This time on Office Hours, I let a few questions piled up at https://pollgab.com/room/brento that required in-depth answers to really do ’em justice. In particular, there was a statistics question that needed demos.
Here’s what we covered:
- 00:00 Introductions
- 00:35 Grogu: Did DBA Brent ever have to support a version of SQL Server that Microsoft had end of lifed? Any tips on discussing this issue with management?
- 04:19 Ahsoka Tano: What are your favorite PASS Summit presentations from prior years?
- 07:11 Can: I have 3000 DB per SQL Server. When I need to restart SQL Services, the service is waiting on Stopping mode.
- 09:22 i_use_uppercase_for_SELECT: When doing filtered statistics do you basically have to use full scan? It appears the where clause for filtered statistics is applied after the tablesample system is applied so you end up sampling the same data as the main statistics which wasn’t seeing this data originally.
- 29:02 Shaheen: Is there info in SQL Server that will tell us why it picked one index over another for a given query?
- 31:58 muppet#1: What’s your favourite SSMS plugin (assuming you use any)?
- 33:02 DBAInProgress: What are some of the major gotchas of using TDE?
- 34:22 Dmitriy: How would I begin to optimize a reporting query that has no WHERE clause, but has a bunch of LEFT JOINs to a bunch of other tables via foreign keys?
- 35:55 Columnstore newbie: I have a clustered columnstore index. The segments look good but selecting TOP 1000 order by UtcDate is still very slow. Why?
