Querying the Execution Plan for Subtree Cost - Take 2

A few weeks ago I shared a query for probing the execution plan for the Estimated Subtree Cost value.

That query was rather simple but it got the job done. In that post I also explained how, although there are several ways of finding missing index suggestions, they were not always very helpful. After posting I was thinking that it would be helpful to combine the two, or at least display the results so that it is easy to determine which execution plans include missing index warnings.

Querying the Execution Plan XML for Subtree Cost

There are quite a lot of great tools and queries available to the DBA that aid in troubleshooting and performance tuning. The list is incredibly long, but I mainly use RedGate's SQL Monitor (which has come on in the last 5 years in leaps and bounds), Glenn Berry's DMV queries, Adam Machanic's sp_whoisactive and Brent Ozar Unlimited's sp_Blitz suite (predominantly sp_Blitz and sp_BlitzIndex). In writing this, I see that Brent Ozar has expanded his offering, so it looks like I have a weekend project ahead of me!

SQL 2016 Database Scoped Configuration: MAXDOP

SQL Server 2016 Microsoft expanded the range and flexibility of database level configuration settings. There are some new features, as well as features which that can be set at more than just the server or database scope. Now, among other things, we are able to:

• set certain options both at the instance and the database level (MAXDOP)
• set instance level trace flag settings at the database level (LEGACY_CARDINALITY_ESTIMATION)
• the Trace Flag 1117 is now set at the FILEGROUP LEVEL
• the brand new database level CLEAR PROCEDURE_CACHE

A summary of the new options can be found on the Technet blog.

In this post I am going to focus on the MAXDOP feature. Having read a few articles and posts about this feature it wasn't clear to me exactly how this feature worked. Especially in regard to the relationship between the instance level setting and the database level setting.

Master Data Services Model Cloning

Yes! My company uses Master Data Services or MDS. Yes! It's a pain. And yes! I needed to learn how to use and administer it. For anyone that hasn't had this well meant tool forced upon on them, it feels ... (if I'm being kind) ... a little unfinished.

Based on my experience MDS administration requires an awkward mix of command line tools, the MDS web interface, and even some sql scripts and stored procs executed directly on the database. There are a few strange quirks, like the fact that individual users or groups can be set as administrators, but there is a single super-user. This super-user can only be an individual and not an AD group (although I'll be happy to be proved wrong on this!). Users in the System Administration role can deploy models, but still need to be given permissions to view and edit those models by the super-user (this has changed in SQL 2016 - multiple super-users are supported as well as AD groups are now supported as super-users).

Check for SQL Server Connectivity

My company is revamping its domain structure and tightening up network security. I am one of the guinea pigs and therefore have been put into the new domain and OU groups. Obviously, we are experiencing some teething problems. As a result I can't access things, namely SQL Server instances that I used to be able to connect to from my work station.

I needed to provide the network team a list of the instances I could no longer access and I certainly wasn't going to try to connect to each one manually. We have a very large estate and it would have taken me ages.

Cleanup SQL Server Logins

Cleaning up after someone leaves. We all do it right? Our companies all have water tight policies and procedures to handle that, or maybe they have policies but not procedures. Or maybe they just kind of don't.

Well, if your company is anything like most of the places I've worked, it's somewhere in the middle. And from a DBA perspective, especially on non-production environments, your list of logins and users becomes a bit long and messy after a while.

Compress your database tables

SQL Server has offered data compression since SQL 2008. It is a very good feature but, as with almost everything SQL Server, it needs to be deployed judiciously. Data compression has been written about and reviewed by many SQL experts who are much more qualified and knowledgeable than me, so I won't go into it too much.

First, a quick summary of what the feature does as well as its benefits and effects:
As the feature's name suggests, when it is enabled on a table and/or its index(es), SQL server compresses (or shrinks) the data before it is written to the database. There are two types: ROW and PAGE.

Query To Find SQL Audit Details

I've been deploying a SQL Server auditing solution for our new SQL 2014 estate. I know, we're a few years behind. But if you saw some of the legacy systems I work with you'd understand that migrating to SQL 2014 is quite a coup!

Anyway, I found a really great audit solution that I have used as a base on Colleen Morrow's blog. I will detail what I have added to her solution in a later post. But, for now, I would like to share a query that displays some basic information about existing audits, both at the server and database level.

I put this query together for the purposes of documentation. Our 2014 estate is getting large fast. And with so many installations, we need to keep track of what is installed and configured on all our servers.

Restarting a service with PowerShell

We use Redgate's SQL Monitor to keep an eye on our mission critical servers. I like the product. It is reasonably user-friendly, reliable and it helps my team monitor our servers efficiently. I've used other products before and I think SQL Monitor compares favourably. No, I'm not paid by RedGate!!!

With that in mind, my team has been struggling a bit with an ongoing issue: every Monday morning one of the SQL Monitor processes was running amok and maxing out CPU! This was strange, as the product is generally pretty stable. After some investigation, we realised that the security team had started performing regular vulnerability scans over the weekends. Part of this test involved intentional failed login attempts. SQL Monitor can't seem to handle this, and the web server, specifically the "xsp4" process goes haywire and trashes the CPU.

Databases With No Recent Backups

I've written the script below to easily identify databases that have not had backups taken within a certain amount of time. This could be useful if you need to build your own monitoring. However, even if you use a third party monitoring tool,as I do, it is helpful to run this query on the server for which I've received an alert.

Many of the servers I manage have a large number of databases. The alerts I receive for overdue backups are database specific. But, in general, if one database is overdue, chances are something has happened on the server. So this query can give me a quick overview.