This is the final 3.1.x HSM patch.

Article:TECH8618  |  Created: 2000-01-02  |  Updated: 2001-01-25  |  Article URL http://www.symantec.com/docs/TECH8618
Article Type
Technical Solution


Environment

Issue



This is the final 3.1.x HSM patch.

Solution





This patch includes the following HSM patches:

patch P0800012 - solaris, solaris_dm, solaris_26, solaris_dm_26, hp_dm
  Corrects a problem where migreconstruct will not build files
  or directories with characters like ", ` , <space> , etc in the name.

patch P0800024 - solaris, solaris_dm, solaris_dm_26
  Corrects a memory leak problem in migcopy.
  The more files migcopy makes copies of the more memory
  migcopy needs.

patch P0800030 - solaris_dm, solaris_dm_26, hp_dm
  Fixes a problem where migcopy hangs waiting for a source
  volume that is locked by another completed migcopy
  that was writing to the volume.

  This is only a problem on DMAPI versions of HSM.

patch P0800035 - solaris, solaris_dm, solaris_26, solaris_dm_26, sgi_dm
  Fixes tar to prevent core dumps when restoring to
  HSM managed filesystems.
  This problem only occurs when tar trys to issue and error to HSM
  log files.

   This patch updates the NetBackup tar command.

patch P0800037 - sgi_dm
   Fixes migout so that, when migration fails because
   DMAPI attributes cannot be set, managed regions are not left on the file.

patch P0800040 - auspex, hp_dm, sgi_dm, solaris, solaris_dm,
                 solaris_26, solaris_dm_26
   Corrects a problem with the HSM nb method.
   If HSM is making two copies and one of the copies is the nb method,
   HSM may only make a copy of the inode and not the data.
   This patch replaces the NetBackup bpbkar command.

patch P0800046 - sgi_dm
 On IRIX platforms, HSM is unable to migrate files when the filesystem
 is 100% full because there is no space available to set DMAPI attributes.
 With their patch 3037, SGI provided a mechanism to reserve filesystem
 space for use in this situation.  This patch takes advantage of the
 mechanism provided by SGI patch 3037.

 NOTE:  SGI patch 3037 MUST be installed before this patch is installed;
 if this code is installed without the SGI patch, there is the potential
 for kernel panics/crashes.

patch F0800047 - solaris_dm_26
 Change migstage command to restore files from nb method
 volumes by giving a list of these files to bpbrestore so
 only one bpbrestore is executed.

  This will help reduce netbackup mountings and positioning
 of volumes.


  NOTE: If migstage is used to cache a file from an nb volume
       migstage will wait until the data is restored before
       migstage completes. To get around this run migstage
       in the background.

  This patch replaces the NetBackup tar command.

patch P0800049 - solaris
   This patch prevents a system deadlock problem in mig.c with the
   Solaris 2.5.1 OS. This patch is for HSM 3.1.1 only and Solaris 2.5.1.
   This patch introduces a new mutex (mig_pollock) which is used to
   synchronize/serialize process doing a poll wait (select system call)
   and a process trying to wake up a polling process, instead of using
   the mutex mig_monitor. With this code HSM process do not hold the
   poll head lock over long term IO requests. The new mutex is acquired
   with a mutex_tryenter() sys call which may return EAGAIN to migd if
   the mutex is not available.

patch P0800051 - solaris
 Fixes intermittent cache failures on Solaris caused by a Device Busy
 condition on a tape drive after positioning using SCSI locate block.

  Corrects a problem where migstage leaves files
 in migration/data if there is a Busy error
 when a file is cached.

patch P0800056 - solaris_26
  When upgrading to solaris 26 kernel version the .PAIN file
  must be converted to a 26 kernel format.

   The 3.1.1 Solaris_26 CD is missing the routines needed to
  convert the .PAIN file.

  NOTE: the pfconvert in patchP0800056.solaris_26.tar.Z has
        been changed to handle .PAIN files that become
        larger then 2Gbytes.

patch P0800059 - solaris_dm_26
  If 2 copies are being made to the same method type and there
 is only one drive for the method only one of the copies is made.

  The making of the 2nd copy (migcopy) does a deadman timeout
 never making its copy.

  If migrc -R <hsmname> is now run the 2nd copy will be made.

patch P0800061 - sgi_dm. solaris_dm, solaris_dm_26
  When upgrading to HSM 3.1 the <hsmname>.IHAND files
  must be converted to a HSM 3.1 format.

   The 3.1.1 solaris_dm_26, solaris_dm, and sgi_dm packages are
  missing the routines needed to convert the <hsmname>.IHAND files.

patch P0800067 - solaris, solaris_26
       Solaris panics when files in the premigration directory
       /<fs>/migration/data are (pre)migrated and cached. Files
       in premigration directory will be cached if migcopy is
       run to make copie(s) to secondary storage media after the
       migrate command is used to migrate the premigration file.
       If this condition exists, then this patch must be installed
       to prevent system panic.
       

       With this patch attempts to cache files under this condition
       will be denied with errno set to EINVAL. This error is
       reported to the user as, "Protocol family not supported" or
       EPFNOSUPPORT. Some Unix utilities, such as tail, may even
       ignore this error condition. In any case, before files can be
       cached when this has happened, the system adiminstrator should
       make the files in premigration "regular" files. See option #2
       below (WORKAROUND) on how to do this. Care should be taken to
       prevent loss of data when doing this.

       If this patch documentation is unclear,
       please call VERITAS support for help.

Workaround:
       Options available:

       1. Do not migrate files in the premigration directory.

       2. If the files in premigration are migrated, then before
       accessing any migrated data file cache all files in premigration
       directory with a request to update the files.

       Example:

       Assume user file /<fs>/<somedir>/foo is a migrated file with
       file handle 100e. The corresponding file to this in premigration
       would be 3e8M100e, assuming the machine ID is 3e8. If this file
       is also premigrated, then cache 3e8M100e and update it. Or cache
      this file and then change the flags field in the .PAIN file
      entry to 0. This can be done with the following command:

      pfprint -f 0 /<fs> inode#

      Where, inode # is the inode number of 3e8M100e.

patch P0800068 - auspex, hp_dm, sgi_dm, solaris, solaris_dm,
                 solaris_26, solaris_dm_26
  HSM is unable to cache some files from optical platters.
  The problem is caused by an incorrect compression of the FHDB.
  The FHDB entry for a problem file is created incorrectly if
  during the making of a copy on the op platter the size of
  the granule header increases to a size greater the 512 bytes.

   This may happen if the path of a problem file is arround 74 bytes long.

   This patch can not correct an existing FHDB but will prevent further
  incorrect FHDB entries.

   To correct the FHDB for a problem file do the following:

    1) run migopscan -n on all the volumes the problem file
      resides on. If this patch has been installed. migopscan can be run
      without using the -n option.

    2) backup the FHDB, this just a safeguard.

    3) replace existing FHDB op entries for the problem file with
      the FHDB.<volume> entries for the problem file.

    4) the problem file will now be able to be cached..

 
  A problem file can have one of the following errors in the HSM log file
  when a cache fails.

    1)
09/09 09:44:18 migin[27031]: starting to reload file 3e8M101b
09/09 09:44:18 migin[27031]: exec /usr/openv/hsm/bin/admincmd/migcopy
09/09 09:44:19 migcopy[27032]: Starting
09/09 09:44:19 migcopy[27032]: exec /usr/openv/volmgr/bin/tpreq
09/09 09:44:32 migcopy[27032]: INFO: Mount DSC22A READ ONLY on /usr/var/openv/hsm/workdir/test2.1012.R
09/09 09:44:39 migcopy[27032]: WARNING: crc error in source. Handle 3e8M101b. FHDB crc 37972 != 7599 Computed
09/09 09:44:39 migcopy[27032]: ERROR: No valid granule found.
09/09 09:44:39 migcopy[27032]: INFO: Unmount vol=4114, delay=180, force=0, mount point=/usr/var/openv/hsm/workdir/test2.1012.R
09/09 09:44:39 migcopy[27032]: ERROR: Failed to copy x101b
09/09 09:44:39 migcopy[27032]: ERROR: copy_for_method() ret=1
09/09 09:44:39 migcopy[27032]: Finished 1
09/09 09:44:39 migin[27031]: ERROR reload of 3e8M101b failed, orig_owner=root path=/test2/acg/1234567890/1234567890/1234567890/1234567890/1234567890/A1234567

    2)
09/09 09:45:56 migin[27056]: starting to reload file 3e8M102f
09/09 09:45:56 migin[27056]: exec /usr/openv/hsm/bin/admincmd/migcopy
09/09 09:45:57 migcopy[27057]: Starting
09/09 09:45:57 migcopy[27057]: INFO: Mount DSC22A READ ONLY on /usr/var/openv/hsm/workdir/test2.1012.R
09/09 09:46:32 migcopy[27057]: INFO: Unmount vol=4114, delay=180, force=0, mount point=/usr/var/openv/hsm/workdir/test2.1012.R
09/09 09:46:32 migcopy[27057]: exec /usr/openv/volmgr/bin/tpreq
09/09 09:47:13 migcopy[27057]: INFO: Mount DSC22B READ ONLY on /usr/var/openv/hsm/workdir/test2.1013.R
09/09 09:47:19 migcopy[27057]: ERROR: zero length source granule header len= 0
09/09 09:47:19 migcopy[27057]: ERROR: No valid granule found.
09/09 09:47:19 migcopy[27057]: INFO: Unmount vol=4115, delay=180, force=0, mount point=/usr/var/openv/hsm/workdir/test2.1013.R
09/09 09:47:19 migcopy[27057]: ERROR: Failed to copy x102f
09/09 09:47:19 migcopy[27057]: ERROR: copy_for_method() ret=1
09/09 09:47:19 migcopy[27057]: Finished 1
09/09 09:47:19 migin[27056]: ERROR reload of 3e8M102f failed, orig_owner=root path=/test2/acg/1234567890/1234567890/1234567890/1234567890/1234567890/T1234567


patch P0800069 - sgi_dm, solaris_dm, solaris_dm_26
migsweep always invokes readdir() instead of selectively invoking
readdir64() when the latter is appropriate.  This causes it to not
select "large inode" files for migration.

patch P0800071 - solaris_dm_26
 The HSM migattr driver may cause a panic on
 VxFS 3.2.5 running HSM 3.1.1 or
 VxFS 3.2.2 running HSM 3.1.1.

patch P0800077 - solaris, solaris_26
       Solaris panics and the stack frame trace shows
       mig_makemigrated. The problem is caused by
       mig_makemigrated using too much stack frame space.

Additional Notes:
   This patch replaces patch P0800049 adding Solaris 2.6 support
   and also replaces patch P0800067.
   

patch P0800079 - auspex, hp_dm, sgi_dm, solaris, solaris_dm,
                solaris_26, solaris_dm_26
  HSM is unable to cache some files from optical platters
  or tapes.
  The problem is caused by an incorrect compression of the FHDB.
  The FHDB entry for a problem file is created incorrectly if
  during the making of a copy on the platter or tape the size of
  the granule header incleases to a size greater the 512 bytes.
  This may happen if the path of a problem file is arround 74 bytes long.
  This patch can NOT correct an existing FHDB but will prevent further
  incorrect FHDB entries.
  To correct the FHDB for a problem file do the following:
   1) run migopscan -n or migtscan -n on all the volumes the problem
      file resides on. If this patch has been installed migopscan or migtscan can be run
      without using the -n option.
   2) backup the FHDB, this just a save guard.
   3) replace existing FHDB op entries for the problem file with
      the FHDB.<volume> entries for the problem file.
   4) the problem file will now be able to be cached.

   NOTE: If a volume was consolidated while the FHDB was wrong
         for files on the volume, the resulting volumes will
        not have the correct data on them. This can result in
        loss of data.

   A problem file can have one of the following errors in the HSM log file
  when a cache fails.

    1)
09/09 09:44:18 migin[27031]: starting to reload file 3e8M101b
09/09 09:44:18 migin[27031]: exec /usr/openv/hsm/bin/admincmd/migcopy
09/09 09:44:19 migcopy[27032]: Starting
09/09 09:44:19 migcopy[27032]: exec /usr/openv/volmgr/bin/tpreq
09/09 09:44:32 migcopy[27032]: INFO: Mount DSC22A READ ONLY on /usr/var/openv/hsm/workdir/test2.1012.R
09/09 09:44:39 migcopy[27032]: WARNING: crc error in source. Handle 3e8M101b. FHDB crc 37972 != 7599 Computed
09/09 09:44:39 migcopy[27032]: ERROR: No valid granule found.
09/09 09:44:39 migcopy[27032]: INFO: Unmount vol=4114, delay=180, force=0, mount point=/usr/var/openv/hsm/workdir/test2.1012.R
09/09 09:44:39 migcopy[27032]: ERROR: Failed to copy x101b
09/09 09:44:39 migcopy[27032]: ERROR: copy_for_method() ret=1
09/09 09:44:39 migcopy[27032]: Finished 1
09/09 09:44:39 migin[27031]: ERROR reload of 3e8M101b failed, orig_owner=root path=/test2/acg/1234567890/1234567890/1234567890/1234567890/1234567890/A1234567

    2)
09/09 09:45:56 migin[27056]: starting to reload file 3e8M102f
09/09 09:45:56 migin[27056]: exec /usr/openv/hsm/bin/admincmd/migcopy
09/09 09:45:57 migcopy[27057]: Starting
09/09 09:45:57 migcopy[27057]: INFO: Mount DSC22A READ ONLY on /usr/var/openv/hsm/workdir/test2.1012.R
09/09 09:46:32 migcopy[27057]: INFO: Unmount vol=4114, delay=180, force=0, mount point=/usr/var/openv/hsm/workdir/test2.1012.R
09/09 09:46:32 migcopy[27057]: exec /usr/openv/volmgr/bin/tpreq
09/09 09:47:13 migcopy[27057]: INFO: Mount DSC22B READ ONLY on /usr/var/openv/hsm/workdir/test2.1013.R
09/09 09:47:19 migcopy[27057]: ERROR: zero length source granule header len= 0
09/09 09:47:19 migcopy[27057]: ERROR: No valid granule found.
09/09 09:47:19 migcopy[27057]: INFO: Unmount vol=4115, delay=180, force=0, mount point=/usr/var/openv/hsm/workdir/test2.1013.R
09/09 09:47:19 migcopy[27057]: ERROR: Failed to copy x102f
09/09 09:47:19 migcopy[27057]: ERROR: copy_for_method() ret=1
09/09 09:47:19 migcopy[27057]: Finished 1
09/09 09:47:19 migin[27056]: ERROR reload of 3e8M102f failed, orig_owner=root path=/test2/acg/1234567890/1234567890/1234567890/1234567890/1234567890/T1234567

patch P0800080 - auspex
  migsweep no longer selects sparse files for migration.  The migrate
  command has been modified so that use of the "-F" (force) option will
  over-ride and force the migration of a sparse file.

patch P0800081 - sgi_dm
  migsweepihand is getting a negative extents value returned
  from fcntl . This causes files to not be purged.
  The error message looks like

 09/17 21:09:51 migpurgeorder.sh[22186]: Starting
09/17 21:10:08 migsweepihand[22196]: ERROR: fcntl returned negative extents on IHAND file.
09/17 21:10:08 migsweepihand[22196]: WARNING read_ihand failure, hsmname=WS_DAT2

patch F0800083 - solaris_dm, solaris_dm_26
 Change to migstage command to restore files from nb method
 volumes by giving a list of these files to bpbrestore so
 only one bpbrestore is executed.

  This will help reduce netbackup mountings and positioning
 of volumes.


 This patch replaces the NetBackup tar and bpbkar commands.

  NOTE: If migstage is used to cache a file from an nb volume
       migstage will wait until the data is restored before
       migstage completes. To get around this run migstage
       in the background.

patch P0800085 - auspex, hp_dm, sgi_dm, solaris, solaris_dm
                solaris_26, solaris_dm_26
  Corrects a problem where HSM only allowed
  the handle for migrated files to reach ffffff .
  This allowed for only 16,777,215 migrated files.
  The patch allows handles to go to ffffffff or
  4,294,967,295 files.

patch P0800095 - auspex, solaris_dm, solaris_dm_26
   This patch corrects the following two problems:

   1) migcopy core dumps when HSM handle from tape does
      not match HSM handle in FHDB.

    2) migsweep fails to find some files.


patch P0800098 - auspex, hp_dm, sgi_dm, solaris, solaris_dm
                solaris_26, solaris_dm_26

   This patch corrects the following problem :

   migcopy core dumps when there is an ASYNC I/O problem.

 
patch P0800102 - auspex, hp_dm, sgi_dm, solaris, solaris_dm
                solaris_26, solaris_dm_26
   This patch corrects the following problems :

   Corrects problem with ASYNCH I/O when there is a
   missing granule on the ad volume.

   Corrects a problem where defered labeling of volumes
   would only use the default pool name HSM.

patch P0800104 - auspex, hp_dm, sgi_dm, solaris, solaris_dm
                solaris_26, solaris_dm_26

  Corrects a possible problem with optical platters
  where HSM should be using a block size of 1024 but
  instead uses a block size of 512. This may cause
  zero length granule headers.

patch P0800106 - auspex
  This patch prevents a migd hang on a system
  call open_byhandle .

patch P0800109 - auspex, hp_dm, sgi_dm, solaris, solaris_dm
                solaris_26, solaris_dm_26

  This patch corrects the following problems:

 Corrects a problem in mignospace where mignospace
 causes migbatch run but not select any files to migrate.

  Corrects a problem where miglock does not always
 update the pid in the running file.

  Corrects a problem where migd may core dump if
 a filesystem is made inactive while mignospace is
 running.

  Corrects a problem where migloc core dumps if the
 volume the file resides on is no longer found
 in the VOLDB.

 Corrects a problem where migcopy selected ad method volumes
 that were marked as full or dead.

  Corrects a problem where migcopy does not assign
 optical platters before executing migreg -F .
 May cause migcopy do execute migreg for each plater
 even if the platter ia already assigned to another HSM.

patch P0800115 - auspex, hp_dm, sgi_dm, solaris, solaris_dm
                solaris_26, solaris_dm_26

 This patch corrects the following problem:

 Changes migreg to make sure the volume being registered is not
 registered to any other hsmname.
 Before this change if a volume was not assigned in the media manager
 and the volume was registered to an hsmname the volume could
 could be registered to another hsmname.

patch F0800119 - sgi_dm
 This is a change to migstage to restore files from nb method
 volumes by giving a list of these files to bpbrestore so
 only one bpbrestore is executed.

  This will help reduce netbackup mountings and positioning
 of volumes.


  NOTE: If migstage is used to cache a file from an nb volume
       migstage will wait until the data is restored before
       migstage completes. To get around this run migstage
       in the background.

  This patch replaces the NetBackup tar and bpbkar commands

patch P0800121 - solaris, solaris_26
   This patch contains a fix to bpbkar so that Netbackup will only
   back up the inode for cached unmodified files.
   Without this patch Netbackup will backup the data.

   This patch is only for HSM .PAIN versions.

patch P0800122 - auspex, solaris, solaris_dm, solaris_26, solaris_dm_26
                solaris_dm33, solaris_dm33_26

 If a scsi reset happens while migcopy and migtlabel
 are writting to a tape they will over write good data
 on the tape.
 This fix causes migcopy and migtlabel to not write
 a tape if the cuurent position indicates the tape
 is at load point.


patch P0800123 - sgi_dm
This patch is required to run HSM 3.1.2 on an IRIX 6.5 platform.

patch P0800137 - auspex
 If a file is cached while copies are being made
 migcopy and other commands that reference the
 file being cached may hang.
 A cache of a file that was not purged may leave
 the IHAND entry incorrect and the file may have
 incorrect bindings on it.

patch S0800143 - hp_11 64 bit only
 HSM 3.1.4 on HP-11.0 does not support 64 bit kernels.  A 64 bit version
 of the migattr driver is required and changes are needed in install.hsm to
 configure a kernel with the correct driver.

 This patch installs the correct 64 bit version of migattr.o in
 directory /usr/openv/hsm/bin/hp_11_64/migattr.o.  It also
 installs a new version of install.hsm in
 /usr/openv/hsm/bin/install.hsm.  The new install.hsm will use the
 64 bit migattr.o when building the HSM aware HP-UX kernel.

Installation Instructions:

1. Make sure that the system is running a 64 bit kernel.
  Do this by executing the command:
       /usr/bin/getconf _CS_KERNEL_BITS
  If this command returns the string   64  then a 64 bit kernel is running correctly.
  If the comman returns the string   32
  then a 32 bit kernel is running.

2. Make sure that the special DMAPI bugfix patch from
  HP, PHKL_64HSM_2 (Unofficial_test_patch) is installed.  This HP
  patch supersedes PHKL_17869 ( PM/VM/UFS/async/scsi/io/DMAPI/JFS/perf cumul).
  If patch PHKL_17869 is the latest HP kernel patch installed, it will
  not successfully run HSM on a 64 bit kernel.

3. Add the dmapi and vxportal drivers to the kernel as documented in
  the initial installation instructions.

4. Make sure that Media manager is installed and configured.

5. Attempt to install HSM.  During the installation of HSM there will be a
  failure configuring a new kernel.   The messages will look like this:
 > Installing HSM...
 > ar: creating /usr/conf/lib/libhsm.a
 > Executing   "/usr/sbin/config /migration/system"
 > Compiling conf.c...
 > Loading the kernel...
 > ld: Unknown input file type: "/usr/conf/lib/libhsm.a"
 > Fatal error.
 > *** Error exit code 1
 >
 > Stop.
 > config: make did an exit(1)
 > Replacing /stand/vmunix with /migration/build/vmunix_test.
 > cp: cannot access /migration/build/vmunix_test: No such file or directory
 > Saving /usr/openv/netbackup/bin/bpbkar to /usr/openv/netbackup/bin/bpbkar.PREHSM  > Replacing /usr/openv/netbackup/bin/bpbkar ...
 > Saving /usr/openv/netbackup/bin/tar to /usr/openv/netbackup/bin/tar.PREHSM
 > Replacing /usr/openv/netbackup/bin/tar ...

 > Will this installation use ONLY the "ft", "nb" and/or "ad" methods? [yn](n):

  The failure  >> ld: Unknown input file type: "/usr/conf/lib/libhsm.a" <<
   occurs because the released HSM 3.1.4 package is unaware of the possibility
  of a 64 bit kernel.

   The HSM 3.1.4 install process continues beyond this failure.  When the
  question is asked:
    Will this installation use ONLY the "ft", "nb" and/or "ad" methods?
  The installation can be interrupted at this time, or answer the question.
  This step will be repeated after the patch is installed.

6. Get the patch to the /tmp directory on the server.
   cd /tmp
   Acquire patchS0800143.README via ftp
   Acquire patchS0800143.hp_11.tar.Z via ftp
   Acquire Vrts_patch.install via ftp

7. Install the new binaries.
   cd /tmp
   /bin/sh ./Vrts_patch.install

8. Create the new HP-UX 11.0 boot file by executing:
  /usr/openv/hsm/bin/install.hsm -i
  This creates a new boot file.  This was attempted earlier in step
  5.  This time no error should occur since the installation process
  is now aware of what is needed and has the correct 64 bit binary
  file to use.

9. The system will have to be rebooted using the new kernel in
  /stand/vmunix before HSM can be used.

patch P0800146 - auspex, hp_dm, sgi_dm, solaris, solaris_dm
                solaris_26, solaris_dm_26

   This patch corrects a possible core dump
  of migout.

patch P0800147 - solaris_26
  Fix to the NetBackup tar command to prevent tar from assigning the wrong
  length to migrated files when they are restored to an HSM managed filesystem.

   This patch replaces the NetBackup tar command.

patch P0800158 - hp_dm, sgi_dm, solaris_dm, solaris_dm_26
 This is a change to migstage to restore files from nb method
 volumes by giving a list of these files to bpbrestore so
 only one bpbrestore is executed.

  This will help reduce netbackup mountings and positioning
 of volumes.


  NOTE: If migstage is used to cache a file from an nb volume
       migstage will wait until the data is restored before
       migstage completes. To get around this run migstage
       in the background.

 This patch replaces the NetBackup tar and bpbkar commands.

patch P0800160 - sgi_dm
   On SGI the use of the -N option with migcons when
   consolidating tapes will cause the consolidation to fail.

Installation Instructions:

patch P0800165 - solaris_26

  . pfinit -e fails when trying to expand a pain file on solaris 5.6
   ERROR: <filesystem_path>.PAIN Cannot statfs

  . pfinit, pfprint, and pfcheck all print incorrect statfs
   info on solaris 5.6 .

patch P0800170
     If HSM is caching a lot of files and there is contention
     for drives and volumes HSM will sometimes cache the file
     from the 2nd copy instead of the 1st copy.

      When this happes the foolowing messages will be in the log file :

      ERROR: SOURCE <volume> mount problem
     ERROR: mt_mount() hard failure, adandoning SRC mount <volume>


Download Instructions:

Download 31_HSM.<server_platform>.tar to the /tmp directory.

       server_platform is: auspex, hp_11, hp_dm, sgi_dm,
            solaris_dm33, solaris_dm33_26,
            solaris_26, solaris_dm_26, solaris, solaris_dm.

      NOTE: platofmrs with the _dm are for  the DMAPI version of HSM.
      To determine if DMAPI or UFS HSM is installed:
            cat /usr/openv/hsm/version

      This will contain the HSM server_platform installed
      on a system.
      
Extract 31_HSM.<server_platform>.tar
      /bin/tar xvf 31_HSM.tar

      This will create the files: patchJ0800168.<server_platform>.tar.Z,
      patchJ0800168.README, patchP0800177.<server_platform>.tar.Z,
      patchP0800177.README and Vrts_patch.install

Please read the entire README files carefully before installing.


Attachments

31_HSM.solaris_dm33_232105.tar (1.3 MBytes)

Legacy ID



232105


Article URL http://www.symantec.com/docs/TECH8618


Terms of use for this information are found in Legal Notices