Hi
We have recently migrated from SQL 2012 to SQL 2016 on new more powerful servers. Now we started to experience huge "Estimated Recovery Time" and "Redo Queue Size" that takes most of the day to recover from. Estimated Recovery Time is often around +30000 seconds.
We have set up a job that monitors the "Redo Queue Size" and know that the problem appears shortly after a nightly index reorganize job starts.
We never used to se this on the old servers, so we just don't understand why this happens? It seems to be an SQL 2016 issue.
Setup information:
HP Proliant DL580, 4 x CPU E7-8893 V3, 2TB RAM, 2 x 2TB NVME Intel fusion card. Dedicated 10gbe network for Sync.