Setting Min and Max Server Memory Setting

Configuring the min and max server memory settings for SQL Server has a number of things to consider prior to making changes. Before we jump into those considerations, let’s learn what the min and max server memory options do.

 

MIN SERVER MEMORY

The min server memory option specifies the minimum amount of memory utilized by the SQL Server buffer pool. It may seem that this should mean that when SQL Server starts it will automatically grab the memory equal to the min server memory value. However, this is not what happens. Instead, SQL Server acquires system memory as needed to process a workload. Consequently, if the SQL Server workload remains low, it is possible that the min server memory value is never reached. However, once it is reached SQL Server will not release an amount of memory that would take it below that minimum setting.

For example, let’s suppose your server has 128 GB of RAM and the min server memory option were set to 32 GB for SQL Server. SQL Server does not immediately allocate 32 GB of RAM upon serve startup. It will acquire memory as needed for the workload. This means that if there is little activity in the database it may take some time before SQL Server allocates 32 GB of RAM. Once this happens, SQL Server will not release memory that would reduce the allocation below the min serve memory setting.

 

MAX SERVER MEMORY

To a certain point the max server memory setting does do what it sounds like. Keep in mind this setting is applicable only to the buffer pool. SQL Server can allocate some memory beyond this max number as it consumes memory for things outside of the buffer pool. Also, realize that this setting does not account for memory allocated to other components of SQL Server that may be installed and running, such as Reporting Services or Analysis Services.

 

The default value for Max Server Memory is 2,147,483,647 in MB. That number is rather large, 2,047.99 GB to be exact. So what happens if the machine SQL Server is installed on does not have that much memory and the default was not changed? As previously mentioned, SQL Server will continue to acquire additional memory based on workload. As a result, SQL Server could make requests of Windows that puts the box under memory pressure as it essentially attempts to choke the OS by leaving less and less memory for the OS. There are mechanisms in the OS that allow it to page out to disk some information in the SQL Server buffer pool so Windows can recover that memory for itself. This hurts SQL Server performance because disk IO is far slower than keeping operations in memory. Among other places, details on how this process works can be found here, which is a blog post discussing the Lock Pages In Memory feature. I will save that setting for discussion a later post.

 

The point I want to make here is that you can calculate how much memory the Windows OS and other applications may need and then subtract that from the physical memory available. There are recommendations on the internet for such things. Whatever number you arrive at would then be set as the Max Server Memory setting in SQL Server. The Available MB perfmon counter should then be judiciously monitored to ensure that there is adequate free memory left on the box, at least 1-2 GB. That will allow plenty of memory for that person who just insists on using an RDP connection to do his or her work or for a new process to fire off without bringing the server to its knees.

 

For example, let’s say that your server has 128 GB of RAM. You may decide to allocate 12 GB for the OS. Perhaps you have a vendor app that needs 4 GB. You take 128-(12+4) and that leaves 112 GB of RAM. If you are running SQL Server 2008 R2 then the Max Server Memory setting can be set to a max value of 64 GB. That is per SQL instance. SQL Server 2012 Standard supports a maximum of 64 GB and SQL Server 2014 Standard supports 128 GB. If SQL Server 2014 Standard were installed in this scenario there would be a maximum of 112 GB that could potentially be allocated to the SQL Server buffer pool for the machine described above.   Setting this limit at 112 GB, followed by the aforementioned monitoring of the Available MB performance monitor counter, would help ensure that SQL Server does not hog memory from the Windows operating system.

 

Now that there is a foundational understanding of what the Min Server Memory and Max Server memory settings are and how they work, let’s look at how to set these options in SQL Server.

 

First, there is the GUI. Open SQL Server Management Studio and connect to a SQL Server instance. Right click on the SQL Server instance name and then choose Properties. Now select the Memory page in the left pane  Manually change the values in the Minimum and Max Server Memory boxes.  Remember that this is in MB.  Then click OK in the lower right of the dialog box.  Also, remember that when you make changes in SSMS you can click the Script button at the top of the dialog and T-SQL for the particular changes will be generated for you.

 

Min and Max Server Settings

 

Second, these settings can be changed using T-SQL.  Here is an example script that sets the minimum to 12 GB and the Max to 32 GB.

 

EXEC sys.sp_configure N’show advanced options’, N’1′ RECONFIGURE WITH OVERRIDE

GO

EXEC sys.sp_configure N’min server memory (MB)’, N’12288′

GO

EXEC sys.sp_configure N’max server memory (MB)’, N’32768′

GO

RECONFIGURE WITH OVERRIDE

GO

EXEC sys.sp_configure N’show advanced options’, N’0′ RECONFIGURE WITH OVERRIDE

GO

 

I encourage you to review these settings on your SQL Server instance and consider whether they need to be adjusted.

Use Windows Groups to Manage SQL Server

Ever come across a SQL Server that had 5, 10, or more individual logins for users in it? I have a few times and it made me wonder if whoever had done such a thing realized how much harder they were making the administration of the server by doing this. I suppose some folks start out with one or two people that they give special permissions or access directly to the SQL instance and underlying databases. Their logic probably goes something like, “It’s just Tommy and Suzie. Those are the only people who will need access for a long time.”

Initially I have to agree with this thought. It does seem simpler when you only have a couple of people to simply create a login and corresponding user for those people. However, SQL Server access and permissions rarely stay that simple. Hopefully you are in a growing company. A growing company will eventually employ additional people who either truly need or believe that they need access to the database. If you are already in the habit of granting direct access using individual logins and database users, it will be tempting to continue this pattern. Before you know it, one day you connect to SQL Server and there is an ever increasing list of individual logins that suddenly catches your attention, and most likely there are a variety of permissions applied to those logins and users. This becomes a management nightmare.

In steps Windows groups. In a typical environment there will be Developers, one or more DBA’s and perhaps folks who write SSRS reports or who do ad hoc type reporting. These groups most likely all need different permission sets. Create a Windows group in AD, or have your System Admin do it. (You are on good terms with her, right?) Ask the SysAdmin to add the developers to a MyDomain\Dev group. Ask her for MyDomain\SQLAdmins and perhaps MyDomain\Reporting.

Now in SQL Server add these Logins and map them to appropriate database Users. For the Dev group maybe you give them only Public permissions in production at the SQL instance and db_datareader to various databases. Now, do the same for MyDomain\Reporting. Why have two groups with the same permissions you ask? Well, it’s a management thing. See if you throw them all into the same group when people on the Dev team need slightly different permissions for something now that permission change would be applied to your reporting people too, and you may not want that.

Back in SQL Server 2008 the BUILTIN Administrators group was removed from SQL Server as a group that was granted automatic access to the SQL Server instance. This was done to enhance security. Given this change, you might want to create a group specifically for your DBA’s and grant them the sysadmin server role. This is where MyDomain\SQLAdmins come in. You can grant them the needed access to manage the SQL instance without making those folks admins on the machine. The Windows SysAdmins can retain control of the machine without automatically being granted control in the SQL instance to do things that the DBA’s would frown upon. Everyone is happy, right?

Now let’s suppose that the company has a new project and three Developers are being hired. In the past the DBA might make three separate logins and map those logins to database users. Now that MyDomain\Dev exists just ask your System Administrator to add those new developers to that group while they are in Active Directory creating Windows accounts for the new employees and presto, permissions are set before the new employees show up on the first day. Prod permissions are set and you have of course already added the MyDomain\Dev group to the development SQL instance in the sysadmin server role so permissions are set there as well. All of that without ever touching SQL Server to grant individual permissions to people. You’ve just saved yourself some time so you can go do something else.

I hope this has been insightful and encourages you to start thinking ahead now, even while your SQL Server environment is still relatively small. I think starting with Windows groups from the beginning will save you time and headache later.

3 Additional Tips for Using SQL Server Management Studio

In an earlier post I listed three tips for using the SQL Server Management Studio user interface. I want to offer three additional tips for SSMS.  I think you will find that these ideas will bring additional efficiency and functionality to your use of the interface.

The first tip is related to getting more viewing space after running a query. Query results obviously take up the bottom portion of the SSMS editor window. I have often found myself then clicking in the query editor window and scrolling to view various parts of my query, all the while the results portion of the screen is using a good portion of the window. Some time back an associate told me how to easily address that annoying issue. Simply click CTRL and R to collapse the query results so that the full page is again used for viewing the query text.

The second tip is related to finding objects in Object Explorer. Ever found yourself trying to locate a particular stored procedure that has a name that is a mile long and your brain only remembers part of the procedure name? What about trying to find all the tables related to Orders or Shipping? As it turns out, there is a nice way to search for such things.

Drill down into the object level you are searching through, such as Databases > AdventureWorks2008R2 > Tables. Right click the Tables folder and select Filter. Then select Filter Settings. You are then presented with the dialog below. Enter the name description in the Value column and click OK. This filtering also works for the Views, Stored Procedures and Functions folders.

Filter Search on a Table

 

The third tip relates to using color to visually identify what SQL Server instance the query window is connected to. I once saw a Developer who had five SQL Server Management Studio windows open at once. When another Developer asked him why he answered that it was so he could maintain separate connections to different SQL instances and not be as confused as if he had all the connections in one running instance of SQL Server. It was, I suppose, an admirable attempt to ensure he was running the query against the right SQL instance, but it caused me to remember that color can be used for identifying the SQL instance.

Here is how it works. When SSMS is launched the user is presented with a connection dialog.

SSMS Connect to Server Dialog2

 

Next select Options and the below appears.

 

SSMS Connection Properties Dialog

Click on the Select button, choose a color and then click OK and then Connect in order to connect to the SQL instance. Now open a new query window. The color you just selected will appear along the bottom of the SSMS editor window. To do this after you have already opened SSMS, click Connect > Database Engine > Options from the Object Explorer pane. This type of color customization can also be done with registered servers. For more information on this feature click this link: https://msdn.microsoft.com/en-us/library/hh213617(v=sql.120).aspx

 

I hope these additional tips on SSMS prove useful. Feel free to leave me a comment or another customization that you like to use.

You Really Should Stay Current on Cumulative Updates

 

I have encountered a Log Shipping error that I want to detail as an encouragement for staying current on SQL Server cumulative updates. First, a little background. The environment is SQL Server 2014 Standard edition with a physical machine as the primary and two secondaries, both virtual machines backed by a flash storage system.   SQL Server is on service pack 1, cumulative update 1. Log shipping had been configured in this environment for many months, but SQL Server 2008R2 was updated to SQL Server 2014 some months prior. The Log Shipping Restore job leaves the database in the Standby Mode so it can be queried for reporting purposes.

 

One evening I received the following error via database mail: “SQL Server Alert System: ’Severity 021’ occurred on \\MySecondary. An error occurred during recovery, preventing the database ‘MyDatabase’ from restarting. Diagnose the recovery errors and fix them or restore from a known good backup.” This was immediately followed by another email advising that the Log Shipping restore job had failed. Yet a third email was sent for error 9004 severity 16, explaining that an error occurred processing the log for “MyDatabase.” The message informed that I could restore from backup or rebuild the log. I connected to the SQL instance and saw that the database was in the restoring mode.

 

A quick internet search turned up the following two links.

https://support.microsoft.com/en-us/kb/2987585

https://support.microsoft.com/en-us/kb/2015753

 

The symptoms section of the first link seemed to fit my situation fairly well. I was unfamiliar with the fsutil command line utility. When I executed it against my environment the results indicated that the Bytes per Sector value was 512 and Bytes per Physical Sector returned “Not Supported.” The support article gave me two workarounds for addressing the issue.

  • Move the transaction log file at the destination to a drive that has “Bytes per Physical Sector” set as 512 bytes.Note The Standby file can still be located on the drive that has “Bytes per Physical Sector” set as 4096 bytes.
  • Restore the log backups without using the standby option. Instead of the STANDBY option, use the WITH NORECOVERY option during the restore operation.

 

Since my database was still in the restoring mode and the second URL indicated I had a damaged transaction log, I could not offline the database and relocate the file per the first recommendation. I also could not confirm how the bytes per sector value on the drive was configured anyway. The first article indicates that the issue was fixed in SQL Server 2014 SP1 CU2 and SQL Server 2012 SP2, CU7. Remember, I mentioned that the environment was on CU1 of SQL Server 2014 SP1. Unfortunately, that put me one update behind with respect to being protected from this situation.

 

Looks like I had one good course of action – restore from backup. I disabled the restore job, restored from a recent full backup and the most recent differential. I left the database in the restoring mode so I could re-enable the Log Shipping restore job. This would get the secondary back to Standby mode.

 

If you want to know when the latest updates for SQL Server are available, then I recommend saving the links below in your favorite web browser.  As this post demonstrates, be sure to read what the latest are about as soon as possible to see if they apply in your situation.

 

http://sqlserverbuilds.blogspot.com/

http://sqlserverupdates.com/

http://sqlblog.com/blogs/aaron_bertrand/default.aspx

 

3 Reasons Not to Install Databases On the C Drive

We’re probably all familiar with the fact that SQL Server by default will install not only the program binaries but all your databases on the C drive as well.  Accidental DBA’s, i.e. System Administrators and Developers who inherited the maintenance of a database, may not be aware of how bad it can be if someone simply clicked the “Next” button all the way through the SQL Server install.  There are surely many reasons not to accept this default during the install, but I will start with three.

First, from a purely practical standpoint, I think installing to the default path makes backups and restore scripts unduly difficult.  Which is easier to remember, “C:\Program Files\Microsoft SQL Server\MSSQL12.MSSQLSERVER\MSSQL\Backup” or something simple like “D:\SQLBackups”? Which one would you rather have to type or copy from Windows Explorer after browsing the path?  With the second one it is so easy you don’t have to spend time browsing and copying it.  You just quickly type it out.  For the restore script would you rather type “C:\Program Files\Microsoft SQL Server\MSSQL12.MSSQLSERVER\MSSQL\Data\SomeDB.mdf” or “D:\SQLData\SomeDB.mdf”?  The same point applies here – Simplicity.  Why make it hard on yourself?

Second, if you don’t put your databases on the C drive then database growth can’t crash the entire computer by filling up the C drive.  We all know that databases grow and that growth has to be monitored and managed.  If the System Administrator misses those emails from Database Mail or her favorite monitoring tool, then she will be awakened first thing in the morning by a company end user who can’t do their regular work only to discover that the entire machine is experiencing paralysis.  If the database files are not on the C drive with the OS then at least database growth will not be the culprit of server paralysis from a full C drive.

Third, if you install all your databases to the C drive, then you are inviting performance issues.  Databases and the OS will be fighting for drive resources.  Disks only spin so fast.  If you have the OS and 10, 14, 50 or more database files all trying to read and write to the same drive, then your C drive is going to be spinning like a yo yo.  Not only is there contention between the OS and the database files, but the database files will be fighting each other for disk usage too.

Let me provide some real numbers from an environment I have seen.  I had a system with 12 databases all on the C drive.  That number includes 4 system databases and 8 user databases.  Latency was in the thousands for several of the databases.  I moved the user databases and TempDB off to another drive and left the System databases on the c drive.  I know I can move those too, but one step at a time.  See the comparisons below.

Database FileName Overall Latency Overall Latency Next Morning Overall Latency 3 Weeks Later
Master.mdf 8330 8287 4
Mastlog.ldf 2409 1607 0
Model.mdf 8129 No Change 3
Modellog.ldf 2881 No Change 1
msdbdata.mdf 32446 16031 5
msdblog.ldf 31270 22969 0
TempDB.mdf 254 Moved to Fusion IO card Moved to Fusion IO card
Templog.ldf 155 Moved to Fusion IO card Moved to Fusion IO card

 

In some cases there was an immediate impact for latency.  Clearly, removing the activity of the eight user databases made a huge difference in performance.  As you can imagine, the user databases performed much better from a latency perspective as well, especially since I put them on a Fusion IO card.

To see latency for your drives you can run the following query from Paul Randal:

 

/*

How to examine IO subsystem latencies from within SQL Server


*/

SELECT
[ReadLatency] =
CASE WHEN [num_of_reads] = 0
THEN 0 ELSE ([io_stall_read_ms] / [num_of_reads]) END,
[WriteLatency] =
CASE WHEN [num_of_writes] = 0
THEN 0 ELSE ([io_stall_write_ms] / [num_of_writes]) END,
[Latency] =
CASE WHEN ([num_of_reads] = 0 AND [num_of_writes] = 0)
THEN 0 ELSE ([io_stall] / ([num_of_reads] + [num_of_writes])) END,
[AvgBPerRead] =
CASE WHEN [num_of_reads] = 0
THEN 0 ELSE ([num_of_bytes_read] / [num_of_reads]) END,
[AvgBPerWrite] =
CASE WHEN [num_of_writes] = 0
THEN 0 ELSE ([num_of_bytes_written] / [num_of_writes]) END,
[AvgBPerTransfer] =
CASE WHEN ([num_of_reads] = 0 AND [num_of_writes] = 0)
THEN 0 ELSE
(([num_of_bytes_read] + [num_of_bytes_written]) /
([num_of_reads] + [num_of_writes])) END,
LEFT ([mf].[physical_name], 2) AS [Drive],
DB_NAME ([vfs].[database_id]) AS [DB],
[mf].[physical_name]
FROM
sys.dm_io_virtual_file_stats (NULL,NULL) AS [vfs]
JOIN sys.master_files AS [mf]
ON [vfs].[database_id] = [mf].[database_id]
AND [vfs].[file_id] = [mf].[file_id]
— WHERE [vfs].[file_id] = 2 — log files
— ORDER BY [Latency] DESC
— ORDER BY [ReadLatency] DESC
ORDER BY [WriteLatency] DESC;
GO

Well, I hope this has given you something to think about.  Your homework is to find out the latency in your environment, do some research, make changes as needed and then follow up to see if the changes you made helped.  Let me know how the homework turns out.

 

3 Tips for Using SQL Server Management Studio

SQL Server Management Studio, also known simply as SSMS, has many options that can be customized to help you get work done.  In this post I will cover three such options that I have found useful.

The first option discussed will be line numbers. Line numbers can be added to the query editor window to help identify sections of code.  To do this click the Tools menu.  In the drop down then select Options.  Then under Text Editor left click Transact SQL.  Now several options will be available related to what happens inside the text editor screen.  Left click the box next to Line Numbers to enable the display of line numbers next to your code in the editor.

Line Numbers Option in SSMS

 

The second option is query shortcuts. Ever find yourself retyping commands like sp_who2, sp_helpindex from Kimberly Tripp or other handy procedures?  You can actually execute these by assigning them to a keyboard shortcut rather than having to type them each time you want to run the procedure.

To do this, go back to the Tools and Options menu.  Choose Environment, then Keyboard, then Query Shortcuts.  The left column will be labeled Shortcuts.  To the right of the shortcut you wish to use simlpy left click and type the command you want to assign.  Then click OK to assign that shortcut.  Presto!  Now the next time you want to run that favorite metadata proc, just use the assigned shortcut.

SSMS Query Shortcuts

 

The third option is one of my favorite options to adjust. It is the number of recent files and projects that SSMS displays.  This list gives you access to recently opened files and projects.  The default is 10 but I don’t always find that 10 is enough for me.  To adjust this, go to Tools, then Options and Environment.  In the Recent Files section, the number can be adjusted up or down to suit your preference.

 

SSMS Recent Files Option