regarding seconds behind master variable in 10.0.4
hi we setup maria 10.0.4 for multisource replication.
it was replicating from a percona 5.5 master. When we started replication, maria was a day behind its master.
its working fine but we noticed this phenomenon:
when we execute 'show all slaves status' consecutively the 'seconds_behind_master' flips from 0 to XXX and XXX to 0. XXX keeps decreasing though, which is good as its getting closer to the state of its master.
But wanted to know why the 'seconds_behind_master' comes down to 0 then again sets to a positive number.
Note: XXX is some positive number.
Thanking you. vishnu rao
Answer Answered by Elena Stepanova in this comment.
Indeed, I also observe seemingly random numbers in
seconds_behind_master field. However, it doesn't seem to be related to multisource replication, it happens with a single master as well, and is also reproducible with MySQL 5.5 and MariaDB 5.5. Thus, I think it's most likely bug http://bugs.mysql.com/bug.php?id=52166 or http://bugs.mysql.com/bug.php?id=53167, which were fixed in 5.6. We need to either merge the bugfixes or fix them on our own. Could you please submit a bug report at JIRA?