Viewing 15 posts - 1,126 through 1,140 (of 1,518 total)
I use this:
USE dbName;
SELECTid
,indid
,OBJECT_NAME(id)
,[name]
,rowcnt
,rowmodctr
,STATS_DATE(id, indid) AS Date_LastUpdated
FROM sys.sysindexes WITH ( NOLOCK )
WHERE indid > 0 AND indid < 255 AND
OBJECTPROPERTY(id, 'IsUserTable') = 1 AND -- system tables excluded
INDEXPROPERTY(id, name,...
June 4, 2008 at 8:54 am
sotn (6/4/2008)
June 4, 2008 at 7:49 am
sotn (6/4/2008)
I have a 4GB box SQL2005 Std Edition if I have the /3GB switch and AWE is NOT...
June 4, 2008 at 7:26 am
GSquared (6/3/2008)
Marios, shouldn't your first test have created a unique clustered index, not just a clustered index, for purposes of this test? Or did I miss something?
Good catch, thanks!...
June 3, 2008 at 3:29 pm
Hmm, the results are a lot less striking if instead of a UNIQUE constraint, we specify a CLUSTERED PRIMARY KEY:
CREATE TABLE #t1 (c1 int, c2 int, c3 char(5000));
CREATE CLUSTERED INDEX...
June 3, 2008 at 2:07 pm
Steve Jones - Editor (6/3/2008)
That's an interesting scenario. I haven't dealt with it, but my thoughts are that you enable AWE on all instances. With 2-3GB for each instance, I...
June 3, 2008 at 12:51 pm
george sibbald (6/3/2008) Marios,
looks like it could be an 'it depends ' situation. I have always worked to using the /3Gb switch if RAM under 16GB. This is because 1GB...
June 3, 2008 at 8:30 am
GilaMonster (6/3/2008)
george sibbald (6/3/2008)
so surely /3Gb switch ok up to 16GB RAM, remove for anything over that.:)
Nope. Even under 16GB you risk starving the kernal of memory.
Personally I would...
June 3, 2008 at 7:59 am
colin Leversuch-Roberts (6/2/2008)
x32 -...
June 2, 2008 at 6:12 pm
Stress (6/2/2008)
Marios Philippopoulos (6/2/2008)
What I’ve found in large production systems is that you cannot guarantee that data on every index is held in the...
June 2, 2008 at 6:03 pm
from the blog of Colin Leversuch-Roberts http://sqlblogcasts.com/blogs/grumpyolddba/:
What I’ve found in large production systems is that you cannot guarantee that data on every index is held in the dmvs at all...
June 2, 2008 at 4:02 pm
You may well have stumbled on a bug here -
I will do some more searching on the web for similar info.
June 2, 2008 at 2:56 pm
I think sys.dm_db_index_usage_stats is sensitive ONLY to index operations in the WHERE clause:
eg. index operations taking place in a JOIN are not counted.
June 2, 2008 at 2:51 pm
Stress (6/2/2008)
I have seen lots of examples on the 2005 index usages DMV, but one thing has eluded me : it only seems shows access stats for the dbo....
June 2, 2008 at 2:22 pm
colin Leversuch-Roberts (6/2/2008)
June 2, 2008 at 12:59 pm
Viewing 15 posts - 1,126 through 1,140 (of 1,518 total)