Not all of the questions y’all post at https://pollgab.com/room/brento are hard – some of ’em can be answered in just a line or two:
Q: Hany: Hello Brent, Who’s “Brent Ozar” in the Azure world?
Microsoft, and I wrote why here.
Q: Neutron Jack: Do you see any concerning TSQL deprecations in SQL 2022?
No, neither deprecated nor discontinued.
Q: MuleDonkey43: Have scalar functions improved much in SQL 2022?
Q: Looking For An Anti-Histagram: Our databases are many years old and over time have a lot of auto generated stats on the tables. I’m sure many are not used, is there an easy way to tell?
Not accurately, no.
Q: Yakira: Is there good way to identify the top 10 SQL queries using the most worker threads?
What’s the problem you’re trying to solve? If you’re working on THREADPOOL, I have a training class on that.
Q: Youssef: Do you have any recommended books for learning the XML behind a SQL query plan?
The XML, no.
Q: Ronaldo: Since SQL monitoring is broken in latest SQL 2022 CTP, do you foresee Microsoft kicking current monitoring vendors to the curb and creating their own commercial monitoring software?
You mean Microsoft System Center?
Q: Juan: Do you have any tips / precautions when using a table as a queue with RCSI?
Not RCSI specifically, but read this.
Q: Mashrur: Hello Brent, I completed your column store index training recently and it blew my mind. On that training you mentioned you only scratch the surface of table partitioning. You please kind enough to provide some advance reference (blog, YouTube, Paid) regarding table partitioning.
Thanks, glad you liked it! Sure, here you go.
Q: Kagamine Len: What are your thoughts about requiring a formal retention policy for all new tables created on production? Must have, nice to have, optional?
That’s determined by your company’s compliance department.
Q: Enrique: What are your thoughts on the new Json enhancements for SQL 2022?
I think it’s dumb to spend $7,000 USD per CPU core to use SQL Server as a file server. Microsoft thinks it’s a great idea. Go figure.
Q: Kol Dar: Is unbalanced parallelism a mainstream problem that SQL DBA’s should be regularly on the lookout for?
No. Use the methodology I teach you in the first module of Mastering Server Tuning.
Q: Joe: Good day to you Brent! The hardware provider allocated and presented 196 GB to our SQL 2016 Standard VM. After reading your BOU Weekly Links, June 6th Edition email it got me thinking. Would there be any benefit to setting the Max Memory above the 128 GB limit?
I don’t have experience with that. If you regularly need over 128GB RAM, you probably need Enterprise.
Q: Shlumiel: Are include fields part of your 5 x 5 suggestion for NC indexes?
Q: Dominique B: Hi Brent, wanted to start by “I’m a big fan of your work” but that may sound a bit cheesy 😉 I’m reading about the memory optimize tempdb and want to validate my understanding… This feature will make the old “rule” for the number of tempdb data file an old story isn’t it ?
No, and I explain why in this module of my Fundamentals of TempDB class.
Q: Eric Swiggum: There was a recent infrastructure outage, once resolved a SQL transactional replication was not applying changes, however we were not alerted. Our Undistributed Commands alert failed us too. Should I monitor “Not Running” statuses for these publications too?
Every now and then, on live webcasts, a question makes me lose my mind. I rant and rave and throw things. And I realize, as I write this, that I’ve never actually done that in text format.
So here it comes, Eric.
Should you monitor something that caused an outage?
Is that what you’re asking me?
Why bother? Screw it. Who cares? It’s not like outages matter, or that it’s your job. Forget monitoring. Just let the outages keep happening, and let your phone keep ringing. Life is meaningless. Death comes for everyone. Eat Arby’s.
And that concludes another episode of Office Hours.
Hey Brent, this is Eric S., I may have worded that question wrong, following the infrastructure outage (not database related) it seemed that the Log reader was “hung”, so the job wasn’t technically failing, however changes weren’t being applied. But as I recall replication monitor clearly displayed a “Not Running” status, I was just thinking of trying to add another custom alert that would catch this, for example checking the output of sp_replmonitorhelpsubscription. Sorry for the confusion, no reply needed.
Heisenbrent. I could even hear the voice of Walter White as I read that. 🙂