I get the call, you get the call, everyone gets the call. “Hey, my app/procedure/query/report is running slow.” Now what do you do? You go to my full day session at SQL Rally, that’s what. Assuming you vote for it.
I didn’t post the abstract I submitted for the SQL Rally before because I thought that it would be redudant. However, since it’s not right off the voting page (unless they updated it since I voted), if you’re interested, here’s what I thought I would do for a day. If it sounds good to you, please go here and vote for it.
One of the most common problems encountered in SQL Server is the slow running query. Once a query is identified as performing poorly, DBAs and developers frequently don’t understand how to diagnose the problem and often struggle to fix the problem. This one day seminar focuses exclusively on these two topics. Attendees will learn how to identify the queries that are performing badly and learn how to fix them. We will start by learning how to gather performance metrics, both server and query metrics, using tools available directly from Microsoft such as performance monitor, DMVs and Profiler. From there we’ll move into learning how the optimizer works and how it uses statistics to determine which indexes and other database objects can assist the performance of a query. The session takes considerable time to show exactly how to generate and read execution plans, the one best mechanism for observing how the optimizer works. We’ll then look at other DMVs that can assist you when performance tuning queries. With all this knowledge gathered, we’ll move into looking at common performance problems, how they evidence themselves in the metrics and execution plans, and how to address them. Finally, we’ll explore advanced methods for solving some of the more difficult query performance problems introducing such concepts as query hints, plan guides and plan forcing. Through all of this, best practices and common techniques will be reviewed. Attendees will go home with a working knowledge of query performance tuning, a set of methods for identifying poorly performing queries, scripts to assist in these processes and the knowledge of how fix performance problems in their own systems.
Although I would prefer that you voted for me, it’s more important that you vote at all (same thing as in real life). Please go here and select the sessions that you want to see.
One of the national sponsors for SQL Satuday is Confio Software. Just because they sponsor SQL Saturday, it’s worth checking out their product, but they also host the speaker’s dinner at these events. So, having eaten off their dime, I felt obligated to take a look. I’m glad I did.
Confio Ignite is a monitoring software that keeps real-time and historical track of the performance of your SQL Server (and Oracle and DB2) database servers. It’s focus is on wait states and queues, a very common method for troubleshooting performance.
You can get a trial download from their web site to run for a couple of weeks yourself. Everyone’s first impression of a software, after the web site of course, is when go to do the install. The install routine for Ignite was very easy. It’s also a bit of a shock. It actually uses web pages for configuration. This makes sense when you consider that the software originally comes from Oracle and Unix systems. Anyway, despite the bit of weirdness of using a web page for configuring software, it’s a very easy install. Once the install is complete, it’s also really simple to add servers for the software to monitor.
I put it to work on a number of development servers to start (I’m not plugging some unknown software into Production without a shake-down. I may be stupid, but I’m not crazy). It’s agentless and it began running queries against the systems immediately. It’s storing the data to my local desktop. That data is available in real-time, a historical view, or a trending view, showing how performance is changing over time. It’s all through a web site, not dissimilar to Quest’s Foglight. You can see a list of servers, their immediate state, action choices and some other general info.
Clicking a server, after it’s had time to gather data, put’s you into the Trend screen. It’s laid out very nicely. You get to break down performance by SQL, Waits, Programs, Databases, Machines, or Users. Then it shows information by Total Wait, Average Wait or Typical Day. The chart gives you a visual representation of queries running against the system, by default numbered and assigned by an internal hashing system that Ignite uses (no, I don’t know what the internals look like, I’m inferring that from the information presented). Down at the bottom of the screen you can see Top Query Problems, which shows you queries that are running slow or using too many resources according configurable settings, or Resources, your standard breakdown of CPU, Memory Paging, Disk Queue and the rest, or SQL Text, literally the queries that have been running on your system.
Almost everything has a drill down that takes you to a lot more information, and there are pop-ups that show you bits of detail.
I’ve only been running it for a few days, but I can already see how it could be very useful. I’m very happy with the focus on waits and on queries. So often when you’re looking at these tools queries are either an after thought or ignored completely. You frequently have to spend time monitoring your queries on your own, while that very expensive monitoring software is tracking page life expectancy and buffer cache hit ratio, both useful measures, but you still need to know which query is causing the problems. Ignite will get to that information right away and still collect the other as well.
So far, I’m pretty pleased with the software. I’m going to switch this over to a production system in the next few days and see what I can see with Ignite running against a real system. More to come.
I’ve blogged in the past about the nHibernate project that has been going on at my company for the last eighteen months. Prior to today, I have only seen the database generated by nHibernate. But today I finally started seeing some TSQL code. My first impressions… oy!
There are two levels of concern coming out of the gate. First, it appears that some of the programming decisions, completely independent of nHibernate, are going to cause problems. Second, it appears we’re going to be hitting issues with nHibernate.
First, the programming decision. I’ve been using Confio to monitor this server for a few days now (look for upcoming posts on my eval of Confio). Only one day has captured any real activity from the nHibernate team (and yes, I’m basically spying on these guys because they are absolutely insistent that my team stay out of their way). The longest running query was two calls to this (names have been changed to protect my job, there are no innocents):
FROM Books WITH (updlock,rowlock)
WHERE BookID = 42
What they’re doing is locking the row so that no other user can get at it while it’s being edited. Total lock time for the first day was 38 seconds for two calls. I’ve posted about this previously, so I won’t go into it again, but this is potentially a disaster.
On to nHibernate. The next query was pretty benign:
WHERE OfficeID = 42
Unfortunately this table, as created out of nHibernate, has no primary key, no clustered index or any other type of index, so this is just a table scan. But it’s only on 109 rows… in dev, not production, oh, and most of the 109 rows have a null value for OfficeID, but it’s all good I suppose… until I notice that this query also had 38 seconds of wait time, but it was called 394,652 times… oy (and no, I’m not even one of the chosen, but that’s a wonderful expression for this situation). Looking at the data in cache, this query has been called, since it was created in cache on the 2nd, 598351 times with a total elapsed time on the server of 5429689ms. The average then is 9ms, but the max was 29296 or 29 seconds.
The next query up looked like this:
INNER JOIN Table2
ON Table1.ChildID = Table2.OtherID
WHERE tabl2.ParentID IN (@p1,@p1,@p2…@p99)
Yep. 99 parameters passed and used against an IN statement. I can’t post the execution plans on this stuff without talking it over with the boss, but suffice to say, it’s two table scans and a merge join to get the data back. Oh, and this was called 1,138 times with a minimum elapsed time of 976ms. Yeah, minimum time to call this proc is 1 second and it’s been called over a thousand times.
It went down hill from there. I’m looking at other queries, one that hits seven tables, but instead of using JOINs uses a bunch of sub-selects in the WHERE clause. It has a minimum run time of 9 seconds. Luckily it’s only been called 64 times.
This is a horror show.
A lot of benefit can be reaped if we’re able to go in and create some indexes on these tables, but that’s going to be an issue anyway because I’m seeing lots of implicit data conversions on the parameters being passed in, as I noted a couple of years ago. At this point in time, I would say, at best, I’m very underwhelmed by what nHibernate is producing. To be perfectly honest, I’m concerned for the health of the project.
Quest Connect 2009, taking place in October 21 for 24 hours, looks like it’s going to have 64 different sessions, live and recorded, by a variety of the names in the industry. It’s another chance to dig in and learn the details on a variety of topics from some of the top names in the business. Can you say Tom LaRock? How about Tim Ford? I know you want to hear from Brent Ozar. Those are just some of the featured speakers. There are a whole slew of others, it’s worth pursuing, and did I mention, the price is right.
I recorded a session for them last night. It’s on the basics of understanding execution plans.