First Responder Kit Release: Yukon Ho!

We’re going to Alaska this month, so try not to find any bugs, okay?

To celebrate, I’m channeling childhood nostalgia, and quoting a verse from one of my many Taco-Bell-dripping-stained Calvin and Hobbes tomes.

We’ll never have to clean a plate,
Of veggie goops and goos
Messily we’ll masticate,
Using any fork we choose!

When I was a kid, the word “masticate” was always good for a chuckle.

Until I found a dictionary.

You can download the updated FirstResponderKit.zip here.

sp_Blitz Improvements

  • #962@RichBenner – the man, the myth, the guy who cares about trace file names. Thanks to him, you’ll never have to worry about one being folded, spindled, or mutilated.
  • #1008: We added a debug mode. Actually, we added two debug modes. Using @Debug = 1 will print out CheckIds as they run, and @Debug = 2 will print out CheckIds and dynamic SQL used in some checks. This will help us identify errors and slow-moving queries. You’re welcome.
  • #1023: We noticed that there were some undeclared character lengths, and fixed them. Who cares, right?
  • #1029: We’ll only give you warnings about CMEMTHREAD when you have multiple NUMA nodes. Spiffy!
  • #1041: When we look at DBCC commands that have been run in the default trace, we ignore anything run by common monitoring tools. By “common” we mean “good”. If your monitoring tool isn’t on the list, perhaps this is a good time for silent reflection about the choices you’ve made.
  • #1042: The only thing worse than knowing fill factor is low is not knowing how low it is. Now you’ll know, and you can begin the long march to the bottom of your Desk Whiskey® bottle.
  • #1043: Endpoints; ah, Endpoints. Those verdant fields. It’s fine if sa owns them. Now you’ll know that too.
  • #1049: The memory dumps check was missing a BEGIN and END. Using advanced control-flow technology, we can now avoid running syntax when our Super Cool Consultant Criteria® isn’t met.

sp_BlitzCache Improvements

  • #1030: We’ll give you a holler if there’s a high variation between the rows estimated to come out of a query’s SELECT operator, and how many rows are returned on average by a query. This is v1, so if you run into anything weird, please let us know!

sp_BlitzFirst Improvements

  • Nothing this time around

sp_BlitzIndex Improvements

  • Nothing this time around

sp_BlitzWho Improvements

  • #1017: Not all blocking would be flagged. Sometimes none of it would be. The whole thing stunk! Stunk! Like the way those zit popping videos probably smell.
  • #1031: DMVs are awful. Has anyone else noticed that? One day you’re cruising along with a positive elapsed time, the next day SQL Server decides you’re a negative number. Yeah. Cool. Thanks, SQL Server.

sp_DatabaseRestore Improvements

  • Nothing this time around

sp_BlitzBackups

  • #977: In the humble words of @TheUsernameSelectionSucks “this is broken”. Now it’s fixed, with the healing powers of QUOTENAME. It turns out that thing you only ever used in dynamic SQL is good for other things, too.
  • #1039: Moving files broke really easily, and didn’t take into account if you were restoring a database with a different name. Now it breaks far less easily, and takes name changes into consideration.

sp_BlitzQueryStore

  • #908: Added filtering for plan_id and query_id. Why? Because I needed it for 24HOP.
  • #1030: Same row variation item that got added to sp_BlitzCache
  • #1034: Future-proofing for SQL Server 2017. Some new columns got added with query metrics in them.

sp_AllNightLog and sp_AllNightLog_Setup

There were like 10 changes here, and I’m far too hungry to write something funny for all of them. If you’re interested, here’s the full list of issues in this release.

Installer Scripts

Two turned into three! Yes, three! There are now two installers for the core Blitz scripts. One has sp_BlitzQueryStore in it, and the other doesn’t. The installer for all scripts includes everything. I’m not making a special one that doesn’t have sp_BlitzQueryStore in it for that one.

Deprecated

sp_BlitzRS and sp_BlitzTrace are no longer under active development. Whatever version of them you have is the final version. It would take a really, really impressive code change to get us to take them out of retirement. If you want them, you can still find them here.

You can download the updated FirstResponderKit.zip here.

Previous Post
What is the biggest mistake you made in production?
Next Post
Why Missing Index Recommendations Aren’t Perfect

2 Comments. Leave new

  • Between the last release and the latest August edition, I had fixed all the warnings about Alerts for Corruption, Sev 19-25, etc. Now sp_Blitz is reporting that there are no alerts. I’ve looked and they still appear to be there. What am I missing?

    Reply
    • Erik Darling
      August 2, 2017 3:31 pm

      A few of the checks were missing second BEGIN/END blocks. We have it fixed in the dev branch, and will push out a minor bump later this week. Just waiting to see if anything else comes in.

      Thanks!

      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.

Menu
{"cart_token":"","hash":"","cart_data":""}