Failed disk question

general discussion
1

Hi,

I’ve got Urbackup installed on a Ubuntu x64 VM, the O/S is installed on a 8GB Parition along with the Urbackup server. The datastore for target backups is a 300GB disk mounted as /mnt/backups/. Recently the 300GB drive started causing problems and failed. I swapped this out and installed a new disk. The problem appears to be now that the database remembers previous backups and expects to find them under /mnt/backups however the new disk is blank. Triggering full and incremental backups seems to copy some files but I am not sure that it’s doing a new full where the previous files were. I’m assuming the logic is something along the lines of:

  1. List files on the client
  2. Calculate check sum of file on client and compare this with what the server has.
  3. If this exists on server adjust sym link current date directory else transfer file and point to it.

So with the above am I safe to assume that the next incremental will fill in the blanks for all missing data from the client?

If this is the case it might be nice to have Urbackup do some sort of self checking / house keeping in the clean up window to get rid of old backups that no longer exist from the database?

I also noticed that DAEMON_TMPDIR is defined twice in /etc/init.d/urbackup_srv so (in my case) the primary setting wasn’t being respected, it needs to be changed in two places.

Thanks.

2

Thanks for the hint on the init.d . If you have a /etc/default/urbackup_srv you should change the tmp dir there btw.

The cleanup is a good suggestion. It just has to somhow detect, when the backup storage is only temporarily unavailable (disk not mountable or something). In this case it shouldn’t delete all info about backups.

Currently it will download all missing files from the client, even for incremental backups.