Full backup finishing with errors

[cirrus]

For a few days we have a problem with zmbackup. It ends with errors.

mailbox.log shows (example entry) for almost every account:

2020-05-10 03:11:35,314 INFO [FullBackupThread] [name=some-address@ourdomain.com;] index - OpenLuceneIndex impl=NIOFSDirectory,dir=/opt/zimbra/index/0/492/index/0
2020-05-10 03:11:35,314 INFO [FullBackupThread] [name=some-address@ourdomain.com;] mbxmgr - Mailbox 492 account c2169c42-1942-47f6-8383-9e6b4fd06e8a LOADED
2020-05-10 03:11:35,314 INFO [FullBackupThread] [name=some-address@ourdomain.com;] cache - initializing folder and tag caches for mailbox 492
2020-05-10 03:11:35,357 INFO [FullBackupThread] [mid=492;] backup - redo log file sequence is 1796 at full backup for some-address@ourdomain.com
2020-05-10 03:11:35,358 INFO [FullBackupThread] [mid=492;] backup - Full backup started for account some-address@ourdomain.com (c2169c42-1942-47f6-8383-9e6b4fd06e8a) mailbox 492
2020-05-10 03:11:35,358 INFO [FullBackupThread] [mid=492;] mailbox - Putting mailbox 492 under maintenance.
2020-05-10 03:11:35,359 INFO [FullBackupThread] [mid=492;] backup - Number of blobs to backup for mailbox 492: 55
2020-05-10 03:11:35,359 INFO [FullBackupThread] [mid=492;] mailbox - Ending maintenance on mailbox 492.
2020-05-10 03:11:35,359 INFO [FullBackupThread] [mid=492;] mbxmgr - Mailbox 492 account c2169c42-1942-47f6-8383-9e6b4fd06e8a AVAILABLE
2020-05-10 03:11:35,362 INFO [BackupSetWorkerThread-BLOB] [name=some-address@ourdomain.com;mid=492;] mailbox - got blob list for mailbox 492 (55 blobs)
2020-05-10 03:11:35,362 INFO [FullBackupThread] [] backup - Account some-address@ourdomain.com in backup set full-20200510.005505.244: All pending file IO completed (0 out of 0)
2020-05-10 03:11:35,364 INFO [FullBackupThread] [] mailbox - Putting mailbox 492 under maintenance.
2020-05-10 03:11:35,376 INFO [BackupSetWorkerThread-INDEX] [name=some-address@ourdomain.com;mid=492;] backup - Stored 44 index files
2020-05-10 03:11:35,450 INFO [FullBackupThread] [] backup - Account some-address@ourdomain.com in backup set full-20200510.005505.244: All pending file IO completed (44 out of 44)
2020-05-10 03:11:35,450 INFO [FullBackupThread] [] mailbox - Ending maintenance and marking mailbox 492 as unavailable.
2020-05-10 03:11:35,458 INFO [FullBackupThread] [] backup - Account some-address@ourdomain.com in backup set full-20200510.005505.244: All pending file IO completed (44 out of 44)
2020-05-10 03:11:35,458 INFO [FullBackupThread] [] backup - Full backup finished for account some-address@ourdomain.com (c2169c42-1942-47f6-8383-9e6b4fd06e8a) mailbox 492
2020-05-10 03:11:35,458 WARN [FullBackupThread] [] backup - Error occurred during full backup for account some-address@ourdomain.com; aborted backup and deleting /opt/zimbra/
backup/tmp/full-20200510.005505.244/accounts/c21/69c/c2169c42-1942-47f6-8383-9e6b4fd06e8a
java.lang.NoSuchFieldError: revision
at com.zimbra.cs.backup.BackupAccountSession.storeBlobsImpl(BackupAccountSession.java:412)
at com.zimbra.cs.backup.FileBackupTarget$BackupAcctSession.storeBlobsStage(FileBackupTarget.java:1681)
at com.zimbra.cs.backup.FileBackupTarget$BackupSetWorkerThread.run(FileBackupTarget.java:2996)
2020-05-10 03:11:35,475 ERROR [FullBackupThread] [] backup - Error while backing up account some-address@ourdomain.com: revision
java.lang.NoSuchFieldError: revision
at com.zimbra.cs.backup.BackupAccountSession.storeBlobsImpl(BackupAccountSession.java:412)
at com.zimbra.cs.backup.FileBackupTarget$BackupAcctSession.storeBlobsStage(FileBackupTarget.java:1681)
at com.zimbra.cs.backup.FileBackupTarget$BackupSetWorkerThread.run(FileBackupTarget.java:2996)


zmcontrol -v: Release 8.8.15.GA.3869.UBUNTU16.64 UBUNTU16_64 NETWORK edition, Patch 8.8.15_P9.

Any help, please?

[fs.schmidt]

Hello,

Since you are a Network edition user, If you have the support active you could open a ticket with Zimbra.

It seems that the backup for a single account is failing. I would try to reindex this account.

You are using Zimbra 8.8.15. Give the real time backup module a try. It works much better than this legacy backup.

[00a]

This just started happening for us as well. All accounts seem to be affected.
Full backups are not able to complete.

Code: Select all

$ zmcontrol -v
Release 8.8.15.GA.3829.UBUNTU16.64 UBUNTU16_64 NETWORK edition, Patch 8.8.15_P9.

[00a]

After looking at this further - The DB checked out. Mail accounts were fine.

So I think this was due to the last apt update and ZImbra perhaps moving systems toward BackupNG.

So I went into the administration console and initiated that process in a new directory and in the initiation it made a full backup.
The full backups took about 20% less space with BackupNG vs the old full backup.

I couldn't find documentation about migrating from zmschedulebackup to "zxsuite backup." Do you need to run both? Is there a way to check that backupNG backups have taken over ....

[BradC]

The full backups took about 20% less space with BackupNG vs the old full backup.

I've been playing with both backup systems recently. One thing to be aware of from a DR point of view is a restore with BackupNG requires re-creating any ZCO profiles and re-syncing Activesync devices. The Zimbra NE Backup doesn't appear to require this and you pretty much end up with a completely functional server with only downtime to the clients as the obvious "issue".

If you have many Outlook MAPI clients that might increase support load with a DR restore. The same issue comes up with a side-by-side migration, but then the existing NE backup can't do that anyway.

[zimbraLuis]

Hello,

I have the same problem since path9 with 2 account with the old module backup.

Do you have resolve this issue ?

[grantntourbus]

Having this exact same issue. Was working fine before the last patch 8.8.15 P9. Pretty sure it even worked for a few weeks after I patched.

Now I get the same error from the Full Backups:
username@domain.com: revision
java.lang.NoSuchFieldError: revision
at
com.zimbra.cs.backup.BackupAccountSession.storeBlobsImpl(BackupAccountSessio
n.java:412)
at
com.zimbra.cs.backup.FileBackupTarget$BackupAcctSession.storeBlobsStage(File
BackupTarget.java:1681)
at
com.zimbra.cs.backup.FileBackupTarget$BackupSetWorkerThread.run(FileBackupTa
rget.java:2996)

on all accounts.

The incremental backup does not throw any errors.

Would like this Full backup to work, or an explanation as to why it stopped. Is there some conflict with using both backup methods?

[grantntourbus]

I have sent a support request in to Zimbra regarding this error. Will post what their response is.

[justjack]

Hi there!
We have exactly the same problem... Is there any updates? Did zimbra support manage to help you? (I've just registered so I cannot send PMs yet). I would appreciate any help!

[phoenix]

Hi there!
We have exactly the same problem... Is there any updates? Did zimbra support manage to help you? (I've just registered so I cannot send PMs yet). I would appreciate any help!

Why don't you raise a support case?