Quantcast
Channel: Zimbra Forums
Viewing all articles
Browse latest Browse all 749

Administrators • Large mailbox backup strategy with zimbraBackupSkipHsmBlobs

$
0
0
I have many users with large mailboxes, average 50GB and more. I use one mailstore to handle 100 users and the /opt/zimbra/store close to 4TB. Autogroup backup strategy is used. This still take average 6 hours to backup 15 users every day. I have 5 mailstore servers all will begin backup at the same time at 1am. Disk I/O is busy every night. Those at different timezone will experience slowness during backup time.

What do you suggest I could improve? I can't add in faster storage box. This is the best config I have for now.

I am thinking of zimbraBackupSkipHsmBlobs. First I setup /secondary mount via NFS to move mail older than 1 year over. This should cut down primary volume down to 1.5TB or less. Then the backup will ignore the /secondary. Sounds good right? But I worry someone wiped his entire mailbox and I could not restore back his data after 1 year.... how to avoid this? My storage box is TrueNAS and I could do snapshot for this. Has anyone done like this?

I have tried S3 briefly but it didn't turn out satisfactory. The space usage in S3 is extremely high. It is also slow to browser bucket. May be my Minio setup was wrong. This setup caused many timeout to Zimbra and we constantly have strange support problem like mailbox locked out. The only solution seemed to restart mailboxd. I have moved out from this config.

Appreciate all good suggestions.

Thanks.

Statistics: Posted by yeak — Fri Mar 08, 2024 12:50 am



Viewing all articles
Browse latest Browse all 749

Trending Articles