SQL Saturday Colorado Springs (SQLSat #584)
It’s been a while since I’ve attended a SQL Saturday — the last one was in 2013 in OKC. I blogged about that one too, but lost the post when the old site went away. At any rate, it was great to be back. SQL Saturday is an excellent opportunity to catch up with what’s happening in the SQL community, and it aligns with my goal for the year to focus more on development. Although being the introvert that I am, meeting new people is still a difficult task.
I attended three morning sessions and two afternoon sessions, including Steve Jones’ DevOps session — which was the main reason I decided to go. The snow the day before made me question whether to bother, but it was worth the trip.
Sessions
Stats and Cardinality Estimator — Susantha Bathige (blog/twitter)
Susantha started things off with a discussion about the new cardinality estimator (circa SQL 2014) and then a deep dive into statistics. I’ll be honest — this was a bit over my head. It ended up being a little too technical, too early in the day. I picked up on some of it, but most of the discussion went over my head. This is clearly an area I need to work on.
Visualize Your Transaction Log — Brian Hansen (blog/twitter)
Brian gave a solid beginner session on the fundamentals of the transaction log. Much of this I already knew, but he got into the basics of VLFs, which was a good primer for the afternoon VLF session. I was torn between this and the advanced AG troubleshooting session — glad I chose the primer.
Bringing DevOps to the Database — Steve Jones (blog/twitter)
Steve offered a great session on DevOps and version control for database people. I walked out with concrete ideas on how to apply this at work and how it fits into my own transition toward database development. Now I just need to actually learn version control software.
Accelerating DevOps and TDM Using Data Virtualization — Tim Gorman (twitter)
Tim talked about thin provisioning databases using a hardware appliance. It’s a fascinating concept and I can see the benefit — the idea that you can thin provision databases to ease storage costs and avoid the bloat of test/dev environments is compelling. My question: is this something most shops can afford right now? Hell, I wish we had something like this already.
Get Into the Goldilocks Zone: Managing SQL Transaction Log VLF Growth — Michelle Poolet (twitter)
Michelle covered managing transaction logs and VLF growth with hard numbers and algorithm details. I’m glad I attended, but it wasn’t the revelation session I was expecting — turns out I’d already learned more about the subject than I realized. One key takeaway: adjust your t-log auto-grow settings instead of leaving them at defaults, pre-size your logs, and set a sensible auto-grow increment. Good advice.
Overall
It was a great day of learning. For the most part there was plenty of room — only one session ran out of seats. I noticed quite a few people wearing smartwatches and Fitbits, which I think is excellent. We spend too much time sitting. Although it could just be that these are data nerds who like to measure things.
I won SQL Toolbelt from Redgate’s raffle, which I was thrilled about. Not a bad way to end the day.
Key Takeaways
- SQL Saturday is worth attending even when the weather is bad — you’ll walk away with something useful from almost every session.
- The DevOps for DBAs track continues to grow; source control for database changes is no longer optional thinking, it’s a real path worth investing in.
- Pre-size your transaction logs and set a sensible auto-grow increment — don’t rely on SQL Server defaults.
- It’s OK to attend a session that goes over your head. It at least tells you what you need to learn next.