increase in waits when index monitoring is on

From: Dba DBA <>
Date: Wed, 8 Feb 2012 14:35:47 -0500
Message-ID: <>

DB Version:
I want to turn index monitoring on in spots to see if some of our indexes are being used. We have an very active OLTP, so before doing that I am running some tests of index monitoring.
Has anyone else noticed this? This is fairly large for the simple tests I am running

tool: runstats
I have a database all to myself

Using one of our tables.

2 indexes
1 column non-unique
4 column unique

I made a copy of this table with the same indexes. So one table would have index monitoring turned on and one did not have index monitoring turned.

The increase in waits happens on inserts. I am inserting 50 records at a time as a test.

I have run this test many times and the waits are consistent. I also dropped and created new tables. This increase is consistent. I also change the test so that the 2nd insert is to the table with index monitoring.
I am not turning index monitoring on during the test. I have 2 different tables. One with index monitoring and one with out (doing that would skew the results)

Run 1 is on a table with index monitoring on 1 non-unique index Run 2 does not have index monitoring on.

  LATCH.library cache lock               214          32        -182
  LATCH.row cache objects              321          93        -228
  STAT...recursive calls                     304           2        -302

Received on Wed Feb 08 2012 - 13:35:47 CST

Original text of this message