Closed adamdriscoll closed 2 months ago
The problem is the size of the JobOutput table. Since changing to a row per output message, the table causes jobs to delete very slowly because of all the cascading deletes. We've implemented nibbling deletes on the jobs table to account for this. This will delete jobs in chunks and allow other queries to process in between the chunks.
Version
4.3.0
Severity
Low
Environment
msi
Steps to Reproduce
I had a jobs table with 250 records in it and ran the server-wide groom command manually. It took 8 seconds to complete. This was causing locks on the table that was causing jobs to fail to run.
Expected behavior
Actual behavior
Additional Environment data
No response
Screenshots/Animations
No response