mdw_purge_data_ long runtime
-
31-10-2019 - |
質問
I have a SQL Server 2008 R2 that is running several databases and has also been configured as our central MDW.
When the management data warehouse is created it creates a number of agent jobs one of which is mdw_purge_data_[DB NAME]
that goes through and purges data that exceeds the retention period, as far as I am aware. This job is currently running and its duration has increased form a few hours to days. This was initially an issue in RTM but was supposedly fixed in Service Pack 1 (we are running SP2). I have tried following an amendment of the stored procedure it calls as per this link:
http://www.toddbaker.org/blog/2010/12/17/sql-2008-mdw-fixing-long-running-purges/
This does not seem to have helped. I then read another blog post suggesting creating two indexes on the sql_handles
columns of two tables in the MDW, snapshots.query_stats
and snapshots.notable_query_text
. When I try and create these I end up with a timeout and the index fails to create. This suggests that there is far too much information in there.
Has anyone got any pointers about reigning in the mdw_purge_data_ job
as I think it may well be impacting the performance of the server as a whole.
正しい解決策はありません