[SOLVED] Increased disk usage post 7.0 upgrade

[sdrury]

Hi,

Have searched the forum and also raised a support ticket with Zimbra for this issue.

Is anyone else experiencing major increase in hard disk usage after upgrading to 7.0?

e.g
We have been pre 7.0 been sitting around 69% disk usage until just after the 7.0 upgrade this suddenly increased to 86% then about a week later jumped to 94%!

I am at a loss as to what has caused this but have a suspicion about the new feature in 7.0 that allows a user to recover deleted email from the trash folder. This must have an effect on the amount of email stored on the server?

Regards,
Sean.

[phoenix]

I doubt very much that the 'recover' feature would consume that much disk space but you never know. Have you looked at the usage for your backup sessions, are they being removed correctly or does it contain more sessions than it should?

[LMStone]

Originally Posted by sdrury

Hi,

Have searched the forum and also raised a support ticket with Zimbra for this issue.

Is anyone else experiencing major increase in hard disk usage after upgrading to 7.0?

e.g
We have been pre 7.0 been sitting around 69% disk usage until just after the 7.0 upgrade this suddenly increased to 86% then about a week later jumped to 94%!

I am at a loss as to what has caused this but have a suspicion about the new feature in 7.0 that allows a user to recover deleted email from the trash folder. This must have an effect on the amount of email stored on the server?

Regards,
Sean.

You don't say from which version of ZCS you upgraded, but if it was an older 6.0x version, the default settings for the backups changed to remove hard links. The net effect is that each backup is independent, and if you keep 30 days of backups, that 10MB PowerPoint attachment will take up 300MB. Search these forums for the "--noZip" option and you'll see what's what.

The Recover function may also be contributing; if you hadn't previously tracked how much daily inbound email you get (and what percentage of that is deleted by users) you might be surprised.

So, I agree with Bill that both of these areas bear investigation.

Hope that helps,
Mark

[sdrury]

Originally Posted by phoenix

I doubt very much that the 'recover' feature would consume that much disk space but you never know. Have you looked at the usage for your backup sessions, are they being removed correctly or does it contain more sessions than it should?

Hi Bill,

We have backups back to the 10th February - should only be 30days worth right?

Regards,
Sean.

[sdrury]

Originally Posted by LMStone

You don't say from which version of ZCS you upgraded, but if it was an older 6.0x version, the default settings for the backups changed to remove hard links.

Hi Mark,

Upgraded from the latest version of 6 at the time. I have searched through the release notes and not found anything that relates to changes in how the backup operate. Only something about mysql dumps being able to be included in backups.

Will now search the forums as you said, but as I mentioned to Bill earlier we appear to have backups older than 30days on the system.

Regards,
Sean.

[LMStone]

Originally Posted by sdrury

Hi Mark,

Upgraded from the latest version of 6 at the time. I have searched through the release notes and not found anything that relates to changes in how the backup operate. Only something about mysql dumps being able to be included in backups.

Will now search the forums as you said, but as I mentioned to Bill earlier we appear to have backups older than 30days on the system.

Regards,
Sean.

Hmmm. Backups older than 30 days are still around?

I'd do as the zimbra user "crontab -l | grep backup" to make sure all three backup lines are in there (weekly full, daily incremental, and daily purge). You should see something like what we use:

zimbra@viognier:~> crontab -l | grep backup
0 1 * * 6 /opt/zimbra/bin/zmbackup -f --noZip --includeHsmBlobs -a all --mail-report
0 1 * * 0-5 /opt/zimbra/bin/zmbackup -i --noZip --includeHsmBlobs --mail-report
0 0 * * * /opt/zimbra/bin/zmbackup -del 12d
zimbra@viognier:~>

You can purge backups older than thirty days manually by running as the zimbra user:

/opt/zimbra/bin/zmbackup -del 30d

Keep us posted!

Hope that helps,
Mark

[sdrury]

Hi Mark,

This is what I get:

zimbra@mail:~$ crontab -l | grep backup
0 1 * * 6 /opt/zimbra/bin/zmbackup -f -a all
0 1 * * 0-5 /opt/zimbra/bin/zmbackup -i

As you can see doesn't look like what you would expect - just wondering why this has changed? I certainly have not made any changes to the backup side.

Thanks for your help so far too.

Regards,
Sean.

[sdrury]

Mark,

Looking at your CRON example, you have 3 extra variables for the 2 backup lines: --noZip --includeHsmBlobs --mail-report
Are these required or just variables that you have set for your environment?
Also you choose to keep only 12days of backups?

Regards,
Sean.

[sdrury]

I have now added:

0 0 * * * /opt/zimbra/bin/zmbackup -del 1m

to CRON and manually run it as well which has now dropped our disk back to a respectable size.

I believe that this was the issue and not the dumpster feature as first thought.

Thanks for your help Mark and Bill.

Regards,
Sean.

[sdrury]

Zimbra support just got back to me and also mentioned the following bug report:
Bug 57368 – zmschedulebackup not showing delete jobs

Regards,
Sean.