SQLQueryStress is a simple, lightweight performance testing tool, designed to load test individual queries. It includes support for randomization of input parameters in order to test cache repeatability, and includes basic capabilities for reporting on consumed server resources. If you cannot convince your business of the need of a full option stress test scenario, there is still plan B: Only the database will be stress tested. No third party tool required. This is not ideal but focusing on the database for a stress test scenario is not such a bad idea.
This article describes how to use the SQLIOSim utility to perform stress tests on disk subsystems to simulate SQL Server activity.
Original product version: SQL Server
Original KB number: 231619
For Microsoft SQL Server 2005, SQLIOSim was shipped as a separate download package. Starting with SQL Server 2008, SQLIOSim is included with the SQL Server product installation. When you install SQL Server, you find the SQLIOSim tool in the BINN folder of your SQL Server installation. It is recommended to use these updated versions of the tool to simulate the IO activity on the disk subsystem.
The SQLIOSim utility replaces the SQLIOStress utility. The SQLIOStress utility was formerly known as the SQL70IOStress utility.
This article also contains download information for the SQLIOSim utility.
This article describes the SQLIOSim tool. You can use this tool to perform reliability and integrity tests on disk subsystems. These tests simulate read, write, checkpoint, backup, sort, and read-ahead activities for Microsoft SQL Server. However, if you have to perform benchmark tests and determine the I/O capacity of the storage system, you should use the Diskspd tool.
The SQLIOSim utility has been upgraded from the SQLIOStress utility. The SQLIOSim utility more accurately simulates the I/O patterns of Microsoft SQL Server.
For more information about SQL Server I/O patterns, see SQL Server I/O Basics, Chapter 2.
Note
To help maintain appropriate data integrity and security, we recommend that you perform stress tests of your I/O subsystem before you deploy SQL Server on new hardware. The SQLIOSim utility simulates the read patterns, the write patterns, and the problem identification techniques of SQL Server. To perform these tasks, the SQLIOSim utility simulates the user activity and the system activity of a SQL Server system. The SQLIOSim utility performs this simulation independent of the SQL Server engine.
The SQLIOSim utility does not guarantee or warrant data security or integrity. The utility was designed to provide baseline testing of a system environment. The SQLIOSim utility may expose potential data integrity issues.
For more information about logging and data storage, see Description of logging and data storage algorithms that extend data reliability in SQL Server.
The download package contains two executable files, SQLIOSim.com and SQLIOSim.exe. Both executable files provide identical simulation capabilities. SQLIOSim.com is a command-line tool that you can configure to run without user interaction. To do this, you can use command-line parameters, a configuration file, or a combination of both of these methods. SQLIOSim.exe is a graphical application that accepts no command-line parameters. However, SQLIOSim.exe does load default configuration data from configuration files.
SQLIOSim.com accepts a limited number of command-line parameters to control basic behavior. The configuration file for the SQLIOSim utility provides advanced behavior control. When command-line parameters and configuration file options overlap, the command-line parameters take precedence.
Parameter | Comment |
---|---|
-cfg file | Override the Sqliosim.cfg.ini default configuration file. The SQLIOSim utility returns an error if the utility cannot find the file. |
-save file | Save the resulting configuration in the configuration file. You can use this option to create the initial configuration file. |
-log file | Specify the error log file name and the error log file path. The default file name is Sqliosim.log.xml. |
-dir dir | Set the location to create the data (.mdf) file and the log (.ldf) file. You can run this command multiple times. In most cases, this location is a drive root or a volume mount point. This location can be a long path or a UNC path. |
-d seconds | Set the duration of the main run. This value excludes the preparation phase and the verification phase. |
-size MB | Set the initial size of the data file in megabytes (MB). The file can grow up to two times the initial size. The size of the log file is calculated as half the size of the data file. However, the log file cannot be larger than 50 MB. |
Sample configuration files for various tests can be downloaded from SQL Server support team’s github repo here.
You do not have to use a configuration file. If you do not use a configuration file, all parameters take default values except the data file location and the log file location. You must use one of the following methods to specify the data file location and the log file location:
Note
CacheHitRatio
parameter is 10 percent. This value is expressed as 1000 because 10, divided by 0.01, equals 1000. The maximum value of a percentage parameter is 10000.Boolean
, the valid values that you can assign to the parameter are true and false . Additionally, the values are case sensitive. The SQLIOSim utility ignores any invalid values.MinIOChainLength
parameter must not be larger than the value of the MaxIOChainLength
parameter.The SQLIOSim utility takes the values that you specify in the CONFIG section of the SQLIOSim configuration file to establish global testing behavior.
Parameter | Default value | Description | Comments |
---|---|---|---|
ErrorFile | sqliosim.log.xml | Name of the XML type log file | |
CPUCount | Number of CPUs on the computer | Number of logical CPUs to create | The maximum is 64 CPUs. |
Affinity | 0 | Physical CPU affinity mask to apply for logical CPUs | The affinity mask should be within the active CPU mask. A value of 0 means that all available CPUs will be used. |
MaxMemoryMB | Available physical memory when the SQLIOSim utility starts | Size of the buffer pool in MB | The value cannot exceed the total amount of physical memory on the computer. |
StopOnError | true | Stops the simulation when the first error occurs | |
TestCycles | 1 | Number of full test cycles to perform | A value of 0 indicates an infinite number of test cycles. |
TestCycleDuration | 300 | Duration of a test cycle in seconds, excluding the audit pass at the end of the cycle | |
CacheHitRatio | 1000 | Simulated cache hit ratio when the SQLIOSim utility reads from the disk | |
MaxOutstandingIO | 0 | Maximum number of outstanding I/O operations that are allowed process-wide | The value cannot exceed 140000. A value of 0 means that up to approximately 140,000 I/O operations are allowed. This is the limit of the utility. |
TargetIODuration | 100 | Duration of I/O operations, in milliseconds, that are targeted by throttling | If the average I/O duration exceeds the target I/O duration, the SQLIOSim utility throttles the number of outstanding I/O operations to decrease the load and to improve I/O completion time. |
AllowIOBursts | true | Allow for turning off throttling to post many I/O requests | I/O bursts are enabled during the initial update, initial checkpoint, and final checkpoint passes at the end of test cycles. The MaxOutstandingIO parameter is still honored. You can expect long I/O warnings. |
NoBuffering | true | Use the FILE_FLAG_NO_BUFFERING option | SQL Server opens database files by using FILE_FLAG_NO_BUFFERING true. Some utilities and services, such as Analysis Services, use FILE_FLAG_NO_BUFFERING false. To fully test a server, execute one test for each setting. |
WriteThrough | true | Use the FILE_FLAG_WRITE_THROUGH option | SQL Server opens database files by using FILE_FLAG_WRITE_THROUGH true. However, some utilities and services open the database files by using FILE_FLAG_WRITE_THROUGH false. For example, SQL Server Analysis Services opens the database files by using FILE_FLAG_WRITE_THROUGH false. To fully test a server, execute one test for each setting. |
ScatterGather | true | Use ReadScatter/WriteGather APIs | If this parameter is set to true, the NoBuffering parameter is also set to true. SQL Server uses scatter/gather I/Os for most I/O requests. |
ForceReadAhead | true | Perform a read-ahead operation even if the data is already read | The SQLIOSim utility issues the read command even if the data page is already in the buffer pool. Microsoft SQL Server Support has successfully used the true setting to expose I/O problems. |
DeleteFilesAtStartup | true | Delete files at startup if files exist | A file may contain multiple data streams. Only streams that are specified in the File x FileName entry are truncated in the file. If the default stream is specified, all streams are deleted. |
DeleteFilesAtShutdown | false | Delete files after the test is finished | A file may contain multiple data streams. Only data streams that you specify in the File x FileName entry are truncated in the file. If the default data stream is specified, the SQLIOSim utility deletes all data streams. |
StampFiles | false | Expand the file by stamping zeros | This process may take a long time if the file is large. If you set this parameter to false, the SQLIOSim utility extends the file by setting a valid data marker. SQL Server 2005 uses the instant file initialization feature for data files. If the data file is a log file, or if instant file initialization is not enabled, SQL Server performs zero stamping. Versions of SQL Server earlier than SQL Server 2000 always perform zero stamping. You should switch the value of the StampFiles parameter during testing to make sure that both instant file initialization and zero stamping are operating correctly. |
The SQLIOSim utility is designed to allow for multiple file testings. The File x section is represented as [File1], [File2] for each file in the test.
Parameter | Default value | Description | Comments |
---|---|---|---|
FileName | No default value | File name and path | The FileName parameter can be a long path or a UNC path. It can also include a secondary stream name and type. For example, the FileName parameter may be set to file.mdf:stream2. NOTE In SQL Server 2005, DBCC operations use streams. We recommend that you perform stream tests. |
InitialSize | No default value | Initial size in MB | If the existing file is larger than the value that is specified for the InitialSize parameter, the SQLIOSim utility does not shrink the existing file. If the existing file is smaller, the SQLIOSim utility expands the existing file. |
MaxSize | No default value | Maximum size in MB | A file cannot grow larger than the value that you specify for the MaxSize parameter. |
Increment | 0 | Size in MB of the increment by which the file grows or shrinks. For more information, see the 'ShrinkUser section' part of this article. | The SQLIOSim utility adjusts the Increment parameter at startup so that the following situation is established:Increment * MaxExtents < MaxMemoryMB / NumberOfDataFiles If the result is 0, the SQLIOSim utility sets the file as non-shrinkable. |
Shrinkable | false | Indicates whether the file can be shrunk or expanded | If you set the Increment parameter to 0, you set the file to be non-shrinkable. In this case, you must set the Shrinkable parameter to false. If you set the Increment parameter to a value other than 0, you set the file to be shrinkable. In this case, you must set the Shrinkable parameter to true. |
Sparse | false | Indicates whether the Sparse attribute should be set on the files | For existing files, the SQLIOSim utility does not clear the Sparse attribute when you set the Sparse parameter to false. SQL Server 2005 uses sparse files to support snapshot databases and the secondary DBCC streams. We recommend that you enable both the sparse file and the streams, and then perform a test pass. NOTE If you set Sparse = true for the file settings, do not specify NoBuffering = false in the config section. If you use these two conflicting combinations, you may receive an error that resembles the following from the tool: Error:-Error: 0x80070467 Error Text: While accessing the hard disk, a disk operation failed even after retries. Description: Buffer validation failed on C:SQLIOSim.mdx Page: 28097 |
LogFile | false | Indicates whether a file contains user or transaction log data | You should define at least one-log file. |
The SQLIOSim utility takes the values that you specify in the RandomUser section to simulate a SQL Server worker that is performing random query operations, such as Online Transaction Processing (OLTP) I/O patterns.
Parameter | Default value | Description | Comments |
---|---|---|---|
UserCount | -1 | Number of random access threads that are executing at the same time | The value cannot exceed the following value: CPUCount1023-100 The total number of all users also cannot exceed this value. A value of 0 means that you cannot create random access users. A value of -1 means that you must use the automatic configuration of the following value: min(CPUCount2, 8) NOTE A SQL Server system may have thousands of sessions. Most of the sessions do not have active requests. Use the count(*) function in queries against the sys.dm_exec_requests dynamic management view (DMV) as a baseline for establishing this test parameter value.CPUCount here refers to the value of the CPUCount parameter in the CONFIG section. The min(CPUCount*2, 8) value results in the smaller of the values between CPUCount*2 and 8. |
JumpToNewRegionPercentage | 500 | The chance of a jump to a new region of the file | The start of the region is randomly selected. The size of the region is a random value between the value of the MinIOChainLength parameter and the value of the MaxIOChainLength parameter. |
MinIOChainLength | 1 | Minimum region size in pages | |
MaxIOChainLength | 100 | Maximum region size in pages | SQL Server 2005 Enterprise Edition and SQL Server 2000 Enterprise Edition can read ahead up to 1,024 pages. The minimum value is 0. The maximum value is limited by system memory. Typically, random user activity causes small scanning operations to occur. Use the values that are specified in the ReadAheadUser section to simulate larger scanning operations. |
RandomUserReadWriteRatio | 9000 | Percentage of pages to be updated | A random-length chain is selected in the region and may be read. This parameter defines the percentage of the pages to be updated and written to disk. |
MinLogPerBuffer | 64 | Minimum log record size in bytes | The value must be either a multiple of the on-disk sector size or a size that fits evenly into the on-disk sector size. |
MaxLogPerBuffer | 8192 | Maximum log record size in bytes | This value cannot exceed 64000. The value must be a multiple of the on-disk sector size. |
RollbackChance | 100 | The chance that an in-memory operation will occur that causes a rollback operation to occur. | When this rollback operation occurs, SQL Server does not write to the log file. |
SleepAfter | 5 | Sleep time after each cycle, in milliseconds |
The SQLIOSim utility takes the values that you specify in the AuditUser section to simulate DBCC activity to read and to audit the information about the page. Validation occurs even if the value of the UserCount parameter is set to 0.
Parameter | Default value | Description | Comments |
---|---|---|---|
UserCount | 2 | Number of Audit threads | The value cannot exceed the following value: CPUCount1023-100 The total number of all users also cannot exceed this value. A value of 0 means that you cannot create random access users. A value of -1 means that you must use the automatic configuration of the following value: min(CPUCount2, 8) NOTE A SQL Server system may have thousands of sessions. Most of the sessions do not have active requests. Use the count(*) function in queries against the sys.dm_exec_requests DMV as a baseline for establishing this test parameter value.CPUCount here refers to the value of the CPUCount parameter in the CONFIG section.The min(CPUCount*2, 8) value results in the smaller of the values between CPUCount*2 and 8. |
BuffersValidated | 64 | ||
DelayAfterCycles | 2 | Apply the AuditDelay parameter after the number of BuffersValidated cycles is completed | |
AuditDelay | 200 | Number of milliseconds to wait after each DelayAfterCycles operation |
The SQLIOSim utility takes the values that are specified in the ReadAheadUser section to simulate SQL Server read-ahead activity. SQL Server takes advantage of read-ahead activity to maximize asynchronous I/O capabilities and to limit query delays.
Parameter | Default value | Description | Comments |
---|---|---|---|
UserCount | 2 | Number of read-ahead threads | The value cannot exceed the following value: CPUCount1023-100 The total number of all users also cannot exceed this value. A value of 0 means that you cannot create random access users. A value of -1 means that you must use the automatic configuration of the following value: min(CPUCount2, 8) NOTE A SQL Server system may have thousands of sessions. Most of the sessions do not have active requests. Use the count(*) function in queries against the sys.dm_exec_requests DMV as a baseline for establishing this test parameter value.CPUCount here refers to the value of the CPUCount parameter in the CONFIG section. The min(CPUCount*2, 8) value results in the smaller of the values between CPUCount*2 and 8. |
BuffersRAMin | 32 | Minimum number of pages to read per cycle | The minimum value is 0. The maximum value is limited by system memory. |
BuffersRAMax | 64 | Maximum number of pages to read per cycle | SQL Server Enterprise editions can read up to 1,024 pages in a single request. If you install SQL Server on a computer that has lots of CPU, memory, and disk resources, we recommend that you increase the file size and the read-ahead size. |
DelayAfterCycles | 2 | Apply the RADelay parameter after the specified number of cycles is completed | |
RADelay | 200 | Number of milliseconds to wait after each DelayAfterCycles operation |
The SQLIOSim utility takes the values that you specify in the BulkUpdateUser section to simulate bulk operations, such as SELECT...INTO operations and BULK INSERT operations.
Parameter | Default value | Description | Comments |
---|---|---|---|
UserCount | -1 | Number of BULK UPDATE threads | The value cannot exceed the following value: CPUCount*1023-100 A value of -1 means that you must use the automatic configuration of the following value: min(CPUCount*2, 8) NOTE A SQL Server system may have thousands of sessions. Most of the sessions do not have active requests. Use the count(*) function in queries against the sys.dm_exec_requests DMV as a baseline for establishing this test parameter value.CPUCount here refers to the value of the CPUCount parameter in the CONFIG section.The min(CPUCount*2, 8) value results in the smaller of the values between CPUCount*2 and 8. |
BuffersBUMin | 64 | Minimum number of pages to update per cycle | |
BuffersBUMax | 128 | Maximum number of pages to update per cycle | The minimum value is 0. The maximum value is limited by system memory. |
DelayAfterCycles | 2 | Apply the BUDelay parameter after the specified number of cycles is completed | |
BUDelay | 10 | Number of milliseconds to wait after each DelayAfterCycles operation |
The SQLIOSim utility takes the values that you specify in the ShrinkUser section to simulate DBCC shrink operations. The SQLIOSim utility can also use the ShrinkUser section to make the file grow.
Parameter | Default value | Description |
---|---|---|
MinShrinkInterval | 120 | Minimum interval between shrink operations, in seconds |
MaxShrinkInterval | 600 | Maximum interval between shrink operations, in seconds |
MinExtends | 1 | Minimum number of increments by which the SQLIOSim utility will grow or shrink the file |
MaxExtends | 20 | Maximum number of increments by which the SQLIOSim utility will grow or shrink the file |
The semicolon character (;) at the start of a line in the configuration .ini file causes the line to be treated as a single comment.
The SQLIOSim utility creates separate data files and log files to simulate the I/O patterns that SQL Server generates in its data file and in its log file. The SQLIOSim utility does not use the SQL Server engine to perform stress activity. Therefore, you can use the SQLIOSim utility to test a computer before you install SQL Server.
When you run the SQLIOSim utility, make sure that you specify the same file location that you use for your SQL Server database files. When you do this, the utility simulates the same I/O path as your SQL Server database.
You can enable the compress attribute or the encrypt attribute for the existing test files. You can also enable these attributes for the existing directory where the test files will be created. The corresponding options to enable these attributes are located in the Properties dialog box for a file or for a directory.
By default, the SQLIOSim utility creates test files that have the .mdx and.ldx file name extensions. Therefore, these files will not overwrite existing data and log files.
Warning
Do not specify the actual SQL Server database files for testing. The SQLIOSim utility will overwrite the data with random test patterns, and your actual SQL Server data will be lost.
The SQLIOSim utility creates the error log file in one of the following locations:
The SQLIOSim.log.xml error log contains details about the execution. These details include error information. Review the log carefully for error information and for warning information.
Note
If you experience an error in the SQLIOSim utility, we recommend that you ask your hardware manufacturer to help determine the root cause of the issue.
The SQLIOSim utility accommodates multiple-file-level testing and multiple-user-level testing. The SQLIOSim utility does not require multiple invocations. However, the SQLIOStress utility requires multiple invocations. You can run multiple copies of the SQLIOSim utility if the following conditions are true:
MaxMemoryMB
parameter of each instance provides for a non-overlapping memory region that is sufficient for each instance.The sum of the MaxMemoryMB
parameter for each instance must be less than or equal to the total physical memory. Some testing phases, such as checkpoint simulation, can be memory-intensive and may create out-of-memory conditions when you run multiple copies. If you experience out-of-memory errors, you can reduce the number of utility copies that are running.
In addition to the default Sqliosim.cfg.ini file, the package provides the following sample files.
Sample file | Description | Parameters that differ from the default configuration file |
---|---|---|
Sqliosim.hwcache.cfg.ini | Minimize reads Files are made small to keep them fully in memory No sequential reads | For the AuditUser section and for the ReadAheadUser section: CacheHitRatio=10000 UserCount=0 |
Sqliosim.nothrottle.cfg.ini | Remove I/O throttling Minimize the time to wait to increase I/O volume | TargetIODuration=1000000 AuditDelay=10 RADelay=10 |
Sqliosim.seqwrites.cfg.ini | Minimize reads Files are made small to keep them fully in memory Files are made non-shrinkable No sequential reads No random access Bulk update in large chunks without delays | Shrinkable=FALSE For the AuditUser section, for the ReadAheadUser section, and for the RandomUser section: CacheHitRatio=10000 ForceReadAhead=FALSE BuffersBUMin=600 BuffersBUMax=1000 BUDelay=1 UserCount=0 |
Sqliosim.sparse.cfg.ini | Use only 32 MB of memory Make target I/O duration large enough to enable many outstanding I/O requests Disable scatter/gather APIs to issue separate I/O requests for every 8-KB page Create a 1-GB non-shrinkable file Create a 1-GB non-shrinkable secondary sparse stream in the file | MaxMemoryMB=32 TestCycles=3 TestCycleDuration=600 TargetIODuration=10000 UseScatterGather=FALSE [File1] FileName=sqliosim.mdx InitialSize=1000 MaxSize=1000 Increment=10 Shrinkable=FALSE LogFile=FALSE Sparse=FALSE [File2] FileName=sqliosim.ldx InitialSize=50 MaxSize=50 Increment=0 Shrinkable=FALSE LogFile=TRUE Sparse=FALSE [File3] FileName=sqliosim.mdx:replica InitialSize=1000 MaxSize=1000 Increment=10 Shrinkable=FALSE LogFile=FALSE Sparse=TRUE |