we've had a few issues recently whereby the windows xp64 supervisor appears to become unresponsive. Looking through the logs it appears, for whatever reason, the supervisor service is having problems updating the mysql database. the errors look as below:
QbDatabaseMysql::lock(): GET_LOCK timed out [SELECT GET_LOCK('qube.951job',10)]
this usually manifests when clients are trying to kill jobs or similar. At this point we usually end up having to restart server or similar in order to clear the problem. Only then will qube allow further jobs to be submitted, modified etc..
Looking at the mysql logs we occasionally get entries as below
100511 11:27:53 [ERROR] Fatal error: Can't open and lock privilege tables: Table 'mysql.servers' doesn't exist
although i am not sure whether they are related.
I'm not that savvy with mysql but would have thought that the only user who could edit the mysql tables would be the qube service itself. All client systems could be using the qube_readonly user to parse the sql db direct for gui information but this shouldn't lock the database for writing in my opinon.
Can anyone tell me why this is happening? Is it possible for the qube service to have stale sql sessions which are stuck open thus leading to the problem or does the service only ever open one session, edit the tables, then close (or perhaps fail to close)?
