Hello Yes Nice To Meet You
This month we have an invitation from Kevin Feasel.
Since Kevin is often kind enough to link to my posts via his Curated SQL feed/aggregator, I figured I’d add to the SEO pyramid scheme this month.
As part of my consulting services, I often help clients decide if they could keep a full time database person (there are too many potential titles to list, here) busy for 40 hours a week for the foreseeable future.
If the Magic 8-Ball comes back with a Yes, I’ll also help them write or review their job posting, and screen candidates. If a resume makes it past various detectors and doesn’t end up in the randomly generated unlucky pile, I’ll interview them.
Me. Personally.
Not in-person, though.
There’s not enough armed security in the world for me to get that close to HR.
The Production DBA Question
If someone is going to be in charge of production DBA tasks, I’ll of course ask questions about experience with whatever model is in use, or is to be implemented by the company.
So like, if the company has Availability Groups, or wants Availability Groups, we’ll talk about those.
For flavor, I’ll also ask them why anyone would be so insane as to not just use a Failover Cluster with Log Shipping.
But the real question and answer that tells me if someone knows their business is this: When you set up backups, how often do you take log backups?
If anyone says “every 15 minutes” without any further clarification or qualification, they immediately go into the “No” pile.
See, 15 minute log backups are a meme in the same Pantheon as 5% and 30% for index fragmentation. Neither answer shows any deeper understanding of what exactly they’re doing.
Log backup frequency is a business requirement based on RPO goals (Recovery Point Objective). If your RPO goal is less than 15 minutes of data loss, 15 minute log backups don’t meet that goal.
And anyone who says that Availability Groups help meet RPO goals gets their resume burned in front of them.
The Developer DBA Question
For anyone responsible for performance tuning, I need to make sure that they’re staying current with their learning and techniques.
I’ll show a lot of pictures of queries and query plans, ask about various anti-patterns and tuning techniques, but my absolute favorite is to show them pictures of query plans.
Actual execution plans, as it were.
Why is this important? Because at least half of the people I show this to will totally ignore the operator times and start talking about costs.
If you’re looking at a plan where this much information is available, and all you can dredge up to troubleshoot things are estimated costs, I’ll probably mail you a copy of this.
And I’m not kidding here — it shows a complete lack of attention and growth — what’s the point of calling yourself a performance tuning expert if your expertise peaked in 2008?
There Are Good People Out There
They may not spend every waking moment writing, recording, speaking, or chasing MVP status.
And that’s fine — I don’t care if you partake in that stuff or not — there are plenty of very smart people out there who don’t file quarterly paperwork or schedule posts months in advance.
What I care about is that you aren’t just an automaton seeking a new place to go through the motions in.
Thanks for reading!
Going Further
If this is the kind of SQL Server stuff you love learning about, you’ll love my training. I’m offering a 75% discount to my blog readers if you click from here. I’m also available for consulting if you just don’t have time for that, and need to solve database performance problems quickly. You can also get a quick, low cost health check with no phone time required.
> Since Kevin is often kind enough to link to my posts via his Curated SQL feed/aggregator, I figured I’d add to the SEO pyramid scheme this month.
The ouroboros strategy works, just as the prophecy foretold.
It’s really interesting how far you can get with an open question or fairly simple screenshot, as well as a bit of conversation. There’s a great chance to test assumptions and get to the “why” behind a person’s intuition.
And it’s rather considerate of you to send a copy of a highly-rated book, not just some crummy guide to rock identification.
I’ve enjoyed reading it myself, of course.
So I can get a free book? Man – my youngest would _love_ that book! 😉 I will admit that predicate is making my eyes water a bit.
As for the log backup, that’s usually a good start for “we get something”, but I’d hope a little more discussion would be in order to make sure that the business can recover to the time they need, not just within 15 minutes. And sometimes that SLA is even longer so maybe you don’t need 15 minute log backups. All good things to discuss. I’ve asked about general backup strategies in my interviews of others and I’m mostly looking for some understanding of backups to start. I’ve had some come in and know about backups, but not really anything beyond “I just use the defaults”.
And what _was_ that company that thought mirroring == DR and found out the hard way that it wasn’t? I remember the stories because they were pretty big and folded within a week because of something that affected the primary and was mirrored to the secondary, which resulted in pretty much everything being lost.
Well yes, the discussion is the most important point, but when you follow up to the 15 minute backup line and ask “what if I only want to lose 5 minutes of data?” and they don’t change their tune, you know it’s not a good fit.
I’ve spoken to many companies who equate HA with RPO. It’s quite shocking to them to hear otherwise.
Personally, I would love to go through an interview with you. But on second thought, I don’t think I would make it through without hysterically laughing because your blog posts and videos really crack me up!
The sad thing is that most people who I’ve interviewed don’t know me well enough to know when I’m joking. Or they’re in interview more and taking everything v. seriously.