Hi,

I've been using bacula to backup a number of clients on Rocky9 using
maria 10.5 as the database, and a file based storage. I have 5 clients
with the largest Full backup being 1.1GB and the other a few hundred
MB.

Typically full backups take less than 5 minutes and the
differentials/incrementals take 5-20 seconds.  I recently upgraded the
director from version 11.0.1-5, to the community 13.0.3 package, and
I've started seeing intermittent hanging jobs taking up to an hour.

(I am not entirely sure this started happening after the upgrade,
because I just started paying attention, but it seems so)

Basically the jobs get stuck like this:

Console connected using TLS at 21-Dec-23 02:13
 JobId  Type Level     Files     Bytes  Name              Status
======================================================================
   426  Back Diff         87    142.1 M Backup_bhs Dir inserting Attributes
   427  Back Diff         18    11.31 K Backup_lpl Dir inserting Attributes
   429  Back Diff         10    16.74 K BackupBaculaConfig Dir
inserting Attributes
====

inspecting the bacula and mysql server during the slow jobs, I can see
no particularly high io or cpu, except that the mysql server has 1 CPU
stuck at 100% and there is a long running query:

SELECT Path.Path, File.Filename FROM File JOIN Path USING (PathId)
JOIN b21197077 AS T ON (File.JobId = T.JobId AND File.FileIndex =
T.FileIndex) WHERE File.Filename LIKE ':component_info_%' AND
File.JobId IN (103,419);

After killing the db session, the temporary table still exists, so i
was able to run it from the console, and it does seem to take a long
time;

MariaDB [bacula]> SELECT Path.Path, File.Filename FROM File JOIN Path
USING (PathId) JOIN b21197077 AS T ON (File.JobId = T.JobId AND
File.FileIndex = T.FileIndex) WHERE File.Filename LIKE
':component_info_%' AND File.JobId IN (103,419);
Empty set (51 min 43.039 sec)

The explain looks like this:

MariaDB [bacula]> EXPLAIN SELECT Path.Path, File.Filename FROM File
JOIN Path USING (PathId) JOIN b21197077 AS T ON (File.JobId = T.JobId
AND File.FileIndex = T.FileIndex) WHERE File.Filename LIKE
':component_info_%' AND File.JobId IN (103,419);
+------+-------------+-------+--------+---------------+---------+---------+--------------------+-------+------------------------------------+
| id   | select_type | table | type   | possible_keys | key     |
key_len | ref                | rows  | Extra
   |
+------+-------------+-------+--------+---------------+---------+---------+--------------------+-------+------------------------------------+
|    1 | SIMPLE      | T     | ALL    | idx_b21197077 | NULL    | NULL
   | NULL               | 48180 |                                    |
|    1 | SIMPLE      | File  | ref    | JobId,JobId_2 | JobId_2 | 4
   | bacula.T.JobId     | 826   | Using index condition; Using where |
|    1 | SIMPLE      | Path  | eq_ref | PRIMARY       | PRIMARY | 4
   | bacula.File.PathId | 1     |                                    |
+------+-------------+-------+--------+---------------+---------+---------+--------------------+-------+------------------------------------+
3 rows in set (0.000 sec)

Any idea what could be causing this?

Thanks
Tom


_______________________________________________
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users

Reply via email to