[BackupPC-users] Prioritising user restore jobs

[Matt B. <m...@wh...>]

Hi all,

This problem doesn't seem to follow my understanding of the config...

I have in my status page:

# 23 pending backup requests from last scheduled wakeup,
# 1 pending user backup requests,
# 1 pending command requests,

The 23 pending requests are a full nightly run which seems to have been
hung by a single host.

The 1 pending user request is a restore direct to the originating host.

The 1 pending command request is a BackupPC_link job.

There is also 2 BackupPC_link jobs running.


In my config I have:

$Conf{MaxBackups} = 4;
$Conf{MaxUserBackups} = 4;
$Conf{MaxPendingCmds} = 10;
$Conf{MaxBackupPCNightlyJobs} = 2;

Which makes me think that the restore job should be able to run straight
away, and if I like I can start more backup jobs. The Max*Bakup limits
haven't been hit yet, so why are my restores queued? Similarly the
MaxPendingCmds has not been hit yet. The NightlyJobs is in line with the
status page.

So the question is how do I make sure that my users can recover files
immediately even if the nightly job overran due to an issue.

And while we're here why did one host hang the rest of the queue? When I
came to the cgi interface this morning there was only one job running. I
canceled it and the 2 BackupPC_link started. I cannot see the command of
the 23 pending backups, so I am presuming they are all link jobs?
Several hosts seem to have recent backups in the last day.

So again... can I force BackupPC to start a user restore job even if
there are queued nightly link jobs >10?

Cheers,

Matt