Hi! I’ve got a system with over 234 million records in the service_alarm table in MySQL. I believe this is causing the DBs to sync up when creating failover. Is it safe to completely drop this table? Also, what’s a good method of making sure the table doesn’t get to this point in the future? Thank you!
Hey Luis,
That table is used keep the link of an alarm that has had an impact on a service. It is used to populate the “Services” column in the alarm console and to allow filtering on alarms based on the service information.
From an stability point of view there is no issue in truncating this table, however the above described functionality of viewing historic alarms will no longer contain the service information. Vice versa requesting alarmhistory on a service will no longer include these deleted links.
Since this is a dataminer internal table the only way to keep this healthy is to essentially make sure that the cleaningthread is doing its job in cleaning out this table when related alarms or services are deleted (no timeouts or errors when the query gets launched see SLDatabase[SLDataGateway]/-[SLDataMiner].txt ). Another way is to lower the amount of alarms that can impact a service by reviewing the alarm thresholds or the element inclusions in a service.