cancel
Showing results for 
Search instead for 
Did you mean: 

storage foundation error when upgrading to 5.1 help

Howell_P
Level 3
Partner
Hi, Hope someone can help me on this.. I'm trying to upgrade Storage foundation 5.0 to 5.1 on solaris..The root disk is mirrored and encapsulated and the rest of the volumes are under volume manager but the file system is UFS...Followed the procedure on how to upgrade storage foundation with encapsulated root disk but upgrade was not complete error occurs after the ugrade...I'm also attaching some of the procedure that was capture during the upgrade process.. On the procedure, after running the upgrade_start scripts..A reboot is needed if the root disk is encapsulated..But upon boot up vxconfigd is not running anymore...Is this normal or do i need to execute the upgrade_start scripts in single user mode.. Hope someone can help me on this. Sample of the /etc/vfstab #device device mount FS fsck mount mount #to mount to fsck point type pass at boot options # fd - /dev/fd fd - no - /proc - /proc proc - no - /dev/vx/dsk/bootdg/swapvol - - swap - no nologging /dev/vx/dsk/bootdg/rootvol /dev/vx/rdsk/bootdg/rootvol / ufs 1 no nologging /dev/vx/dsk/bootdg/var /dev/vx/rdsk/bootdg/var /var ufs 1 no - /dev/vx/dsk/bootdg/home /dev/vx/rdsk/bootdg/home /export/home ufs 2 yes - /dev/vx/dsk/bootdg/opt /dev/vx/rdsk/bootdg/opt /opt ufs 2 yes - /devices - /devices devfs - no - ctfs - /system/contract ctfs - no - objfs - /system/object objfs - no - swap - /tmp tmpfs - yes - #NOTE: volume rootvol (/) encapsulated partition c1t0d0s0 #NOTE: volume swapvol (swap) encapsulated partition c1t0d0s1 #NOTE: volume opt (/opt) encapsulated partition c1t0d0s3 #NOTE: volume var (/var) encapsulated partition c1t0d0s4 #NOTE: volume home (/export/home) encapsulated partition c1t0d0s7 /dev/vx/dsk/oradg/oravol /dev/vx/rdsk/oradg/oravol /oradata vxfs 1 yes suid /dev/vx/dsk/datadg/datavol /dev/vx/rdsk/datadg/datavol /data ufs 1 yes largefiles,suid Output of the Upgrade: Checking installed packages on sparcsvr uninstallsf is now ready to uninstall SF. All SF processes that are currently running must be stopped. Are you sure you want to uninstall SF? [y,n,q] (y) Checking sparcsvr for attached volumes VxVM vxsvmdiag ERROR V-5-1-10981 error getting li st: : iod,info config device open fail , specific reason:-1 ....................................... None Cannot check for open volumes on sparcsvr because the vxconfigd process is not in enabled mode. You must affirm that there are no open volumes if you want to continue. Are there any open volumes on sparcsvr [y,n,q] (n) ******************** Uninstalling SF: 100% Uninstall completed successfully on all systems Uninstallation log files, summary file, and response file are saved at: /opt/VRTS/install/logs/uninstallsf-E5EsWp NOTICE: As part of the uninstallation process on Solaris, the current configuration of VxVM was saved in the directory /VXVM-CFG-BAK. This configuration may be used in a future installation of VxVM. If you do not plan to reuse it, you may manually remove this subdirectory. ********************** 1) Veritas Cluster Server (VCS) 2) Storage Foundation (SF) 3) Storage Foundation and High Availability (SFHA) 4) Storage Foundation for Cluster File System (SFCFS) 5) Storage Foundation for Cluster File System/HA (SFCFSHA) 6) Veritas Storage Foundation for Oracle RAC (SF Oracle RAC) 7) Symantec Product Authentication Service (AT) b) Back to previous menu Select a product to install: [1-7,b,q] 2 *************************** Logs are being written to /var/tmp/installer-201004141700Xiy while installer is in progress Starting SF: 92% _______ Estimated time remaining: 0:04 12 of 13 Performing SF configuration .................................................................. Done Starting vxdmp ............................................................................... Done Starting vxio ................................................................................ Done Starting vxspec .............................................................................. Done Starting vxconfigd ........................................................................... Done Starting vxesd ............................................................................... Done Starting vxrelocd ............................................................................ Done Starting vxconfigbackupd ..................................................................... Done Starting vxportal ............................................................................ Done Starting fdd ................................................................................. Done Starting vxdbd ............................................................................... Done Starting odm ................................................................................. Done Storage Foundation Startup did not complete successfully vxdmp failed to start on sparcsvr vxio failed to start on sparcsvr vxspec failed to start on sparcsvr vxconfigd failed to start on sparcsvr vxesd failed to start on sparcsvr ************************** bash-3.00# ./upgrade_finish VERITAS Volume Manager VxVM 5.1 upgrade completion script This script completes the process of upgrading your system from previous releases to the VxVM 5.1 release of the VERITAS Volume Manager and upgrading Solaris to 2.6 or later versions. If you saved old configuration information using upgrade_start before you removed the previous Volume Manager package, then you can run this script after adding the VxVM 5.1 package to restart the Volume Manager on your system NOTE: This can also be used to undo whatever upgrade_start script did. Hit RETURN to continue. Verifying installation of VxVM 5.1 ... /etc/vx/reconfig.d exists. Licensing information: System host ID: 83bcb88d Host type: SUNW,Sun-Fire-V240 Storage Expert VxVM PGR PGR_TRAINING Dynamic Lun Expansion Hardware assisted copy Cross-platform Data Sharing RAID5SNAP Some VxVM 5.1 licenses are already installed. Do you wish to enter another license key; yes/no? [no] The following operation will modify the /etc/system file. The old /etc/system file will be saved in /VXVM5.1-UPGRADE/upgrade_finish_system.SAV Hit RETURN to continue. VxVM vxddladm INFO V-5-1-14412 eventsource is not running copying /kernel/drv/sparcv9/vxdmp.SunOS_5.10 /kernel/drv/sparcv9/vxdmp copying /kernel/drv/sparcv9/vxio.SunOS_5.10 /kernel/drv/sparcv9/vxio copying /kernel/drv/sparcv9/vxspec.SunOS_5.10 /kernel/drv/sparcv9/vxspec Loading VM driver (vxdmp) ... (vxdmp) load complete. Loading VM driver (vxio) ... (vxio) load complete. Loading VM driver (vxspec) ... (vxspec) load complete. copying /sbin/vxconfigd.SunOS_5.10 /sbin/vxconfigd The following operation will modify the /etc/vfstab file. The old /etc/vfstab file will be saved in /VXVM5.1-UPGRADE/upgrade_finish_vfstab.SAV. Hit RETURN to continue. Restoring saved VxVM state files... Restarting the volume manager ... Starting vxconfigd ... failed. VxVM ERROR V-5-2-2939 Operation of the volume manager could not be restored. You may just need to reboot your system and try again. If you have already rebooted your system, consult the Volume Manager documentation for further help. Stopping config daemon if running... Restoring /etc/vfstab... Restoring /etc/system... Restoring /etc/vx/jbod.info... Upgrade operation is incomplete. Consult the Volume Manager documentation if you need further help.
8 REPLIES 8

mightknowsometh
Level 3
Employee Certified

OK, what the upgrade script tries to do is to unencapsulate you boot disk (there a replenty doco on how to do it manually -- http://www.symantec.com/business/support/knowledge_base_results.jsp?SearchTerm=unencapsulate+boot+disk&ddProduct=&pid=&content=all ).

In short, break the mirror (same site just search), unencapsulate, upgrade, .. if all fine , then re-encapsulate, add mirror....     if the upgrade did not go to plan, boot from the mirror and start again....

Javier_E_Rodrig
Level 2
Partner Accredited

Ok based on your vfstab, this is how the vfstab should look when the root disk is not encapsulated:

#device device mount FS fsck mount mount
#to mount to fsck point type pass at boot options
#
fd - /dev/fd fd - no -
/proc - /proc proc - no -
#/dev/vx/dsk/bootdg/swapvol - - swap - no nologging
#/dev/vx/dsk/bootdg/rootvol /dev/vx/rdsk/bootdg/rootvol / ufs 1 no nologging
#/dev/vx/dsk/bootdg/var /dev/vx/rdsk/bootdg/var /var ufs 1 no -
#/dev/vx/dsk/bootdg/home /dev/vx/rdsk/bootdg/home /export/home ufs 2 yes -
#/dev/vx/dsk/bootdg/opt /dev/vx/rdsk/bootdg/opt /opt ufs 2 yes -


/dev/dsk/c1t0d0s1       -                       -               swap    -       no      -
/dev/dsk/c1t0d0s0       /dev/rdsk/c1t0d0s0      /               ufs     1       no      -
/dev/dsk/c1t0d0s4       /dev/rdsk/c1t0d0s4      /var       ufs     2       no     -
/dev/dsk/c1t0d0s3       /dev/rdsk/c1t0d0s3      /opt            ufs     2       yes     -
/dev/dsk/c1t0d0s7       /dev/rdsk/c1t0d0s7      /export/home    ufs     2       yes     -

/devices - /devices devfs - no -
ctfs - /system/contract ctfs - no -
objfs - /system/object objfs - no -
swap - /tmp tmpfs - yes -
#NOTE: volume rootvol (/) encapsulated partition c1t0d0s0
#NOTE: volume swapvol (swap) encapsulated partition c1t0d0s1
#NOTE: volume opt (/opt) encapsulated partition c1t0d0s3
#NOTE: volume var (/var) encapsulated partition c1t0d0s4
#NOTE: volume home (/export/home) encapsulated partition c1t0d0s7
/dev/vx/dsk/oradg/oravol /dev/vx/rdsk/oradg/oravol /oradata vxfs 1 yes suid
/dev/vx/dsk/datadg/datavol /dev/vx/rdsk/datadg/datavol /data ufs 1 yes largefiles,suid

But you also need to comment from the "/etc/system" these two lines:
* rootdev:/pseudo/vxio@0:0
* set vxio:vol_rootdev_is_volume=1

I also recommend that you comment all additional file system entries in the "/etc/vfstab" to avoid falling in single user mode while rebooting in case that one volume is not available.

After this, try a reboot of Solaris and verify with the df -h that is running from unencapsulated disks.

If you succeded at that, then you can upgrade by un-insalling storage foundation (old ) and instaling the new version. After you finish the clean installation just "vxdg import" your disk groups.

To reconfigure the mirrors of you OS I recommned using Solaris Volume Manager to avoid these problems in the future.

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified

I am actually wondering if there's something wrong with the 5.1 upgrade_start and upgrade_finish...
I have in the past manually unencapsulated root disk before upgrading.
When I read about your problem, I had to try upgrade_start & upgrade_finish for myself. I installed SF 4.1 on a test V240 system (Solaris 10), encapsulated & mirrored root disk.
To cut a long story short, I encountered exactly the same problems as you did.

I have found this TechNote and plan to work through this to see if I can get my installation fixed:
http://seer.entsupport.symantec.com/docs/247008.htm

Apologies for the delay in response - my lab time is very limited.

****EDIT*****

I do realize that the TechNote is VERY old, but I will see how much of it is actually still usable.

Howell_P
Level 3
Partner
I will try the procedure based on the technote you've attached...Does this procedure works in you lab?  Would it make a difference if i unencapsulate the rootdisk manually compared to the upgrade_start script that also does the same thing...Is there really a problem with the scripts of SF 5.1. Hope someone can help me on this..

Howell_P
Level 3
Partner
is there a difference if i so it manually and do the upgrade? Does this procedure works for you.. Hope symantec can provide details on how to resolve this problem. The upgrade documentation is fairly straight forward but after running the upgrade_start script (successfully) the documents states that a reboot is needed if the root disk is encapsulated..But after doing the reboot, vxconfigd is not accessible or not running anymore..
Is this a normal behavior of storage foundation after an upgrade_start and reboot?

Thanks,

Howell

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified

Sincere apologies, Howell... I am currently quite busy with customer support calls and haven't had any time to try and fix my lab system.
I am hoping to find time before the weekend.
If you are running out of time to get you system up and running, we can possibly escalate this post to Symantec Support?

Michael_Auria
Level 5
Employee Certified
I'd like to help you get your system successfully running 5.1 

Can you open a case with our support offices ?  If you are based in the US, I would be glad to help you directly.  Otherwise, the next available engineer will be able to assist you.

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified
Hi Howell
I sincerely hope that you managed to get this sorted out?
I only had the opportunity today to attempt the fix as per TechNote mentioned previously.
I am happy to say that my installation is now 100% fine - root disk encapsulated and mirrored.

The TN is VERY old and some of the stuff not relevant anymore (like /etc/vex....). This is what I've done:

  1. backed up /VXVM* (using tar)
  2. pkgrm VRTSvxvm (no patches to remove with patchrm as no 5.1 patches were installed yet; ignored package dependancy messages)
  3. cd <pkgs install folder>
  4. pkgadd -d VRTSvxvm.pkg (answer y to conflict messages) Checked installation goes through fine : Installation of <VRTSvxvm> was successful.
  5. Ran upgrade_finish that is in / directory (NOT the one in ..../dvd1-sol_sparc/scripts - I think this is what I've done wrong previously...). Result of upgrade_finish:
# pwd
/

# ./upgrade_finish
VERITAS Volume Manager VxVM 5.1 upgrade completion script

This script completes the process of upgrading your system from
previous releases to the VxVM 5.1 release of the VERITAS Volume Manager
and upgrading Solaris to 2.6 or later versions.

If you saved old configuration information using upgrade_start before
you removed the previous Volume Manager package, then you can run this
script after adding the VxVM 5.1 package to restart the Volume Manager
on your system

NOTE: This can also be used to undo whatever upgrade_start script did.

Hit RETURN to continue.

Verifying installation of VxVM 5.1 ...
  /etc/vx/reconfig.d exists.

Licensing information:
    System host ID: 83064a8d
    Host type: SUNW,Sun-Fire-V240

Storage Expert
VxVM
FASTRESYNC
DGSJ
PGR
PGR_TRAINING
VVS_CONFIG
Dynamic Lun Expansion
Hardware assisted copy
Cross-platform Data Sharing
RAID5SNAP
Site Awareness

Some VxVM 5.1 licenses are already installed.

Do you wish to enter another license key; yes/no? [no]

The following operation will modify the /etc/system file.  The old
/etc/system file will be saved in /VXVM5.1-UPGRADE/upgrade_finish_system.SAV

Hit RETURN to continue.
VxVM vxddladm INFO V-5-1-14412 eventsource is not running
copying /kernel/drv/sparcv9/vxdmp.SunOS_5.10 /kernel/drv/sparcv9/vxdmp
copying /kernel/drv/sparcv9/vxio.SunOS_5.10 /kernel/drv/sparcv9/vxio
copying /kernel/drv/sparcv9/vxspec.SunOS_5.10 /kernel/drv/sparcv9/vxspec
  Loading VM driver (vxdmp) ... (vxdmp) load complete.
  Loading VM driver (vxio) ... (vxio) load complete.
  Loading VM driver (vxspec) ... (vxspec) load complete.
copying /sbin/vxconfigd.SunOS_5.10 /sbin/vxconfigd

The following operation will modify the /etc/vfstab file.  The old
/etc/vfstab file will be saved in /VXVM5.1-UPGRADE/upgrade_finish_vfstab.SAV.

Hit RETURN to continue.

Restoring saved VxVM state files...

Restarting the volume manager ...
  Starting vxconfigd ... started.


VERITAS Volume Manager Release 5.1 has been successfully restarted on
your system.  You must reboot the machine now to restore the volumes in
the root disk.

Hit RETURN to continue.
Starting volumes ... started.

VERITAS Volume Manager Release 5.1 has been successfully restarted on
your system.