I am seeing that none of the trend tables for any of my Service elements are getting pruned of old data. Can someone please confirm that the DataMiner dbase cleaning process does NOT attempt to clean out data from MySQL trending tables associated with Services. This is a Major issue for many of our installations. It is normal for our systems to have a thousand or more services and these services have a Service Definition Trend template assigned to them with both RT and Avg trending enabled. The database in many of these systems has grown so large, sldmadb folder > 500G in size, that Backups and Optimizations can take days to finish and are now failing and causing the servers to run out of disk space due to the temporary uncompressed sql backup file exceeding the size of the free space on the drive. In looking deeper into this I notice that dbase cleaning thread does not appear to be attempting to clean data from these Service element tables so I have thousands of tables that continue to grow in size.
Thanks
Hi Jeff,
We confirm indeed that we have an open issue for:
- Built-in aggregator elements (supporting the Aggregation rules on Views)
- Enhanced Service elements (so services with an enhanced service protocol linked to it)
on agents that still use MySql as local database.
The impact of this issue is indeed that old data does not get cleaned up like it should.
DataMiner Systems that have Cassandra as local database are not affected.
Regular DataMiner elements on DataMiner Systems that have MySQL as local database are also not affected.
In order to restore from this, I believe it’s best to not wait on a fix, but try to delete the necessary tables through manual queries (e.g. via an automation script iterating over the aggregator and service elements).