SQL Server Extended Availability Groups

Comments: 1 Comment
Published on: April 1, 2018

It may come as no surprise to many that Microsoft has hastened the SQL Server development cycle. Furthermore, it may be no surprise to many that Microsoft has also hastened the patch cycle for SQL Server.

If you were unaware of this, consider this as your notice that Microsoft has indeed hastened the patch cycle. Not only has the patch cycle become more rapid, the idea of Service Packs is more or less a notion of history at this point. Critical Updates (or CUs) is the new norm. This is a pretty good thing due in large part to the rapid improvements that can be made to the product due to Azure.

With all of this considered now, there is some really awesome news. A hint to this awesome news is in the preceding image and title of this post. In a recent CU for SQL Server 2017, Availability Groups and Extended Events both have seen massive upgrades. The upgrades are so big in fact that it is mind blowing. These upgrades were no small feat by any means and it took some major investment and cooperation from the likes of some well known competitors.

Upgrade the first: Availability Groups have now been extended to be able to include nodes from MySQL, PostGres and MariaDB. Frankly, I don’t understand the MariaDB move there but it’s all good. I am 100% on board with the MySQL addition and may have to work really hard to find a use case to include PostGres.

Imagine the realm of possibility this change brings!! First we got SQL Server on Linux and now we can include a predominantly Linux flavored DBMS in a SQL Server High Availability solution. LAMP engineers have got to be losing their gourds right about now over this. Microsoft is taking away every anti-MS premise that has been used in recent years and turning the world on its ears to become more global and reachable in the architecture and DBMS world.

Upgrade the second: In order to help support and troubleshoot AGs on these other platforms, we need some tools. The tools of choice happen to be in the form of Extended Events. While there is nothing quite yet in place on these other platforms to properly monitor an AG, XE is able to capture some MySQL, PostGres and MariaDB information as transmitted across the wire when these platforms are added to an AG. How COOL is that?

If you are really chomping at the bit, I recommend procuring the latest CU that was recently released. You can find that CU from this Microsoft site here.

The Wrap

I am in full support of this new direction from Microsoft. Partnering with other large platforms to provide a supremely improved overall product is very next level type of stuff and frankly quite unheard of in this ultra competitive world. It is so unheard of in fact that this was a nicely crafted April Fools joke. Happy April Fools Day!

Having mentioned Extended Events, if you are interested, I do recommend a serious read from any number of articles posted in 60 day series.

Extended Events File Initialization Failure

It should come as no surprise that I write a lot of articles about Extended Events (XE). This happens to be another article on Extended Events. Truth be told, this article is hopefully something that is more of an edge case scenario. Well, I sure hope that is the case and that it is not a common problem.

One of the recommended methods to trap payload data in an XE session is via the use of the event_file target. Sending data to a file has numerous benefits such as being able to take the trace and evaluate the trace file from a different machine (locally to that machine).

Every once in a blue moon you just may run into various issues with the event_file such as explained here or here. Though slightly different, the net effect is quite similar and should be treated with roughly the same kind of troubleshooting steps.

Configuration Error

As luck would have it, I ran into one of these rare opportunities to troubleshoot an error occurring on a client server. Truth be told, I was unfamiliar with the actual error at first. Here is that error.

Error: 25602, Severity: 17, State: 22.

The preceding error was scraped out of the SQL Server error log. Obviously a little more detail was needed because this error is far from useful without that detail. Looking a little deeper, I found some errors like this.

Msg 25602, Level 17, State 22, Line 43
The target, “5B2DA06D-898A-43C8-9309-39BBBE93EBBD.package0.event_file”,
encountered a configuration error during initialization. Object cannot be added to the event session.

Some very good clues are actually contained in that particular message. Some of these clues include the following: a) the term “target”, b) the term “event_file”, and c) the phrase “event session.” Ok, I get it at this point. One of my Extended Event Sessions I had put on the server and used previously was broken. But, since it had been working and I know I had fetched data from it, I found myself puzzled as to why it might be busted.

The next logical thing to do at this point was to test the various sessions that are stopped and try to figure out which one is causing the problem and see if the error is reproduced. Finally upon finding the session that is failing, I ran into the complete message.

Msg 25602, Level 17, State 22, Line 43
The target, “5B2DA06D-898A-43C8-9309-39BBBE93EBBD.package0.event_file”,
encountered a configuration error during initialization. Object cannot be added to the event session.
The operating system returned error 3: ‘The system cannot find the path specified.
‘ while creating the file ‘C:\Database\XEDROPME\SVRLoginAudit_0_131650006658030000.xel’.

The additional info that I needed is in bold text in the previous text. So, for some reason, there is a problem with the path for the XE trace file output. Going out to the file system to check it out, I found that the client in this case decided to delete the entire folder. How does that happen? Well, it does! When it happens, the XE traces will start to fail and you will no longer capture the intended trace data. Let’s take a look at a simulated reproduction of this issue.

First, I will create a session and then start the session, then validate the session exists and then stop the session.

And here is what I see on my test server when I validate the session exists.

Perfect so far. Now, let’s make that trace output directory disappear. For this demo, you might note that I had created the directory as “C:\Database\XEDROPME”. My intent in the name was obviously to notify the world that the folder was to be dropped.

That statement is easy enough and is performed from my test environment for those getting weary of the use of xp_cmdshell. Now, let’s try to start that session that we knew was previously running.

I re-formatted the output of the error for ease of readability. Otherwise, the output in the preceding image is what will happen when the output directory is dropped. The fix is rather simple at this point – put the folder structure back into place. To read an introductory post about checking if a session exists or not on your server, check this out (a more advanced post is coming soon).

Conclusion

From time to time we will run into various problems supporting Extended Events. This is bound to happen more frequently as we support more varied environments with more hands in the kitchen (so to speak). We need to learn that even small changes can have a ripple effect to other things that may be running on the server. It is worthwhile to perform a little due diligence and clean things up as we make changes – or at minimum to observe the system for a time to ensure no unintended consequences have occurred.

Extended Events is a powerful tool to help in troubleshooting and tuning your environment. I recommend investing a little time in reading the 60 day series about Extended Events. This is not a short series but is designed to provide an array of topics to help learn the tool over time.

Profiler for Extended Events: Quick Settings

Not long ago, I wrote a rather long article about a new-ish feature within SQL Server Management Studio (SSMS) that impacted Extended Events. You can read that book – here! The XEvents Profiler feature is one of those things that you may or may not use. If you consider using the feature, I do believe it is important that you research it a bit and try to learn the pros and cons first.

With that there is a little more about the feature that the aforementioned book did not cover. In fact, this information has pretty much gone ignored and mostly stays hidden under the covers.

Settings

As of SSMS 17.4 we have been given the ability to control XEvents Profiler just a tiny bit more. For what it is worth, we as Database Professionals love to be able to control our database environment. So this teeny tiny bit of new control ability is potentially a huge win, right?

If you are the controlling type, or maybe just the curious type, you will be pleased to know that under “Options” from the Tools menu in SSMS, Microsoft has tucked some new control options to help you configure XEvents Profiler – to a degree. If you open options, you will see this new node.

If you expand the “XEvent Profiler” node (circled in red), you will discover the “options” node. If you click on this “options” node and do a quick comparison (in SSMS 17.4 and SSMS 17.5) you will also find that you don’t need t expand the “XEvent Profiler” node at all because the options are listed in the right hand pane for both nodes and they are exactly the same. So, choose one or the other and you will end up at the same place.

The options that you currently have are:

  • Stop Session on Viewer Closed
  • Toolbar commands stop and restart

You can either set these options to True or False. I recommend you play with them a bit to discover which you really prefer. That said, I do prefer to have the “Stop Session on Viewer Closed” set to true. There is “profiler” in the name of the feature afterall. And if you have read the “book” I wrote about this feature, you would know that the filtering offered by the default sessions of this feature basically turn on the fire hose effect and can have a negative impact on your server. Are you sure you want a profiler style fire hose running on your production server?

Conclusion

There surely will continue to be more development around this idea of an XE style profiler. More development generally means that the product will mature and get better over time. This article shows how there is more being added to the feature to try and give you better control over the tool. We love control so the addition of these options is actually a good thing. Is it enough to sway me away from using the already established, more mature, and high performing tools that have been there for several generations? Nope! I will continue to use TSQL and the GUI tools available for XE that predated the XEvent Profiler.

Some say this is a way of bridging the gap. In my opinion, that gap was already bridged with the GUI that has been available for several years. Some say that maybe this tool needs to integrate a way to shred XML faster. To that, I say there are methods already available for that such as Powershell, the live data viewer, the Target Data viewer, or even my tools I have provided in the 60 day series.

I would challenge those that are still unfamiliar with the XE GUI (out for nearly 6 years now) to go and read some of my articles or articles by Jonathan Kehayias about the power that is in XE as well as some of the power in the GUI.

SQL Server User Already Exists – Back to Basics

One of my all-time favorite things in SQL Server is security. No matter what, it always seems that there is a new way to abuse permissions. When people abuse their access level or abuse the way permissions should be set in a SQL Server environment, we get the pleasure of both fixing it and then trying to educate them on why what they did was wrong and how to do it the right way.

In similar fashion, I previously wrote about some fundamental misconceptions about permissions here and here. I have to bring those specific articles up because this latest experience involves the basics discussed in those articles along with a different twist.

I do hope that there is something you will be able to learn from this basics article. If you are curious, there are more basics articles on my blog – here.

Gimme Gimme Gimme…

It is not uncommon to need to create a login and grant that login access to a database (or associate that login to a database user. In fact, that is probably a fairly routine process. It is so routine, that I have a demo script for it right here.

I even went as far as to include some of the very routine mistakes I see happening on a frequent basis (as referenced by a prior post here).

To this point, we only have a mild abuse of how to set permissions for a principal. Now it is time for that twist I mentioned. This user account needs to be created on a secondary server that is participating in either a mirror or an Availability Group. Most people will take that user account that was just created on the first server and then use the same script to add the account to the secondary server. Let’s see how that might look.

For this example, I will not go to the extent of creating the mirror or AG. Rather, I will pretend I am just moving the database to a new server. So I have taken a backup and then I will restore the database to the new server.

Next, let’s go ahead and recreate the login we created on the previous server.

You see here that I am only going to create the login if it does not exist already. Running the script produces the following for me.

Now, let’s deviate a bit and grant permissions for the login just like so many administrators will do.

It seems pretty apparent that my login that I just created does not have access to the GimmeSA database, right? Let’s go ahead and add permissions to the GimmeSA database and see what happens.

Well, that did not work according to plan right? Enter twist the second.

What I am seeing more and more of, is people at this point will just grant that login (that was just created) sysadmin rights. You can pick up your jaw now. Indeed! People are just granting the user SA permissions and calling it good. This practice will certainly work – or appear to work. The fact is, the problem is not fixed. This practice has only camouflaged the problem and it will come back at some future date. That date may be when somebody like me comes along and starts working on stripping non-essential sysadmins from the system.

There are two legitimate fixes for this particular problem (and no granting sysadmin is definitely not one of them). First you can run an orphan fix with a script such as this one by Ted Krueger. That will map the user that already exists in the database to the login principal (thus the reason for the error we saw). Or, you can prep your environment better by using the SID syntax with the create login as follows.

The trick here is to go and lookup the SID for the login on the old server first and then use that sid to create the login on the new server. This will preserve the user to login mappings and prevent the orphan user issue we just saw. It will also prevent the band-aid need of adding the login to the sysadmin server role.

The Wrap

In this article I have introduced you to some basics in regards to creating and synchronizing principals across different servers. Sometimes we try to shortcut the basics and apply band-aids that make absolutely no sense from either a practical point of view or a security point of view. Adhering to better practices will ease your administration burden along with improving your overall security presence.

This has been another post in the back to basics series. Other topics in the series include (but are not limited to): Backups, backup history and user logins.

Database Corruption and IO Errors

A quick way to have your day turned upside down and rip your gut out with nerves and anxiety is to come in one day to find that users are panicked, applications are not working and the HelpDesk team is curled up in the fetal position in the corner. Why? The sky is falling and everybody thinks the database has blown up.

Calmly, you settle in and check the server and eventually find your way to the error logs to see the following:

Msg 823, Level 24, State 2, Line 1

The operating system returned error 1(Incorrect function.) to SQL Server during a read at offset 0x0000104c05e000 in file ‘E:\Database\myproddb.mdf’. Additional messages in the SQL Server error log and system event log may provide more detail. This is a severe system-level error condition that threatens database integrity and must be corrected immediately. Complete a full database consistency check (DBCC CHECKDB). This error can be caused by many factors; for more information, see SQL Server Books Online.

Suddenly you understand and feel the collective fear and paranoia. What do you do now that the world has seemingly come to an end for your database?

Corruption

What exactly does this error message mean? Well, typically, an 823 error is a very strong indicator that there is some sort of problem with the storage system, hardware or driver that is in the path of the I/O request.

Great! That fear is getting a little heavier knowing what the error represents. This doesn’t bode well for the database. Let’s go ahead and crack out the list of what we can do or check when a problem like this hits:

  1. Check msdb.dbo.suspect_pages
  2. Run a consistency check for all databases on the same volume
  3. Check Logs (SQL, Windows, Storage system) to see if there may be additional info (via different errors/warnings) in close proximity to the 823 error.
  4. Check your drivers
  5. Restore the database

This is where your experience, training, and preparedness come in handy. An experienced data professional will be prepared with database backups (including log backups). So you are not concerned here because all of your backups are reporting successful. As you prep to pull the backups (for the past couple of days just in case) you notice that there are no available backups in your repository. Looking closer at your backup jobs you discover that the backups completed in mere seconds where they normally take hours for this database.

Now that your heart is racing, forehead is beading up with sweat, gut is sinking and the fear is very palpable – what do you do? Time to step through the rest of the steps and pull out your lucky charms, right?

Querying against suspect_pages, you find the table to be completely empty. You know that checkdb runs regularly but maybe it didn’t run last night. That is easy enough to check with a little query from here. Since a consistency check does not seem to have run (as confirmed by the script) and is the second item on the checklist, let’s go ahead and run it now.

Msg 0, Level 11, State 0, Line 0

A severe error occurred on the current command.  The results, if any, should be discarded.

Msg 0, Level 20, State 0, Line 0

A severe error occurred on the current command.  The results, if any, should be discarded.

Crud. Blood pressure and nerves are getting a little more frazzled now. Maybe we can cycle through the database and find which table is causing the problem. Let’s try a checktable of every table in the database. Before doing the checktable, one more check against suspect_pages still shows no rows to be found.

Running the checktable, every table is coming up clean except one. That one table produces the same sort of error as the checkdb and just so happens to be the largest and most critical table to the database. Blood pressure is closing in on critical now. We have a corruption issue that is severe enough that checktable cannot complete, we know we have 823 errors and some sort of IO issue and do not have a backup.

Wait…backup. Let’s try to force a backup and see what happens. We can tell the backup to continue after error so let’s see what happens. Maybe that will allow you to move the database to a different server or different spindles to try and just recover the data.

Msg 3202, Level 16, State 2, Line 1

Write on “E:\SQLBackups\myproddb.bak” failed: 1(Incorrect function.)

Msg 3013, Level 16, State 1, Line 1

BACKUP DATABASE is terminating abnormally.

The situation just does not want to get any better at this point. Time for drastic measures – shut down the SQL Server services and try to xcopy the data and log files to a different server and try to re-attach from there. Anything is worth a shot, right?

Error 1: Incorrect Function

Glad to know it is not just SQL Server throwing the errors – sorta. The corruption is ever present and there is nothing more that can be done, right? All hope is lost. Time to fill out the resume and move on to something else, right? Give it one more shot. A spark of insanity hits and you wonder if a mere query against the clustered index will work, if not then maybe something against any of the indexes to try and spare any data at all.

You rub your luck horseshoe and query the table (yes the table that checktable aborts because of corruption). Lo and behold you get results from this very simple query. How is that possible. On a whim, you drop all the Non-Clustered Indexes and try a fresh backup.

Hallelujah! The backup completes without error. Time to take this backup and restore the database to a completely different server. Then on the new server run a consistency check to determine if it is all clear. To your liking, there is absolutely no corruption at this point so the non-clustered indexes can be recreated (easy to do because you have a script with the index definitions handy).

Wrap

This journey from fire and brimstone and the world ending to the epiphany and then the sweet euphoric feelings of success is not a normal resolution for these types of errors. This sort of thing happened for a client that called trying to get around the 823 errors. We had absolutely no indication whatsoever of where the corruption was beyond knowing we had failing disks at the time. We got lucky in that the non-clustered indexes in this case ended up being stored on the bad sectors and dropping those indexes allowed us to recover the database and make it usable.

When everything is breaking, the unconventional thought (especially without having a checkdb complete to tell you which index or which pages are corrupt) of dropping indexes may just save your bacon. It saved my clients bacon!

As a follow-up item, it is important to regularly check the dbccLastKnownGood for each database. That can be done by following the script in this article.

«page 3 of 118»

Calendar
September 2018
M T W T F S S
« Jul    
 12
3456789
10111213141516
17181920212223
24252627282930

Welcome , today is Friday, September 21, 2018