Paul Randal (8/14/2009)
ok - in that case if it only happens under load with concurrent executions of the SP, then I'd say it's definitely a bug. The index rebuild code is interacting with the new allocation code in tempdb (that my team wrote) that does temp table 'caching'. If you're really motivated, call into Product Support and tell them I think this is a bug. Get me an SR number and I'll help it get escalated past frontline support.If you're not motivated, take out the DBREINDEX - I think it's completely superfluous.
Let me know which way you're going to go - I'll escalate directly to them dev team if you're not going to call support.
Thanks!
Paul,
Thanks a lot for your help here. I'm motivated to get a result as I'm keen to improve my SQL skills and it would be great to know that this isn't my code (or code that I support) which is at fault. Having said that, I'll be away from work for a few weeks starting from a few hours from now and have to get stuff finished. I'm not sure who in my organization would be responsible for liaising with MS on this issue (possibly me, which would take yet more time I can't afford right now).
So if you were to escalate it yourself that would be great, and I've already bookmarked this page so I can keep an eye on any updates. And I'll try and produce a cut down SP which reproduces the problem without needing loads of other code/data which I can't republish.
Thanks again,
Alex.