I've got an issue where a join takes about 3 minutes to run on a development server (2 X 1.4 GHz/ 2GB / Win2K3SP1 / SQL2KSP4 (8.00.2187) standard) and does not finish (even after 10 Hours!) on a monster (8 X ? GHz/ 12GB / Win2K3SP1 / SQL2KSP4 (8.00.2187) clustered Enterprise) data center test machine.
This is a join between a table with 360K rows and another with 5.6 Mil rows. Identical data, schema, indexes, auto-stats, etc. between the two boxes. I see no CXPackets, so the parallel processing issue doesn't seem to apply.
Any ideas what to look for?I'm not a SQLserver user, so this is just a guess from my DB2 background.
Apparently the optimizer chose different access paths in both cases, despite of identical indexes and statistics. Maybe the optimizer got (badly) influenced by the higher performance of the second box?
I guess you can ask the systems for their respective access paths?
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment