Currently we are facing an issue with the alarms flowing into the UIM database. I checked in the USM in UMP and the alarms are delayed. I found the nisqueue.db file size is too big. After restart, the CA UIM Primary Hub the file size was reset but it grows again. How do I know which alarms are stuck in the nisqueue.db file?
- Database settings
Release : 8.5
Component : UIM NAS
The database team executed a change on the CPU Parallelism on the database server because there were too many sessions running which were competing against each other for the CPU.
After changing the database configuration, the alarms were no longer being delayed. The customer monitored the nisQueue db file, and the size increased until 76kb then decreased to 12kb.
The DBA disabled query parallelism on the SQL Server.
They did it ONLINE by setting the SQL Server wide parameter "maximum degree of parallelism" (maxdop to 1). Setting the MAXDOP option to 1 will prevent using more than one processor in executing the query, which means that the SQL Server Engine will use a serial plan to execute the query
Previously it was set to the default of 0. 0 = "unlimited paralellization" and sometimes that can lead to problems.
Parallelism in SQL Server Execution Plan