[Video] Office Hours and an AI-Generated Sweatshirt

Videos
1 Comment

I originally took these questions from https://pollgab.com/room/brento poolside at the Trixie Motel (really), but I forgot to use the good microphone, so I re-recorded ’em from my home studio when I got back:

Here’s what we covered:

  • 01:04 NeedMoney: What’s the fastest/easiest way to make money as SQL Server Consultant (Production work, performance, health check, backups, maintenance)
  • 02:52 LatestGreatest: When would you recommend upgrading to a new SQL Server? Would you suggest when a new relevant feature is released or just when installed version goes end of life?
  • 04:38 PolarExpress: Would you consider/recommend a client to use SQL Server Express edition as viable solution for a small web app?
  • 05:55 ConstantCare: How much Cloud Costs (compute, storage, database) do you pay to run the Postgress database for ConstantCare?
  • 06:30 Right Arm of the Darling One: Batch Mode on Rowstore is a few years old now. Has it served you well?
  • 07:33 SadButTrue: Hey Brent, what is the most efficient way to tell if data in a table has changed since a point in time? I don’t care about what changed, just if it did change. I’m using change tracking now, but having a cursor running CT functions over 700 tables has a high CPU impact.
  • 09:04 Craig: If I have an IF statement statement that only includes a single statement, is there a reason to ever use BEGIN…END? It’s common and seems to just take up space. Also, any reason not to put simple IF statement on one line? e.g. IF NOT EXISTS (SELECT * FROM
  • #Results) RETURN;
  • 10:11 MyTeaGotCold: Do you still predict that Amazon is trying to build Aurora for SQL Server / Oracle?
  • 11:15 Handsome Consultants Aside: Hi Brent. Can you think of anything that SQL Server is the best in the market for? Oracle and Postgres seem to have it beat.
  • 13:14 Hekaton: How do people build SQL boxes with 1 TB of RAM? Every mobo I know of only had 4 slots for RAM and I have never seen a 256 GB RAM stick.
  • 14:10 DBACAT: Hi Brent, have you ever done a mashup of your top X number of roasts you do to people from office hours or other sessions? I think it would be hilarious. It’s probably a lot of work but a “The best of Brent’s roasts” would probably get a lot of views.
  • 16:14 JuniorWannabe: Hi Brent, what are your thoughts on a decent quick check if something changed recently on a SQL server? Sometimes an app stops working and it is easy to look at modified date of stored procedures or functions, but what would you look for?
  • 17:32 chris: Having followed your blog for some time I understand fragmentation is rarely the root of a performance problem these days. But, how do I handle vendors that are insistent on defragmenting the indexes? Further, what does the fragmentation per centage mean with a SAN back-end?
Previous Post
Index Rebuilds Make Even Less Sense with ADR & RCSI.
Next Post
The 6 Best Things Microsoft Ever Did to SQL Server

1 Comment. Leave new

  • JuniorWannabe: my team and I built what we call “CentralRepository” which is a database that stores meta data of all our servers, databases, jobs, users, logins, and tons of other configuration data. It collects data on every server every day and each run is versioned so we can determine how the server/DB was configured at a point in time. Another tool we built was a utility database that site on every server. That also has a ton of data we store about our databases. Examples include: index definition changes, and stored procedure changes. We set an initial baseline and compare that to the definitions on the current day, updating the baseline with the changes for the next day’s run. It’s taken us over 12 years building these tools we started small and grew them from there. We need this because we support over 400 servers and this information is valuable.

    Reply

Leave a Reply

Your email address will not be published. Required fields are marked *

Fill out this field
Fill out this field
Please enter a valid email address.