I got caught out on this one while doing an incremental migration, the disk where /opt/zimbra lives was filling faster than I expected. My bad, https://wiki.zimbra.com/wiki/Zimbra_NG_ ... rk_Edition gives clear directions about what to do before kicking off an import.Both exist in the 8.8.8 product, but when you elect to install the NG component when running the Zimbra installer, no crontab entires for the Traditional NE Backup are set, so it never runs. This also causes an issue IMHO if you run just the BackupNG backups in that /opt/zimbra/redolog/archive never gets emptied. Run an imapsync to migrate 500GB of email from a new customer and you'll find the redolog archive folder has 500GB of redo logs in there. I feel this is a configuration error, so I opened a Support Case as an RFE for it. In the interim, you can run "zmprov mcf zimbraRedoLogDeleteOnRollover TRUE" and restart the mailbox service, and then delete the existing archived redo logs as BackupNG does not use the archived redo logs at all I am told.
The fix was easy (as Mark describes above), re-ran the data import and all was well.
Question: If I'm sticking with Zimbra NG Backup, do I want to keep zimbraRedoLogDeleteOnRollover set to TRUE indefinitely?
Question: From what I can tell the /opt/zimbra/redolog/archive ONLY gets emptied when a traditional backup job completes, is that true?