I’ve got a job that runs on our slave servers to backup a few databases to an NFS share. Every so often, we get a bad backup, usually an incremental, which causes a core dump when we attempt to apply logs. (We wait a week for this just in case.)
My only question here is, once that incremental rollup fails, is the full backup to which we’ve applied a partial recovery now effectively broken? The DB itself seems perfectly healthy, and subsequent backups and apply-log attempts seem to work, so I think I’m fighting off an IO problem more than anything else.
I just want to know if I’m sitting on garbage data at this point, because I can at least tweak my backup schedules to reduce IO contention on the NFS share and avoid these problems, and delete anything that’s no longer useful to me.