04-16-2014 05:20 AM
I've a strange issue with media servers crashing to blue screen when being backed up , but they are working absolutely fine backing up clients.
This is a new environment and the Master and Media servers are all running NetBackup 126.96.36.199
Master is a VM running Windows 2008 R2 standard
4 x Physical Media servers all running Windows 2008 R2 Enterprise ... all have the same hardware config
Client backups work fine, both to MSDP and Tape, as does a backup of the Master server, and Optimised media server Duplications also work fine between the media storage servers.
But when I try and backup the media servers themselves (to any backup target ...MSDP or Tape, either to themselves or to any other media server) the media server being backed up crashes to blue screen.
PAGE_FAULT_IN_NONPAGED_AREA (*** STOP: 0x00000050 ...
I found : http://www.symantec.com/business/support/index?page=content&id=TECH206217 and applied the regedit as the MSDP volume is 64TB, but they still crash to blue screen but now with .
**** STOP: 0x0000007E
I've run a memory test on 2 of the servers and they pass with no issues detected.
This happens every time, regular as clockwork, the media server being backed up will crash as above, yet they backup clients and Duplicate client images perfectly successfully.
I've tried various backup selections in the Policy, as that was how I found TECH206217, but they still crash to blue screen ...even attempting to backup a test folder of < 100kb.....the MSDP drive and MSDP database drive are excluded from all policies..
I'm aware that this is probably not a Netbackup issue as such, but has anyone seen this before with media servers that have large MSDP pools, and are there any recommended Windows settings I should be using ?
..also, this is a new environment ...so backing up the media servers has never worked, it has always resulted in the media server being backed up crashing as above
04-16-2014 05:46 AM
Only seen this many years ago on W2003 server with a specific version of Storage Foundation (4.1!).
Can you confirm that SP1 for W2008 R2 is installed?
If so, I would log a SEV 1 call with Symantec and get them to analyze the crash dumps.
This TN is merely dealing with slow performance when W2008 R2 media servers are backing up themselves, but maybe something useful there?
04-16-2014 05:48 AM
Suggestion: BIOS & HBA driver upgrade.
04-16-2014 06:00 AM
Thanks for the replies ...
The media servers are all on Win2008 R2 with SP1.
The HBA driver is the latest Emulex FC driver downloaded recently. (and I've used these drivers on another environment I've built with SAN presented MSDP pools)
BIOS ...don't know but these servers were only built about 4 weeks ago, so I'd expect these to be recent as well.
04-16-2014 10:29 AM
Has this ever worked before? If this is a happening all of a sudden, did something change like a windows patch or driver upgrades? Patch tuesday was a week ago.
If you do open a case, pm me with the case number.
05-29-2014 03:27 AM
Sounds a little like the problem we had back in the Windows 32bit days with machines hosting multiple Oracle databases
If I remember correctly it had to with the memory not being released fat enough for new allocation.
The server guys changed some memory pool registry keys for the high water marks, sorry but cannot remember the exact keys