Hello,

we're using auto grouped backups to backup a large zimbra installation (zcs-NETWORK-7.2.0_GA_2669.UBUNTU10_64.20120410002323).

We write the backup directly to a NFS mounted NAS, which usually works fine. Now on one of the boxes just stopped during the backup. Last log line:

2013-11-15 07:04:33,877 INFO [FullBackupThread] [name=XXX;mid=18131;] backup - Number of blobs to backup for mailbox 18131: 146
2013-11-15 07:04:33,892 INFO [FullBackupThread] [name=XXX;mid=18131;] mailbox - Ending maintenance on mailbox 18131.
2013-11-15 07:04:33,892 INFO [FullBackupThread] [name=XXX;mid=18131;] mbxmgr - Mailbox 18131 account 99b5723d-86dc-42f7-b792-889b4ef0eed5 AVAILABLE <- last written line of FullBackupThread.

There is nothing obvious in the logs behind this point in time. I believe there was a long running stop-the-world full GC at this time (load peaked).

A complete run looks like this:

2013-11-14 12:28:58,288 INFO [FullBackupThread] [] backup - Finished backing up 3170 of 3170 accounts
2013-11-14 12:28:58,819 INFO [FullBackupThread] [] redolog - Redo log rollover took 531ms
2013-11-14 12:29:10,239 INFO [FullBackupThread] [] backup - Backup set full-20131114.000020.791: All pending file IO completed (0 out of 0)
2013-11-14 12:29:10,239 INFO [FullBackupThread] [] io - AsyncFileCopier is shut down
2013-11-14 12:29:10,239 INFO [FullBackupThread] [] backup - Full backup finished for backup set; label: full-20131114.000020.791
2013-11-14 12:29:14,406 INFO [FullBackupThread] [] backup - full backup took 41332879 millisec

The incomplete full backup remains in backup/tmp also.

How to proceed now? Can i just remove the incomplete full backup and the next run will cover these accounts again?

Regards,
Dirk