Рет қаралды 1,657
I went through your top-voted questions from pollgab.com/room/brento and I didn't have answers for a few of 'em. Here's what we covered:
00:00 Start
00:48 MyTeaGotCold: I increased my RAM from 16 GB to 32 GB and my buffer pool has grown, but my PAGEIOLATCH_SH waits went up by two percentage points. What might that indicate?
01:46 DavDBA: Hey Brent, Sp_Blitz is recommending that I change the target recovery interval from 0 to 60. I want to implement this, but I read in the Microsoft documentation that it can cause extra I/O activity on certain systems. Have you encountered such issues?
02:32 John: A friend of mine has several hundred SQL AG’s used for DR, not HA. These are all manual failover, async commit. They want to remove WFSC, implementing Read-Scale AGs instead. Is this risky on such a large scale? Are there serious gotchas to be concerned about?
03:30 WindowsMillenniumWasAnInsideJob: Conspiracy theory: Microsoft won't take performance issues in SQL Server seriously until Azure SQL DB won't have majority of the market share (vs boxed editions). Only then Microsoft will fix them because it will be in their interests to do so. What do you think about it?
04:55 SQL_TheOcean: Is there a way to copy the Managed instance databases across a different region in Azure.
05:32 Eduardo: What is your opinion of Natural language to SQL query in Azure SQL DB? Is this a game changer?
06:45 Steve E: Hi Brent, When query tuning, I often find myself trying to fix a bad estimate and wondering how SQL Server made an estimate. Is there a trace flag or tool that shows how SQL server came up with estimates on plan operators - what stats it used, what formulas it applied etc. Thanks
07:32 ConfusedDBA: My friend is using spatial index on a table with less than 10k rows. He does not have a problem with this table but often times he hears that the usage of spatial index/data type will cause issues. Are those remarks valid? If yes, what kind of problem he can expect in the future?
08:23 SQLBaller: Have you seen it where when running a stored procedure it will complete in two seconds? However, if you run that same stored procedure with include actual execution plan it takes ten minutes to run.
09:36 Need4Speed: Is there a way to optimize a table strictly used for inserts of archival records? The table currently has 62 million records and only has a primary key. I have noticed that my process for inserts is running a little slow and was wondering if there was anything I could check.
10:21 ChompingBits: My friend has a huge vendor owned application. The vendor's configuration makes use of Resource Governor which frequently kills their OLA DBCC runs, and doing so in a way that creates a false positive for disk errors. Is there a way to verify RG killed a process?
10:57 i_use_uppercase_for_SELECT: Do you have any recommendations or best practices for query store when restoring to lower environments with a different database name? The plans seem to include the database name and fail forcing plans.
11:50 themoderndba: I had you mention in a prior podcast a Microsoft guide for faster data loads to a database. What is the exact name of the guide? I am interest in this for a current ETL task.