[sr-dev] git:master: db_cluster: if a db query fails, mark connection inactive

Juha Heinanen jh at tutpro.com
Wed Apr 11 14:46:56 CEST 2012


Daniel-Constantin Mierla writes:

> I hope to be fixed as part of my most recent commit. The condition to 
> reactivate the connection after inactive_interval was wrong.

yes, now my tests show that db_cluster properly switches to lower
priority server when higher priority dies and then after 5 minutes back
to higher priority server when it is up again.

-- juha



More information about the sr-dev mailing list