From: Martin S. <ma...@li...> - 2012-03-22 12:12:23
|
>>>>> On Wed, 21 Mar 2012 23:01:02 +0100, Lucas B Cohen said: > > Dear list, > > I'm struggling to finish my last round of backups with Bacula 2.4.4, > which I care to do before migrating to the latest version. My problem is > simple to describe: with one of my (multiple, and otherwise working) > media pools, the director fails to find a usable volume for use with a > backup job (thus prompting for the labeling of a new one, which I really > want to avoid). > > I've spent the evening going through debug output and part of the source > in src/dird/next_vol.c and src/cats/sql_find.c. Here is what seems like > the part relevant to the problem of what I'm witnessing in the director > debug output when I ask to run a job: > > Washington-dir: sql_find.c:323-0 fnextvol=SELECT > MediaId,VolumeName,<snip> FROM Media WHERE PoolId=5 AND MediaType='LTO2 > Ultrium' AND Enabled=1 AND VolStatus='Append' ORDER BY LastWritten IS > NULL,LastWritten DESC,MediaId LIMIT 1 > Washington-dir: sql_find.c:331-0 item=1 got=0 > > 'got=0' means the query returned 0 rows. But copying and pasting the > exact same query in fact returns the expected result of the tape I'm > expecting. I can only wildly speculate about what could cause numrows or > sql_num_rows(mdb)'s return value to not reflect the fact that there is > in fact a result returned by the query. Seems very strange. Is it repeatable, i.e. does the backup still fail to find a volume after your manual query has returned that volume? __Martin |