VCB - forgotten snapshots and functionless autodelete during next backup job

[rocker77]

Hello,

I would need resolve my problem with VCB or more precisely with undeleted snapshots. I use VCB in SAN mode and sometimes after backup completes, VCB don't delete snapshot. For this situation I set parameter PREEXISTING_VCB_SNAPSHOT="delete" in config.js so VCB should delete forgotten snapshot in next backup job, but sometime VCB delete snapshot and sometime next backup job fail. When I try delete snapshot manually by snapshot manager after backup job fail it works.

I use VCB 1.1 and VC 2.5 U2 + ESX 3.5 U2

Do you have any experiences with this issue?

Thanks!!!!

[Rubeck]

HI Roc...

I don't know the exact solution to your problem, but shouldn't you be running VCB 1.5 when using VC and ESX U2?

/Rubeck

[rocker77]

VCB 1.5 isn´t requested for VI U2.

[Rubeck]

I know... Was just curious

/Rubeck

[rocker77]

This is is discrussed in thread below

http://communities.vmware.com/message/1074921

[jrVMguy]

I am experiencing the same issues. I read the referenced post but that didn't provide an answer.

I did not experience these issues until after we upgraded the VC.

Environment details:

ESX 3.0.2 and ESX 3.5u2

Virtual Center 2.5u3 - upgrade from 2.0.2

Physical HS20 blade in the Bladecenter - Win2k3 SP2

VCB 1.5 - with TSM integration scripts installed

Combination of IBM bladecenter (HS20 blades) and IBM x3650 hardware

All VMs are on FC SAN (IBM ESS800)

My issues are twofold:

1. The command to unmount the VCB snapshots does not run after the backup completes.

This leaves the snapshots in my backup directory that I have to manually remove.

2. I am seeing "Operation failed: snapshot timed out" errors on some nights.

As you can guess that leaves the snapshot on the VM that I have to manually delete.

I believe I am seeing issue 2 on the night following issue 1. We finally got most of the bugs worked out the start of this week.

I believe if we can get an answer to issue 1 issue 2 will go away - I am hoping.

Can anyone shed some light on the the time out errors or provide me with some direction to start looking ????

Is there a command I can run on the VC server to clean up the snapshots when they are left behind ?????

If one exists, are there any changes that need to be made to it ??? i.e. is the script generic ????

[rocker77]

Exactly same issues as me. I thing that command which should delete snapshot after backup job completes try delete snapshot but crashes on "snapshot timed out".

I solve this by enabling auto delete and auto unmount in config.js for VCB framework.

From yesterday I have to solve another bad thing. During creating snapshot (before backup job) VM froze for a 15 minutes, I got "snapshot timed out"and VM begun run.

During 15 minutes VM didn't serve its resources to network/users!!! :smileyangry:

[jrVMguy]

I check my backups daily - need to make sure they are there.

I am getting the snapshot timed out error when it is trying to create the snapshot. When that happens my backup fails and it deletes all the snapshots that were created sucessfully and NO backup happens. :_| I am going to have to do some digging to see if I can find out why that is happening as that is worse than the post backup script not running.

That is the other issue I am seeing - when my snapshots happen ok the "post backup" script is not running to unmount the snapshots from my proxy. the backup finishes ok but I have to manually clean up the snapshot directory on my VCB proxy server.

The backup happens during off hours since we are not a 24/7 shop so I don't know where to look to see what is happening and the PRE VCB log doesn't give much info.

[rocker77]

You don't have to manually clean up mnt directory. You can do it automatically by VCB before next backup. It works good.

[jrVMguy]

I think I have fixed my issue.

Last week my boss and I looked at the DSM.OPT file and found that it had the POST-BACKUP "script" listed before the PRE-BACKUP "script". After we made that change it seems to have solved my issues. I have been monitoring it since the change and other than the SECOND night where it messed up it has been working as it used to work.

We had seen issues in the past when we have added or deleted VM's from the list to backup and it totally hosed the DSM.OPT file.

At the end of our DSM.OPT file I see the PRE and POST commands listed out - similar to this:

PRESCHEDULECOMMAND "pre-command VMbackup VMNAME1-FullVM VMNAME2-FullVM ....."

POSTSCHEDULECOMMAND "post-command VMbackup"

[sisdip]

Hi Roc,

One question,,, if I add a PREEXISTING_VCB_SNAPSHOT="delete" parameter in config.js , I have to restart the service "VMware Virtual Mount Manager Extended" or not does do anything? Tanks & regards Davide