To resolve this issue, these changes are recommended:
- Implementing new rollup Jobs (one combined, one per vCenter Server).
- Disabling daily, weekly, and monthly rollup jobs:
- Create a new rollup job per vCenter Server
- Add a step for each stored procedure of database with Type Transact-SQL script:
- Past day stats rollup
- Delete past day rolled up records
- Past week stats rollup
- Delete past week rolled up records
- Delete usage stats rollup
- Past month stats rollup
- Delete past month rolled up records
- Schedule rollup jobs so they do not overlap. Set a five minute gap between each vCenter Server rollup job.
An example schedule pattern:
1am – vCenter1, 1.05am – vCenter2, 1.10am – vCenter3….. 2am – vCenter1, 2.05am – vCenter2
- Alter the rollup job schedules again by not interfering with the weekly index maintenance job that runs on the weekend.
Note: To dramatically reduce the chances of database locks, combine VMware default rollup jobs into one job.