Regular Advisor
Keeping OML active message table under 50K messages (best practice)

Hi,

 

   We ran into an issue while I was out for an extended period in which we could not login using the Java GUI to OML.  We'd get a message indicating that the connection attempt timed out.  It was happening to our primary and backup servers.  I pointed out the possibility of it being because of so many messages being loaded and support advised that wasnt the issue.  When I eventually deleted the active messages in the backup server and then the primary (after downloading the critical and major messages), I was able to log back in.

 

   I found out why the active table filled up so fast (2 nodes out of control) but in the meanwhile was wondering if anyone could advise on a best practice for keeping the active message table under 50K messages.  I know about alerts that will appear in the events browser but I'm looking for something we can cron that can run regularly.  Commands like opcack just do all and don't have date filtering.  Others, like opcactdwn, have date filtering but wont help in this scenario. 

 

Thanks in advance,

Heidi