -m = Creates the home directory
    -e = Password expiration date
  

• I'm not recommending this due to possible security reasons, but this is pretty cool. Put this little script into ROOTs .profile and create the link. What this does is let root use YOUR user .profile to set $PATH and other stuff.

  # custom homedir for root users
  UID=`id | cut -d'=' -f2 | cut -d'(' -f1`
  [ "$UID" = "0" ] && {
      USER=`who am i | cut -d' ' -f1`
      echo "Setting up root environment for $USER"
      [ -L /home/root.$USER ] && {
         export HOME=/home/root.$USER
         . ~/.profile
      }
  }
  

  The link: ln -s /home/dburton /home/root.dburton

• Put an alias in roots profile (to speed up things).

  alias ra='/usr/lbin/modprpw -l -k'
  alias ga='/usr/lbin/getprpw -m lockout'
  

  then type "ra account_name" to reset the account then type "ga account_name" to ensure it is complete. You should see this: lockout=0000000

• Configure the .profile, .kshrc, .exrc and .logout files (See my Server Build page).

• For 9000/700 series:

  b p0 ipl
  isl>hpux -is boot disc(;0)/stand/vmunix
  

  • Boot to an alternate kernel in single user mode:

    isl>hpux -is boot disc(;0)/stand/vmunix.prev
    

• For 9000/800 series
  • Search for any bootable devices (LAN, tape, CD or drive).

    sea ipl
    

  • The "?" is the p number given for the bootable item you want to use (after doing your search).

    bo p? ipl    or    bo pri ipl
    

  • Select "YES" if asked to interact with IPL.

    hpux -is (;0)/stand/vmunix    or    "hpux -is"
    

• Run some command forever until you ctrl-c the thing:

  while (true) ; do command ; sleep 2 ; done
  

• Check out a few drives, one right after another:

  for x in 1 2 3 ; do pvdisplay "/dev/dsk/c3t1d${i}" ; done | more
  

• To set the system date (month=mm, day=dd, hour=hh, minute=mm, year=yy):

  date -u mmddhhmmyy
  

set_parms hostname
set_parms timezone
set_parms date_time
set_parms root_passwd
set_parms ip_address
set_parms addl_netwrk

• Change the "/sbin/sh" line for root in the /etc/passwd file to "/sbin/ksh" then do this:

     cp -p /bin/ksh /sbin/ksh
  
  NOTE:
  If you do a "file /bin/ksh" (or /usr/bin/ksh), you'll notice that this shell is
  dynamically linked. In theory, you can't use this shell in single user mode due
  to the needed libs not being mounted. I don't know why, but I have used it with
  no adverse results. You need to test in your environment. In other words,
  use at your own risk.
  
  ADDITIONAL NOTE:
  Just to seal the deal, HP does not support the use of root using ksh at bootup.
  

• Check the /etc/passwd and /etc/group files

    pwck  and  grpck
  

mknod /dev/null c 3 0x000002
chown bin:bin /dev/null
chmod 666 /dev/null

Run "catman" to build the "/usr/share/lib/whatis" file.

Then you can run the "man -k [some command] to see what
manual sections the command is in. For example, if you run
"man -k mkdir", it will show that "mkdir" is in
sections 1 and 2.

To see section 2 you would run "man 2 mkdir".

• Look for big files (see my "find" commands on this page). Also look here:

  /var/adm/crash
  /var/adm/syslog
  /var/tmp
  /tmp
  /var/adm/lp
  /var/mail
  /var/adm/wtmp (to big?)
  /dev/rmt/om (instead of /dev/rmt/0m)
  /var/adm/lp/log
  /var/spool/lp/request/
  /var/spool/cron/tmp
  /var/sam/log
  /var/spool/mqueue
  /usr/etc/yp
  lost+found (found in any mounted file system)
  Did you mount a filesystem *ON* another filesystem (thus hiding what's underneath)
  Remove core files
  Extra kernels saved in /stand
  

1. Open "sam ->Printers and Plotters->LP Spooler-> Save/Restore Spooler Configuration"
2. Go to "Actions"
3. Click on 'Save Spooler Configuration'
4. Exit out of SAM
5. tar cvf /tmp/sam.tar /var/sam/lp
6. Transfer the sam.tar file to system B:/tmp.

1. tar xvf /tmp/sam.tar
2. You may need to change the permissions and ownership of these files. To do that run:
   chmod -R 755 /var/sam/lp/* ; chown -R lp:bin /var/sam/lp/*
3. Open "sam ->Printers and Plotters->LP Spooler-> Save/Restore Spooler Configuration"
4. Go to "Actions"
5. Click on 'Restore Spooler Configuration'
6. Don't forget to restart the scheduler
7. Exit out of SAM

• Log off the terminal before making these changes to the terminal configuration. Then perform the following steps on a 700/96 terminal:
  1. Press the user/system key to get to the configuration screen.
  2. Press f8 (config keys).
  3. Press f5 (terminal config).
  4. Tab to the Terminal Id field.
  5. Type VT100 in this field.
  6. Tab to the TermMode field.
  7. Press f2 (next choice), EM100 appears in the field.
  8. Press f1 (save config).
  9. Use the f8 (config keys) to go to the ansi config screen.
  10. Press f6 (ansi config).
  11. Tab to the backspace Def field.
  12. Press f2 (next choice).
  13. Verify the Backspace Def field now reads BackSpace/Del in the field.
  14. Tab to the EM100 ID field.
  15. Press f2 (next choice) to change the EM100 ID field to EM100.
  16. Press f1 (save config).

• The terminal should now display EM100 below the softkeys.
• This procedure enforced the following settings:

  * terminal id to vt100,
  * termmode to em100,
  * correctly mapped the backspace and delete keys
  * em100 id to EM100
  

• The terminal is now ready to correctly operate in vt100 mode.

• To get Hitachi controller info:

  ioscan -fun
  

• Check the timeout, should be 60:

  pvdisplay /dev/dsk/c?t?d? | grep Timeout
  

• Change the timeout value to 60:

  pvchange -t 60 /dev/dsk/c?t?d?
  

• Check the queue depth, should be 2:

  scsictl -a /dev/rdsk/c?t?d?
  

• Change it to the value of 2:

  scsictl -m queue_depth=2 /dev/rdsk/c?t?d?
  

How many floppies are in a Gigabyte, Terabyte, Petabyte, Exabyte and a Zettabyte
	Gigabyte	Terabyte	Petabyte	Exabyte	Zettabyte
How many MB in a:	1,024	1,048,576	1,073,741,824	1,099,511,627,776	1,125,899,906,842,620
How many 1.44 MB floppies are in a:	711.11	728,177.78	745,654,044.44	763,549,741,511.11	781,874,935,307,378.00
How many feet of floppies are in a:	207.41	212,385.19	217,482,429.63	222,702,007,940.74	228,046,856,131,318.00
How many miles of floppies are in a:	0.04	40.22	41,189.85	42,178,410.59	43,190,692,449.11
	7/11 down the street	St. Petersburg, FL to Plant City, FL	About 5.19 times around the earth	Distance from the Earth to Mars	More than ten times the farthest distance of Pluto from the Sun

• Stuff #1
  • Open a window and run:

    script -a /tmp/STUFF.`date +%Y%m%d`
    

  • Open another window and "tail -f /tmp/STUFF.`date +%Y%m%d`" the file above.
  • Go to the first window and do "ll" or some such thing. EVERYTHING typed AND displayed is in the target file. Even SAM gets displayed correctly if you do a tail -f. If you checkout the file later (cat -v for example), SAM output looks weird. Not sure if you can "really" see SAM window(s) after the fact.
  • How about putting this into a users .profile?

• Stuff #2
  • Restoring Default File Permissions/Ownership:

    Somebody did a "chmod -R 777 /usr" for some stupid reason. Now you need to restore all the files to whatever perms and ownership they originally had.

    I assume your on the server in at least init level 2. Do this chmod so we
    can start fixing this mess.
    
    /sbin/chmod ugo+rx /usr/bin/cut /usr/bin/expand /usr/bin/grep /usr/sbin/swlist   
    
    Then this: comp="/usr"
        (or this if there is more than one directory messed up)
               comp=' -e ^/usr  -e ^/var -e ^/sbin/init.d'
    
    And away we go...
     
       swlist -l file -a type -a path -a mode -a owner -a group | expand | grep \
       $comp | cut -f 2 -d : | /sbin/awk '{if(length($2) < 2 && $4 > 0)         \
       print "/sbin/chmod "$4,$3";/sbin/chown " $5":"$6,$3}' | /sbin/sh
    

Support number: 1-800-USA-4SUN (800-872-4786)

• Get motif version info:

  /usr/ccs/bin/what /usr/dt/lib/lubXm.s0
  

• Free up /var... (Blow away "save" files):

  cd /var/sadm ; find . -name save
  

• To change the openwin logo to our (tif) logo, copy our file to:

  /usr/openwin/server/etc/Solaris.im8.Z
  

• Patch location:

  /var/sadm/patch
  

• Other worth while commands:

  scsiinfo -p
  boot -r
  sysdef -d
  devinfo
  prtconf
  proctool    - Like Glance on HP
  probe-scsi  - Like "ioscan" on HP
  admintool   - Like SAM on HP
  

  Check the Rosetta Stone for other commands similar to HP commands.

Nothing yet........

Ok.. so I put this waaay down at the bottom of this web page. I don't expect anyone to actually toss a buck or two my way as a thank you for all the work I've done on this site (Hmmm... feel guilty yet?), but it sure would be nice.