I'm installing vCenter 4 for production use and have noticed a error.
In the VCDB, there are 3 scheduled jobs created for stats rollup. The Past Week one seems a bit odd.
Its created by the sql code job_schedule2_mssql.sql
It creates 3 steps, ulike the other 2 rollup jobs, but the code runs like this:
Step 1 completes and goes to the next job having run EXECUTE stats_rollup2_proc
Step 2 completes and quits with success having run exec purge_usage_stat_proc
Step 3 exec purge_stat2_proc never runs.
If you use SQL 2005 and slick on the steps tab and move to another tab, SQL warns you about the flow logic problem.
Ive just checked and it was also in the RC Sql Agent.
Noticed this as well when going through the roll up jobs. I'm assuming this is an accident in the release of the product and the logic needs to be corrected in step 2 to proceed to step 3.
I also have this issue and I think it may be causing some other problems within the db.
Here is the structure from vCenter Server 4.0.0 Build 258672
Daily
EXECUTE stats_rollup1_proc
exec purge_stat1_proc
Weekly
EXECUTE stats_rollup2_proc
exec purge_stat2_proc
exec purge_usage_stat_proc
Monthly
EXECUTE stats_rollup3_proc
exec purge_stat3_proc
Initially the vCenter Health status was reporting:
Performance statistics rollup from Past Day to Past Week is not occurring in the database
Performance statistics rollup from Past Week to Past Month is not occurring in the database
Our transaction logs were getting filled within a couple of hours - (60Gb), so had to change the recovery model to simple and do a shrink. I think the scheduled db jobs were failing due to the sizes of the db's involved.
Once I have cleared out some old data with this script (warning it's slow: 6 million rows in 10 hours), the jobs have run and the errors cleared on the vCenter health page. The transaction logs are also not being generated at the same rate.