Old backup files not deleted

Hi,
I’ve been running miab for 18 months now. Great!
Today for the second time (in a row) I have found that after a full backup, the previous full backup and incrementals weren’t deleted.
The result is I ran out of space on the place where I stash my remote archives.

Any ideas what might have changed to cause this, or where I should look for clues?

It is a minor annoyance, and to be honest the setup has been very stable, and so incidents like this give me an excuse to give it the once over more often than I would otherwise do. However the worry is that it is symptomatic of a more important issue.

Thanks

Hello stefbishop,

The following explanation can be found under System -> Backup Status -> Configuration -> Days:

This is the minimum number of days backup data is kept for. The box makes an incremental backup, so backup data is often kept much longer. An incremental backup file that is less than this number of days old requires that all previous increments back to the most recent full backup, plus that full backup, remain available.

I am not sure if this is the proper explanation, but we can see that no maximum age is defined and incremental backups are chain dependent. Perhaps the chain broke or something else required the need for a new full backup. Regardless, I would expect a full backup overlap to occur while MiaB waits for the older backups TTL to trigger.

Have you reviewed the ‘Deleted in…’ column to confirm approximately when the older backups were to be deleted?

Strange, this should never happen if you are using the rsync or S3 methods. rsync would either ignore or overwrite previously transferred files. S3 should never run out of available space. How are you transferring and storing your backups?

Hi, many thanks for your response.

I never checked to see what the backup status said before I manually deleted the outdated backup files from /home/backup/encrypted. So I will next time!
The rsync to the offsite storage was doing what it should. It’s just that there wasn’t enough room for 2 backup cycles, but if there was I would never have been alerted!

Since I’ve not made any changes to the system for many months, I can’t see that the behaviour is likely to be a result of a recent change I’ve made. However the hosting provider did migrate the running system some time ago. Maybe as a result of that there is a memory pointer that has become invalid.
Probably I should just restart miab.

In the meantime it might be worth other users checking what is in /home/backup/encrypted to see if there is only 1 backup cycle and that the backup status report accurately reflects that content. (Since light loaded systems with acres of storage might accumulate a lot before anyone notices!)