September 11, 2009 at 12:56 pm
Wonder if anyone has a comment on this. I inherited a 2000 Standard recently. Was on a single Raid5 set with everything else (OS, shares, etc). 2 disks were lost in the Raid. Was able to get data back through force disk online however lost Master and the backup of Master was too old to feel comfortable with. No problem I thought. All new harddrives, 15k SCSI instead of 10k and twice the size. Put OS on Raid1, TLog on Raid1, Data on Raid5.... front end is Access/VBA... nothing changed on Front End.. but now with the system back up I have an abnormal amount of blocking. The application hasn't changed. Thanks ahead of time for anyone's input.
Jim
September 11, 2009 at 3:54 pm
maybe transaction isolation level has changed ?
September 12, 2009 at 10:25 am
Thank you. As far as I know this is not a global option on SQL Server 2000 but is specified at the query / connection level. Can you enlighten me if I am incorrect ?
September 12, 2009 at 10:46 am
Locking or blocking? The disks have changed so maybe some problem with their configuration? What is the most common wait type, is it IO related?
run perfmon see what response times you are getting in physical disk (logical disk if SAN based)
---------------------------------------------------------------------
September 12, 2009 at 3:33 pm
Hello. Blocking. I ran perfmon, physical counters seem in acceptable ranges. I did find during review today that I have some orphaned users though the server didn't overtly complain about this (and these users own objects). (SID is different in master than it is in the UserDB) Do you think this could impact the server to a point where it has blocking issues? Overall, SQL is running a little bit slower than before. Since I have the weekend I'm almost thinking it wouldn't be a bad idea to just create a new db shell, with new sid's recreate the objects and reload the data.... Thank you again for your feedback and comments.
Maybe compile blocking?
September 12, 2009 at 4:21 pm
Just wanted to say thanks.... the Blocking vs Locking.... you tipped me in the right direction here... and the system "seems" to be fixed. We'll see when the users get to it but it's noticeably faster. Flying as a matter of fact.
September 13, 2009 at 3:01 am
good to hear. what do you think actually fixed it? orphaned users would simply not have been able to connect rather than cause blocking.
---------------------------------------------------------------------
September 13, 2009 at 5:45 pm
The blocking, the sid's, strange #'s for sid's... I recreated the db and all objects, repopulated, etc.... I suspect something was corrupt and I didn't put my finger on any one thing... the db didn't even complain about the sid's/orphaned objects which I found odd. I do appreciate your feedback which is always encouraging.
Viewing 8 posts - 1 through 7 (of 7 total)
You must be logged in to reply to this topic. Login to reply