From: <cba...@us...> - 2003-04-30 06:41:14
|
> I'm trying to figure out why all our backups stop once one of our = > computers receives signal=3DALRM. The ALRM signal was received because = > the user turned off his computer while the backup was running. > > Should the other computers still be backing up ok? It seems that all = > operations have halted. > > The BackupPC_dump process is still showing as running for the computer = > that received the ALRM signal. Any ideas? > > We are running 2.0.0beta1. I believe this is a bug in 2.0.0beta1. A canceled backup doesn't always get cleaned up by BackupPC. It thinks it is still running, but it really has been canceled. As Doug points at, at that point BackupPC_nightly can't run, and new backups won't start until BackupPC_nightly runs. (I actually changed the code in this area for 2.0.0beta1 because I was fixing the opposite problem: BackupPC sometimes wedged when trying to cancel a backup that couldn't be killed.) Until the bug is fixed the only recourse is to stop and restart BackupPC. If this happens often and is annoying, look in bin/BackupPC and remove these three commented lines around line 1030: ##vec($FDread, $Jobs{$host}{fn}, 1) = 0; ##close($Jobs{$host}{fh}); ##delete($Jobs{$host}); Then restart BackupPC. Craig |