What comes after a baker’s dozen of cumulative updates? CU14, that’s what, and it includes a really long list of fixes. I’m only gonna hit the highlights here:
- Non-yielding scheduler when there is a “heavy use” of prepared statements
- Columnstore index build times out after 25 seconds
- sys.fn_hadr_backup_is_preferred_replica returns true for multiple replicas when there can be only one
- Access violation and the server stops unexpectedly when you use Extended Events (SEE ERIN THIS IS WHY I’M #TEAMPROFILER)
- Assertion failure when a trigger does a cross-database transaction in an Availability Group (repro video)
- Patching fails if the SA account is renamed
- “Query results are not as expected when you run a particular query” LOL
- Corruption-sounding error when you back up to a drive with low space
- Filtered nonclustered columnstore indexes on top of a clustered columnstore index may not be maintained correctly (Xzibit called)
- Trace flag 460 throwing false errors on table variables (just blogged about this)
- DBCC STACKDUMP doesn’t
