Re: Bugzilla database problems
- From: "Owen Taylor" <otaylor redhat com>
- To: "Ross Golder" <ross golder org>
- Cc: GNOME Infrastructure <gnome-infrastructure gnome org>
- Subject: Re: Bugzilla database problems
- Date: Thu, 20 Sep 2007 08:58:21 -0400
On 9/20/07, Ross Golder <ross golder org> wrote:
> Getting messages like "undef error - DBD::mysql::st execute failed:
> Table './bugs/longdescs' is marked as crashed and should be repaired" on
> bugzilla.
>
> Running 'mysqlcheck' on that table, but it might take a while...
That usually seems to fix things.
If someone has time, they might look into the operation of the our
scripted mysqlhotcopy in this case ... it doesn't seem to have worked
100% as intended.
/var/lib/mysql-backup/bugs: empty
/var/lib/mysql-backup/bugs_old: database as of 36 hours ago
The fact that the backup is 36 hours old isn't too surprising ... when
the backup ran 12 hours ago, it failed for the same reason as above,
but I'm a little concerned that
/var/lib/mysql-backup/bugs_old didn't get moved back to
/var/lib/mysql-backup/bugs.
That makes me think that the *next* time the backup script runs in 12
hours it will either:
A) Wipe out bugs_old and the last known good backup (could still be
restored from tape)
B) Not do a backup at all until someone manually deletes the bugs_old directory
It's unlikely that a failing bugs database would not get attention for
24 hours, but it's very possible for some of our other databases.
- Owen
[
Date Prev][Date Next] [
Thread Prev][Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]