T-SQL Tuesday #104: Just Can’t Cut That Cord

We all have our favorite scripts, tools or utilities. Those are the things that help make our jobs easier. Some of us may have an unhealthy relationship with some of those scripts (similar in nature to the relationship many have with their phone). Whether or not the need to cut that proverbial cord exists, today we are not discussing the health of that dependence. Suffice it to say, sometimes we simply need to upgrade our scripts. How else can we get better scripts or make our scripts better – by sharing them.

This is precisely the goal Bert Wagner (b | t) seems to have envisioned for the 104th installment of TSQL Tuesday.

If you are interested in reading the original invite, you can find that here.

For this month’s T-SQL Tuesday, I want you to write about code you’ve written that you would hate to live without.

Maybe you built a maintenance script to free up disk space, wrote a query to gather system stats for monitoring, or coded some PowerShell to clean up string data.  Your work doesn’t need to be completely original either – maybe you’ve improved the code in some open source project to better solve the problem for your particular situation.”

There is a high probability that through the sharing of your script, somebody out there can benefit from that script. In addition, it is very likely that somebody will make a suggestion to help make your script better. Worst case (emphasis on worst case here), you have the script stored somewhere with half decent instructions on what it does and making it easily accessible for you to use again and again. Just in case you forget you have it out there – you can google for it again and find it on your own blog ;).

Personally, I have been able to find and re-use some of my older scripts. Not only do I get to re-discover them, but I also get to re-imagine a new use or improvement for the script.

Brief Intermission

A shout out is absolutely necessary for Adam Machanic (twitter) for picking the right blog meme that has been able to survive so long in the SQLFamily. This party has helped many people figure out fresh topics as well as enabled them to continue to learn.

Easy Access

While pondering the topic for today, I had the thought occur about how frequently I post a script on my blog already anyway. An easy out for this topic would have been to re-share one of those old scripts. For instance, I could easily redo a recent article about server access that has a couple scripts demonstrated in it. Or I could go back a few years to my articles about foreign keys (here or here) and space use (here or here). Even more intriguing could be to re-envision some of my articles on Extended Events. But where would the fun in that be?

Rather than take the easy road and rehash something, I have something different. This one goes hand in hand with the numerous articles and scripts I have previously provided on auditing – yet it is different.

Not every shop can afford third party software or even Enterprise edition and so they have to come up with a different way to audit their database instances. One of the problems with a home grown solution is to ensure the data is not stored local to the server (lots of good reasons for that). Here is an example of what I did for one client that happened to have a developer that found a back door that was giving him SA access to the SQL Server Instance and was changing things and trying to cover his tracks – even after being warned.

First the query

This query will be run from a job on a different server that is restricted in access to just a select few people. I do rely on the use of the default trace in this query. I am also reliant upon a little bit of sneaky behavior. If I run this from a separate server, prying eyes are usually unlikely to find that it is running and thus makes it easier to catch them red-handed. In addition, if they discover via some sort of trace and by a lot of luck that it is running, then they have no access to the remote server to alter anything that was captured.

The query does go out to the default trace and pull back any changes to permissions or principals on the server in question. The captured data is then stored in a database that is also restricted to a select few people. Lastly, the captured data can be routinely queried, or automated reports can be created to send email notifications of changes encountered.

The second part of the trickery here is that I am using a linked server to perform the queries (a slight change and I could also do this via powershell which will be shown in a future article). The linked server query uses the openquery format and sends the default trace query to the remote server. Since I am running this from a job on an administrative server that pulls a limited data set, I am not overly concerned with the linked server setup here.

Storing It

Once I query the data, I need to put it somewhere on my administrative server. The table setup for that is very straight forward.

After creating this table, I am ready to store the data. All I need to do is throw the audit query into an agent job and schedule it to run on a regular schedule. For my purposes, I usually only run it once a day.

TSQL2sDay150x150The Wrap

This has been my diatribe about service and giving back to the community. When done properly, there is a natural born effect of enhancing one’s personal life equal in some way to the amount of effort given towards the community.

Oh, and if you are interested in some of my community contributions (which according to Jens Vestargaard is an awesome contribution), read this series I have published.

Summiting that Technical Challenge Part II

Comments: No Comments
Published on: July 6, 2018

Conquering Challenges

Months ago, I posted an article about some of the challenges encountered while migrating from one service level to another for my blog.

As it turns out, I had some long lingering effects that I was delaying fixing because I didn’t want to have an apparent flood of republished posts.

What could possibly be wrong after fixing everything I listed in the aforementioned article? Good thing you asked. Every occurrence of a percent symbol was changed to a new string – “{529e71a51265b45c1f7f96357a70e3116ccf61cf0135f67b2aa293699de35170}”.

Things that make you go huh?

As you can see, this string is hardly useful. I started noticing this string in several articles and it took a while to figure out what the heck was going on. It wasn’t until I was reviewing an old post that had a SQL Script in the post that contained a “LIKE” keyword. An example of one such article is this one about SQL Trace. At that point, the lightbulb went off and I decided to give it a try. Lo and behold, that string truly did represent a percent symbol. I made the appropriate changes and voila, the post looking perfect all over again.

That was one measly little post. I have ~100 articles affected by this problem. Obviously that becomes tedious to change over and over and over. In addition, if I do change them manually one by one, I end up with ~100 articles triggering social media alerts that makes it look like I am re-publishing a bunch of stuff when really it is just editing and fixing the articles. What to do? What to do?

Thank goodness I have half an idea how to update data in a database and my blog is stored in a database. I can run a mass update to replace that string properly wherever it occurs. Let’s try it.

First, just one article to validate…

And if I look at the post after the update, I can confirm that the update succeeded (in my staging database first and foremost). Next, I roll that change to my production database and test again.

After I can confirm that it worked there (you can also see that it worked via this link), then I can work on the mass update.

Now, when I check my blog for that string, I find that there is no more residue left and all seems to be back to being correct.

To fix the string problem, I did use the replace function. Since this is a MySQL database, I had to use that version of it. Notice it isn’t terribly different from the TSQL replace function.

The Wrap

I did not enjoy this journey much at all. Much of the experience was due to outside forces. I can’t do much to control them, but I can do something to fix the net effect of what they caused. I am sure that with all of the problems encountered in that migration, I will find something further that needs to be fixed.

 

Use SSMS with a Different Windows Account – Back to Basics

One of the tasks I find myself doing on a fairly regular basis is running SSMS as a different Windows User. The two biggest use cases for this are: a) to test an account to prove that it is working (or not) and has the appropriate level of access, and b) to use SSMS to connect to a Domain SQL Server from a computer in a different domain (or not on the domain).

In addition to needing to do these tasks for myself, I find that I need to show somebody else how to do the same thing on a fairly consistent basis. Considering the finite keystrokes we all have (which I referenced here), it is time for me to “document” how to do this task.

I will cover two really easy and quick methods to perform this task. One from a command line and the other from the GUI. Both methods will involve a variation of the runas utility.

RUNAS

Let’s start with the easiest of the two methods. In this case, you will need to test windows account (let’s call it a domain account) from a computer which is on the same domain. This requirement allows us to take advantage of the shortcuts from within the GUI to access the runas utility.

To access the runas from Windows, one will first locate the icon for SSMS from the Start Menu, then right click that icon as shown here.

After right clicking the icon, you will see a menu pop up on the screen. Select “Run as different user” from that menu. Once you have selected the appropriate “run as” option, a login prompt will appear as shown here.

Enter the appropriate credentials at the prompt and then SSMS will launch. In this case, I may want to test the account myidomain\domain.useracc. So, I merely need to enter the domain credentials for that account. A caveat here is that the account you are testing will need to have the necessary permissions to “logon” to the workstation in order to launch the app – unlike the second method.

CMD Line

This second method has a few advantages over the GUI method with the biggest advantage being that you can use this method from any machine on the domain or even a machine not joined to the domain (so long as you have the ability to authenticate to the domain). And of course the additional advantage that the account you are testing does not require “logon” permissions on the machine you are using.

Let’s start with the basic command.

I can run that from a command line, or I can throw that into a desktop shortcut (the shortcut method is much more convenient). After I hit “enter” from the command line, I am prompted for a password for the account to be used for that session. Here’s an example of how that would look.

You won’t be able to see the password being typed (don’t fat finger the password 😉 ), but after you enter it successfully and press “enter” then you will see SSMS start to launch. After a successful SSMS launch, you should see something similar to the following:

I have a few things highlighted here of interest. First, in the red box, you will note that the user shown as connected to the server is my “local” test box account instead of the domain account. However, if I verify the authenticated account, I can see that the domain account is indeed accessing the SomeServer SQL Server (as demonstrated by the green box on the right).

The Wrap

Sometimes what may be ridiculously easy for some of us may be mind-blowing to others. Sometimes we may use what we think are common terms only to see eyes start to glaze over and roll to the backs of peoples heads. This just so happens to be one of those cases where launching an app as a different principal may be entirely new to the intended audience. In that vein, it is worthwhile to take a step back and “document” how the task can be accomplished.

Runas should be a very common tool in the toolbox of all IT professionals – not just Data Professionals. Learning how to test different accounts is essential to being an effective and efficient professional that can provide solid results.

If you feel the need to read more about single-user mode, here is an article and another on the topic.

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.

Single User Mode – Back to Basics

In a recent article, I took you on a trip through how to hack (ethically) a SQL Server to regain sysadmin access. In that article, I made quick mention of restarting SQL Server into single-user mode. It only makes sense to show quickly how to get into single-user mode.

Before getting into that, 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.

Single-User

So, what exactly is this single-user mode thing? Single-user mode is basically the official back-door into SQL Server for various reasons such as:

  • Somebody deleted all of the logins that were in the sysadmin role.
  • The sa account is disabled or the password has been forgotten.
  • Somebody deleted any Windows groups that were members of the sysadmin role.
  • All members of the sysadmin role are no longer with the company.
  • You need to restore the master database
  • You want to keep SQL Server all to yourself because you are greedy!

These are some pretty solid reasons to need to be able to use the back door. But how exactly do we get to the back door?

Two Paths

As luck would have it, there are two ways to enable single-user mode. You can either get there by making some changes for the SQL Server service in Configuration Manager, or you can utilize a command prompt. I won’t cover the gui path beyond the gentle reminder that you must remember to undo your change when using that method.

My preferred method is through the command line. Using my SQL Server 2017 as the experiment, I would navigate to the Binn directory for that instance. In this case, as shown in the next image.

Before getting too far ahead of myself, I am going to stop my SQL Server.

Notice, I also queried to find all of my services related to SQL before stopping the MSSQLServer service via the net stop mssqlserver command. We will come back to some net start and net stop commands later.

With the service successfully stopped, I can now restart the service in single-user mode.

And then the validation that we are indeed starting in single-user mode…

But wait, did you notice that bit of trickery on the startup command?

This is a pro-tip for when you must use single-user mode. Inevitably, somebody will steal the single-user connection and you will be locked out of the session. By using an app name after the single-user switch, you are telling SQL Server to only accept connections for that specific application. Since most apps will not be using sqlcmd, you will have far less contention to gain that connection and you will be able to complete your task much easier.

You could also pass something like this instead…

In this case, I would be limiting the connections to a query from SSMS (and not object explorer).

Now that I have a single-user connection, I can add a sysadmin or restore the master database or just sit on it and play devious. It all depends on what your objective for the single-user session happens to be.

More Command Line

Remember that reference to the NET commands? Well, it turns out we can also start SQL Server in single-user via net start. Let’s check it out.

The command is pretty simple:

The effect here is the same as navigating to the Binn directory and starting SQL Server with the sqlservr.exe executable. The big difference is considerably less typing and less verbose output of the service startup.

When using the net start method, you do need to know the service name of the SQL Server instance. To get that, I do recommend the following powershell script.

This will produce results similar to the following.

From the results, I can pick the SQL Server service and then pass that to the net start command fairly easily.

The Wrap

Starting SQL Server in single-user mode should be a tool every data professional holds in the bag. This is an essential tool that can be used in multiple scenarios and ensure you are able to fully maintain and control your server. I have shown how to get to single-user mode via two command line methods and mentioned a GUI method. The nice thing about the command line methods is that you don’t have to remember to undo the startup switch like you do with the GUI method.

If you feel the need to read more about single-user mode, here is an article and another on the topic.

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.

Lost that SQL Server Access?

As a data professional can you recall the last time you needed to support a SQL Server instance for which you had no access? What if you used to have access and then that access magically disappeared?

I know I run into this dilemma more than I would probably like to. It is rather annoying to be under a crunch to rapidly provide support only to discover you are stuck and have to wait on somebody else who hopefully has access.

It’s one thing to not have access in the first place. This is usually an easy fix in most cases. The really unpleasant access issue is the one when you have confirmed prior access to the instance and then to be completely locked out. More succinctly, you have lost that SQL access!

Whoa is Me!

All hope is now lost right? OK, that isn’t entirely true. Or is it? What if everybody else from the team is also locked out and there is no known sysadmin account. In essence everybody is locked out from managing the instance and now you have a real crisis, right? Well, not so fast. You can still get back in to the instance with sysadmin access. It should be no real secret that you could always restart the SQL instance in single-user mode. Then again, that probably means bigger problems if the server is a production server and is still servicing application requests just fine.

What to do? What to do?

Restart Prohibited

If you really cannot cause a service disruption to bounce the server into single-user mode, my friend Argenis Fernandez (b | t) has this pretty nifty trick that could help you. Truth be told, I have tested that method (even on SQLExpress) several times and it is a real gem. Is this the only alternative?

Let’s back it up just a step or two first. Not having access to SQL Server is in no way the same thing as not having access to the server. Many sysadmins have access to the windows server. Many DBAs also have access to the Windows server or can at least work with the sysadmins to get access to the Windows server in cases like this. If you have admin access to windows – then not much is really going to stop you from gaining access to SQL on that same box. It is a matter of how you approach the issue. Even to restart SQL Server in single-user mode, you need to have access to the Windows server. So, please keep that in mind as you read the article by Argenis as well as the following.

Beyond the requirement of having local access to the server, one of the things that may cause heartburn for some is the method of editing the registry as suggested by Argenis. Modifying the registry (in this case) is not actually terribly complex but it is another one of those changes  that must be put back the way it was. What if there was another way?

As luck would have it, there is an alternative (else there wouldn’t be this article). It just so happens, this alternative is slightly less involved (in my opinion). Let’s start with a server where I don’t have SQL access (beyond public) but I do have Windows access.

We can see on this SQLExpress instance on the TF server that my “Jason” does not exist. Since I don’t have access, I can’t add my own account either. Time to fix that. In order to fix it, I am going to create Scheduled task in Windows that will run a SQLCMD script from my C:\Database folder. The folder can be anywhere, but I generally have one with scripts and such somewhere on each server that I can quickly access.

From here, you will want to click on the “Change User or Group” button to change it to an account that does have access to SQL Server. The account that I use is not a “user” account but rather it is a “system” account called “NT AUTHORITY\SYSTEM” that is present all the way through SQL Server 2017.

To locate the “NT AUTHORITY\SYSTEM” account, just type “SYSTEM” into the new window and click “Check Names”. The account will resolve and then you can click OK out of the “Select User or Group” window.

With the account selected that will run this task, we can now focus our attention on the guts of the task. We will now go to the “Actions” tab.

Click the new button, and here we will configure what will be done.

I do recommend putting the full path to SQLCMD into “Program/Script” box. Once entered, you will add the following to the parameter box.

If you do not have an instance, then just the server name will suffice after the -S parameter. The -i parameter specifies the path to the SQL script file that will be created and placed in the C:\database directory (or whichever directory you have chosen).

That is it for the setup of the task. Now let’s look at the guts of the script file.

Save that into a script document named myscript.sql in the aforementioned directory and then execute the windows task. After executing the Windows task, it is time to verify if it worked or not.

Boom! From no access to a sysadmin in a matter of seconds. Here is that quick verify script – generalized.

The Wrap

Losing access to a SQL instance is never a desirable situation – for the DBA. When the people that are supposed to have access, lose that access, all hope is not lost. There are plenty of methods available to regain the requisite access to manage the server. Today, I shared one such method that I view as being extremely easy. If you lose access, I would recommend taking the steps shown in this article to regain that access.

While not in the back to basics series, I do recommend checking out my other posts in that series. Some topics in the series include (but are not limited to): Backups, backup history and user logins. I would also recommend reading this audit article. If you are able to elevate your permissions, then obviously anybody with server access can elevate their permissions too. For that reason, you should regularly audit the permissions and principals in SQL Server.

«page 1 of 118

Calendar
July 2018
M T W T F S S
« Jun    
 1
2345678
9101112131415
16171819202122
23242526272829
3031  

Welcome , today is Sunday, July 15, 2018