Known Issue: Slow startup on systems with many alerts

We have become aware of an issue that affects systems that store alerts to the alert log, and results in gradual increase of system startup time.

The issue is caused by the lack of a database index on the CLEAR_TIME column of the alert log table. It can be fixed by creating an index on that column, which can be performed through SQL or the administration tool of the database system in use.

This problem is most often encountered in the form of SQLTags reporting STALE status for some amount of time after the system has been restarted. However, it can also affect groups, causing execution to halt until the alert log operations have completed.

Again, the solution is to create an index on the CLEAR_TIME column of the alarm log table. This can be accomplished with the following query (for MySQL, other database systems may have slightly different syntax):

CREATE INDEX clear_ndx ON alarmlog(clear_time)

Note that the alarm log table name may be different on your system. Also, it is advisable to first verify through an administration tool that the column is not already index on your system.