Vikas Rana https://docs.microsoft.com/archive/blogs/vikas_rana/feed.xml Wed, 17 Apr 2019 03:13:05 GMT en-US hourly 1 Query Store Entering Error State https://docs.microsoft.com/archive/blogs/vikas_rana/query-store-entering-error-state Wed, 27 Jun 2018 14:22:03 GMT https://blogs.msdn.microsoft.com/vikas_rana/?p=315 sp_query_store_consistency_check stored procedure within the affected database. Please remember if QDS in ERROR state in that case it may affect Automatic plan correction feature and may cause performance issue if you are using plan correction feature extensively. For CRITICAL environment it’s very important to automatically detect this situation and mitigate immediately. You can consider creating a scheduled job to check the QDS status and take appropriate action based on actual state. Here is the sample code we have created which you can test\customize based on your environment\requirement. As always please test following code in TEST\UAT environment before deploying on production. If you would like to improve following T-SQL code then please update here: https://github.com/vikasrana2006/Query-Store-Entering-Error-State/blob/master/QDS_ERROR_STATE_CORRECTION.sql  
IF EXISTS (SELECT * FROM sys.database_query_store_options WHERE actual_state=3)
BEGIN
BEGIN TRY
ALTER DATABASE [QDS] SET QUERY_STORE = OFF
Exec [QDS].dbo.sp_query_store_consistency_check
ALTER DATABASE [QDS] SET QUERY_STORE = ON
ALTER DATABASE [QDS] SET QUERY_STORE (OPERATION_MODE = READ_WRITE)
END TRY
BEGIN CATCH
SELECT 
    ERROR_NUMBER() AS ErrorNumber 
    ,ERROR_SEVERITY() AS ErrorSeverity 
    ,ERROR_STATE() AS ErrorState 
    ,ERROR_PROCEDURE() AS ErrorProcedure 
    ,ERROR_LINE() AS ErrorLine 
    ,ERROR_MESSAGE() AS ErrorMessage;
END CATCH;  
END
    Vikas Rana | Twitter | Linkedin | Escalation Engineer Microsoft IGTSC
 ]]> Query Store Best Practices for Mission Critical or Busy SQL Server https://docs.microsoft.com/archive/blogs/vikas_rana/query-store-best-practices-for-mission-critical-or-busy-sql-server Fri, 20 Oct 2017 16:18:19 GMT https://blogs.msdn.microsoft.com/vikas_rana/?p=305 Query Store feature helping lot of SQL Admin in isolating top resource consuming queries quickly and also troubleshooting plan choice regressed scenario which is difficult to isolate or time consuming process. Now with powerful report and using few clicks you can check top resource consuming query and force optimum plan. In this article I would like to highlight some of best practices if you are enabling Query Store on mission critical and high transactions Server So as you already know that Query Store have following capture Mode for data collection: All – Captures all queries. This is the default option. Auto – Infrequent queries and queries with insignificant compile and execution duration are ignored. Thresholds for execution count, compile and runtime duration are internally determined. None – Query Store stops capturing new queries.   QDS store runtime stats in cache until they are persisted on disk asynchronously, you could track memory usage using memory clerk USERSTORE_QDSSTMT & CACHESTORE_QDSRUNTIMESTATS. So write operation from cache to disk is very fast and you may not see large memory allocation against these memory clerks however workload with a lot of frequent ad hoc/unparameterized queries would make cache to grow. If are running QDS with capture mode “ALL” on busy Server or mission CRITICAL Server then consider following best practices to save cache memory and reduce database recovery time:
  1. Considering switching capture mode to “Auto” if you are not troubleshooting any ongoing issue for which you needs to capture all queries stats.
  2. If your workload contain lot of ad-hoc batch then considering using “auto” capture mode as capturing detail for ad-hoc workload will not provide much benefit.
  3. Enable Trace Flag 7745 which Forces Query Store to not flush data to disk on database shutdown. Note: Using this trace may cause Query Store data not previously flushed to disk to be lost in case of shutdown. For a SQL Server shutdown, the command SHUTDOWN WITH NOWAIT can be used instead of this trace flag to force an immediate shutdown.
  4. Enable Trace Flag 7752 to Enables asynchronous load of Query Store. Note: Use this trace flag if SQL Server is experiencing high number of QDS_LOADDB waits related to Query Store synchronous load (default behavior).
References Best Practice with the Query Store https://docs.microsoft.com/en-us/sql/relational-databases/performance/best-practice-with-the-query-store Query Store Usage Scenarios https://docs.microsoft.com/en-us/sql/relational-databases/performance/query-store-usage-scenarios https://docs.microsoft.com/en-us/sql/t-sql/database-console-commands/dbcc-traceon-trace-flags-transact-sql   Vikas Rana | Twitter | Linkedin | Support Escalation Engineer Microsoft India GTSC]]> Query Data Store Forced Plan behavior on AlwaysOn Readable Secondary https://docs.microsoft.com/archive/blogs/vikas_rana/query-data-store-forced-plan-behavior-on-alwayson-readable-secondary Fri, 13 Oct 2017 19:10:08 GMT https://blogs.msdn.microsoft.com/vikas_rana/?p=275 Question: For example if you have Query Data Store (QDS) enabled for user database participating in Always On Availability Groups and you Forced Plan for specific query, what happened if same query running on readable secondary will it use Forced Plan? Answer: QDS is not supported on Readable Secondary, so though you have Forced Plan on Primary Replica you “may” see different plan on secondary database because QDS do not force same plan on readable secondary. Please note, SQL do not create Plan Guide to Force Execution Plan in QDS. QDS needs to store runtime stats in databases which is not possible when database in readable state. I am sharing diagram from following article for your reference: https://docs.microsoft.com/en-us/sql/relational-databases/performance/how-query-store-collects-data   Question: Will QDS retain FORCED Plan information when Database failover from Primary replica to secondary Replica? Answer: Yes, QDS store Forced Plan information in sys.query_store_plan table, so in case of failover you will continue to see same behavior on new Primary. select * from sys.query_store_plan where is_forced_plan=1 Question: How QDS Force Query Plan? Is it creating Plan Guide to Force Plan? Answer: As I mentioned earlier QDS do not use plan guide, instead of it use “USE PLAN” query hint which takes an xml_plan as an argument, so QDS basically pull this XML plan information from query_plan column from sys.query_store_plan table. You can refer following diagram to know how SQL SERVER force Plan.       Vikas Rana | Twitter | Linkedin | Support Escalation Engineer Microsoft India GTSC]]> Log Reader Agent Fail on Always On availability Groups configured on Multi-subnet https://docs.microsoft.com/archive/blogs/vikas_rana/log-reader-agent-fail-on-always-on-availability-groups-configured-on-multi-subnet Wed, 09 Aug 2017 16:48:27 GMT https://blogs.msdn.microsoft.com/vikas_rana/?p=187 Consider following scenario:
  • You have Always On availability Groups configured on Multi-subnet.
  • You have AlwaysOn Listener configured and running with RegisterAllProvidersIP 1 to support Modern application which use MultiSubnetFailover in connection string.
  • You have Transactional replication configured with remote distributor.
  • You have redirected Original Publisher to the AG Listener Name using “sp_redirect_publisher” stored procedure
In this scenario Log Reader Agent may fail and log following error message: Error messages: Unable to query the redirected server 'SQLAGLS' for original publisher 'SQLPUB' and publisher database 'PUBDB' to determine the name of the remote server; Error 7303, Error message 'Error 7303, Level 16, State 1, Message: Cannot initialize the data source object of OLE DB provider "SQLNCLI11" for linked server "[xxxx]".'. (Source: MSSQL_REPL, Error number: MSSQL_REPL21879) Get help: https://help/MSSQL_REPL21879 Errors were logged when validating the redirected publisher. (Source: MSSQL_REPL, Error number: MSSQL_REPL22037) Get help: https://help/MSSQL_REPL22037   So in this scenario Log Reader Agent failing because replication use OLE DB provider to connect to Publisher Server which does not support MultiSubnetFailover yet. So when RegisterAllProvidersIP set to 1 for Listener resource in that case it register all dependent IP in DNS. So when OLEDB provider query SQLAGLS then DNS may return IP which is not online hence Log Reader Agent may fail. I would recommend reading following blog to understand more about RegisterAllProvidersIP & HostRecordTTL parameters and it’s impact on connection in multi-subnet scenario: https://blogs.msdn.microsoft.com/alwaysonpro/2014/06/03/connection-timeouts-in-multi-subnet-availability-group/ Solution So in this current scenario if we set RegisterAllProvidersIP to 0 which mean return only active IP in that case Log Reader Agent will work fine however during AG failover it will cause connectivity issue because client operating systems will not resolve the AG Listener name to the new IP address until the currently cached entry expires and modern app will not be able to take advantage of  MultiSubnetFailover. If we reduce HostRecordTTL value to some lower number then this will increase the traffic on DNS. So what is our best option? Considering this scenario instead of using existing AG Listener we can create a dedicated new Listener for replication and set following parameters: Get-ClusterResource SQLAGLSREPL| Set-ClusterParameter HostRecordTTL 120 Get-ClusterResource SQLAGLSREPL |Set-ClusterParameter RegisterAllProvidersIP 0 So RegisterAllProvidersIP 0 will ensure AG Listener return only active IP and HostRecordTTL 120 will restore connectivity quickly (120 seconds or less) in case of AG failover to different subnet. Considering you are using this listener only for Replication hence setting HostRecordTTL to lower number will not cause large amount of traffic to DNS.   Vikas Rana | Twitter | Linkedin | Support Escalation Engineer
Microsoft India GTSC]]> SQL SERVER Transactional and Merge Replication internal and work flow explained https://docs.microsoft.com/archive/blogs/vikas_rana/sql-server-transactional-and-merge-replication-internal-and-work-flow-explained Tue, 30 May 2017 16:30:06 GMT https://blogs.msdn.microsoft.com/vikas_rana/?p=186
  • Different types of Replication in SQL Server
  • Transaction replication work flow and troubleshooting latency – tips
  • Merge replication workflow and troubleshooting agent failures
  • Understanding Merge replication metadata
  • Replication with SQL Server Availability groups
  • Demos for replication issues
    • You can view this video on channel9 using following URL https://channel9.msdn.com/Blogs/Developer-Support-Series-PGI-Sessions/Developer-Support-Series-SQL-Server-Replication   Vikas Rana | Twitter | Linkedin | Support Escalation Engineer Microsoft India GTSC]]>         Troubleshoot 586 error message id with state 5 - The prepared statement handle 1 is not valid in this https://docs.microsoft.com/archive/blogs/vikas_rana/troubleshoot-586-error-message-id-with-state-5 Fri, 31 Mar 2017 13:55:33 GMT https://blogs.msdn.microsoft.com/vikas_rana/?p=185 Recently I worked with one of customer where in jTDS application getting an error 586 and in Profiler Trace we could see an exception with error message id : 586 State 5 with below text: The prepared statement handle 1 is not valid in this context.  Please verify that current database, user default schema, and ANSI_NULLS and QUOTED_IDENTIFIER set options are not changed since the handle is prepared. I was able to reproduce this issue and could see following error message in SSMS: Msg 586,Level 16, State 5, Procedure sp_execute, Line 1 [Batch Start Line2] The prepared statement handle 1 is not valid in this context.  Please verify that current database, user default schema, and ANSI_NULLS and QUOTED_IDENTIFIER set options are not changed since the handle is prepared.  Here is the script I used to reproduce this issue: use tempdb go declare @p1 int exec sp_prepare @p1 output,N'@pp varchar(255)',N'SELECT @pp' select @p1 as PrepHandle exec sp_execute @p1, @pp='This is 1st Execute' use master exec sp_execute @p1, @pp='This is 2nd Execute' go   use tempdb go declare @p1 int exec sp_prepare @p1 output,N'@pp varchar(255)',N'SELECT @pp' select @p1 as PrepHandle exec sp_execute @p1, @pp='This is 1st Execute' SET ANSI_NULLS OFF exec sp_execute @p1, @pp='This is 2nd Execute' go

     

    So whenever you are tracking 586 error message with state 5 please check in Profiler Trace if application changing the database context or changing ANSI_NULLS. Happy Learning!!!

      

    Vikas Rana (@vikasrana_dba)

    Support Escalation Engineer Microsoft India GTSC

     

     

    ]]>         Utility Control Point reporting incorrect CPU utilization when Server has more than 40 logical processor https://docs.microsoft.com/archive/blogs/vikas_rana/utility-control-point-reporting-incorrect-cpu-utilization-when-server-has-more-than-40-logical-processor Wed, 22 Mar 2017 09:46:17 GMT https://blogs.msdn.microsoft.com/vikas_rana/?p=155 num_processors column of sysutility_mi_cpu_stage_internal table contain incorrect value. So server has 80 Logical processor however we could see only 40 being reported, so due to this all calculation was going wrong.   SELECT [num_processors] ,[cpu_name] ,[cpu_caption] ,[cpu_family_id] ,[cpu_architecture_id] ,[server_instance_name] ,[virtual_server_name] ,[physical_server_name] ,[instance_processor_usage_percentage] ,[computer_processor_usage_percentage] FROM [msdb].[dbo].[sysutility_mi_cpu_stage_internal]   We followed below workaround on SQL SERVER 2016 SP1 to resolve this issue: 1. On each enrolled machine go to the sysutility_mi_collect_and_upload SQL Server Agent Job. 2. Edit the Stage Data Collected from PowerShell Script step 3. In the PowerShell script find the line [Int32] $numProcessors = [System.Environment]::ProcessorCount (line 741 in 2016 SP1 copy of the script) 4. Replace this line with [Int32] $numProcessors = (@(gwmi Win32_Processor) | measure -property NumberOfLogicalProcessors -sum).Sum Note: Please take msdb and entire script backup before you modify this code. If you edit the code incorrectly then this can cause UCP job to fail. I have tested this workaround only with SQL SERVER 2016 SP 1 only.   Vikas Rana Twitter| Linkedin | Support Escalation Engineer Microsoft India GTSC]]> Msg 13575 OR Msg 13542 error message while adding period columns https://docs.microsoft.com/archive/blogs/vikas_rana/msg-13575-or-msg-13597-error-message-while-adding-period-columns Tue, 29 Nov 2016 14:46:58 GMT https://blogs.msdn.microsoft.com/vikas_rana/?p=145 1st error message: Msg 13575, Level 16, State 0, Line 32 ADD PERIOD FOR SYSTEM_TIME failed because table 'temporaltest.dbo.CUSTOMERINFO' contains records where end of period is not equal to MAX datetime. 2nd error message: Msg 13542, Level 16, State 0, Line 8733 ADD PERIOD FOR SYSTEM_TIME on table ' temporaltest.dbo.CUSTOMERINFO' failed because there are open records with start of period set to a value in the future. This issue can occur for memory optimized table as well as non-memory optimized table. You can reproduce this issue using following T-SQL: /************************************************************************************************************************/ CREATE DATABASE [temporaltest] CONTAINMENT = NONE ON PRIMARY ( NAME = N'temporaltest', FILENAME = N'E:\temp\temporaltest.mdf' , SIZE = 8192KB , MAXSIZE = UNLIMITED, FILEGROWTH = 65536KB ), FILEGROUP [imoltp_mod] CONTAINS MEMORY_OPTIMIZED_DATA DEFAULT ( NAME = N'imoltp_mod1', FILENAME = N'e:\temp\imoltp_mod1' , MAXSIZE = UNLIMITED) LOG ON ( NAME = N'temporaltest_log', FILENAME = N'E:\temp\temporaltest_log.ldf' , SIZE = 8192KB , MAXSIZE = 2048GB , FILEGROWTH = 65536KB ) GO use temporaltest GO CREATE TABLE [dbo].[CUSTOMERINFO] ( [ID] [bigint] IDENTITY(1,1) NOT NULL, [EMPID] [int] NOT NULL, CONSTRAINT [CUSTOMERINFO_primaryKey] PRIMARY KEY NONCLUSTERED HASH ( [EMPID] )WITH ( BUCKET_COUNT = 128) )WITH ( MEMORY_OPTIMIZED = ON , DURABILITY = SCHEMA_AND_DATA ) GO insert into [CUSTOMERINFO] ([EMPID]) select 1 GO select * from [CUSTOMERINFO] GO ALTER TABLE [CUSTOMERINFO] ADD PERIOD FOR SYSTEM_TIME (SysStartTime, SysEndTime), SysStartTime datetime2 GENERATED ALWAYS AS ROW START HIDDEN NOT NULL DEFAULT GETUTCDATE(), SysEndTime datetime2 GENERATED ALWAYS AS ROW END HIDDEN NOT NULL DEFAULT CONVERT(DATETIME2, '9999-12-31 23:59:59.99999999'); /************************************************************************************************************************/ WORKAROUND If you are seeing same symptoms in that case you can break your T-SQL in 4 parts as shown below which should help you to resolve this issue: /************************************************************************************************************************/ ALTER TABLE [CUSTOMERINFO] ADD SysStartTime datetime2 NOT NULL DEFAULT GETUTCDATE(), SysEndTime datetime2 NOT NULL DEFAULT CONVERT(DATETIME2, '9999-12-31 23:59:59.99999999'); GO ALTER TABLE [CUSTOMERINFO] ADD PERIOD FOR SYSTEM_TIME (SysStartTime, SysEndTime); GO ALTER TABLE [CUSTOMERINFO] alter column SysStartTime ADD HIDDEN; GO ALTER TABLE [CUSTOMERINFO] alter column SysEndTime ADD HIDDEN; /************************************************************************************************************************/

     

    Vikas Rana (@vikasrana_dba) Support Escalation Engineer Microsoft India GTSC ]]>         SQL SERVER 2016 installation failing due to MSVCR120 missing Update (EXCEPTION_ILLEGAL_INSTRUCTION) https://docs.microsoft.com/archive/blogs/vikas_rana/sql-server-2016-installation-failing-due-to-msvcr120-missing-update-exception_illegal_instruction Wed, 23 Nov 2016 16:52:18 GMT https://blogs.msdn.microsoft.com/vikas_rana/?p=85 Detail.txt (01) 2016-10-10 17:25:48 Slp: Sco: Returning service process id 0 (01) 2016-10-10 17:25:48 Slp: Configuration action failed for feature SQL_Engine_Core_Inst during timing ConfigRC and scenario ConfigRC. (01) 2016-10-10 17:25:48 Slp: The service did not respond to the start or control request in a timely fashion. (01) 2016-10-10 17:25:48 Slp: The configuration failure category of current exception is ConfigurationFailure (01) 2016-10-10 17:25:48 Slp: Configuration action failed for feature SQL_Engine_Core_Inst during timing ConfigRC and scenario ConfigRC. (01) 2016-10-10 17:25:48 Slp: System.ComponentModel.Win32Exception (0x80004005): The service did not respond to the start or control request in a timely fashion. (01) 2016-10-10 17:25:48 Slp:    at Microsoft.SqlServer.Configuration.Sco.Service.StartService(String[] startParams) (01) 2016-10-10 17:25:48 Slp:    at Microsoft.SqlServer.Configuration.SqlEngine.SqlServerServiceSCM.StartSqlServer(String[] parameters, Boolean withFailpoint) (01) 2016-10-10 17:25:48 Slp:    at Microsoft.SqlServer.Configuration.SqlEngine.SqlServerStartup.StartSQLServerForInstall(String sqlCollation, String masterFullPath, Boolean isConfiguringTemplateDBs, Boolean isSlipstreamRunning) (01) 2016-10-10 17:25:48 Slp:    at Microsoft.SqlServer.Configuration.SqlEngine.SqlEngineDBStartConfig.ConfigSQLServerSystemDatabases(EffectiveProperties properties, Boolean isConfiguringTemplateDBs, Boolean useInstallInputs) (01) 2016-10-10 17:25:48 Slp:    at Microsoft.SqlServer.Configuration.SqlEngine.SqlEngineDBStartConfig.DoCommonDBStartConfig(ConfigActionTiming timing) (01) 2016-10-10 17:25:48 Slp:    at Microsoft.SqlServer.Configuration.SqlEngine.SqlEngineDBStartConfig.Install(ConfigActionTiming timing, Dictionary`2 actionData, PublicConfigurationBase spcb) (01) 2016-10-10 17:25:48 Slp:    at Microsoft.SqlServer.Configuration.SqlConfigBase.PrivateConfigurationBase.Execute(ConfigActionScenario scenario, ConfigActionTiming timing, ConfigBaseAction action, Dictionary`2 actionData, PublicConfigurationBase spcbCurrent) (01) 2016-10-10 17:25:48 Slp:    at Microsoft.SqlServer.Configuration.SqlConfigBase.SqlFeatureConfigBase.Execute(ConfigActionScenario scenario, ConfigActionTiming timing, ConfigBaseAction action, Dictionary`2 actionData, PublicConfigurationBase spcbCurrent) (01) 2016-10-10 17:25:48 Slp:    at Microsoft.SqlServer.Configuration.SqlConfigBase.SlpConfigAction.ExecuteAction(String actionId) (01) 2016-10-10 17:25:48 Slp:    at Microsoft.SqlServer.Configuration.SqlConfigBase.SlpConfigAction.Execute(String actionId, TextWriter errorStream) (01) 2016-10-10 17:25:48 Slp: The following is an exception stack listing the exceptions in outermost to innermost order (01) 2016-10-10 17:25:48 Slp: Inner exceptions are being indented (01) 2016-10-10 17:25:48 Slp: (01) 2016-10-10 17:25:48 Slp: Exception type: System.ComponentModel.Win32Exception (01) 2016-10-10 17:25:48 Slp:     Message: (01) 2016-10-10 17:25:48 Slp:         The service did not respond to the start or control request in a timely fashion. (01) 2016-10-10 17:25:48 Slp: (01) 2016-10-10 17:25:48 Slp:     HResult : 0x80004005 (01) 2016-10-10 17:25:48 Slp:     Error : 1053 (01) 2016-10-10 17:25:48 Slp:     Data: (01) 2016-10-10 17:25:48 Slp:       SQL.Setup.FailureCategory = ConfigurationFailure (01) 2016-10-10 17:25:48 Slp:       WatsonConfigActionData = INSTALL@CONFIGRC@SQL_ENGINE_CORE_INST (01) 2016-10-10 17:25:48 Slp:       WatsonExceptionFeatureIdsActionData = System.String[] (01) 2016-10-10 17:25:48 Slp:     Stack: (01) 2016-10-10 17:25:48 Slp:         at Microsoft.SqlServer.Configuration.Sco.Service.StartService(String[] startParams) (01) 2016-10-10 17:25:48 Slp:         at Microsoft.SqlServer.Configuration.SqlEngine.SqlServerServiceSCM.StartSqlServer(String[] parameters, Boolean withFailpoint) (01) 2016-10-10 17:25:48 Slp:         at Microsoft.SqlServer.Configuration.SqlEngine.SqlServerStartup.StartSQLServerForInstall(String sqlCollation, String masterFullPath, Boolean isConfiguringTemplateDBs, Boolean isSlipstreamRunning) (01) 2016-10-10 17:25:48 Slp:         at Microsoft.SqlServer.Configuration.SqlEngine.SqlEngineDBStartConfig.ConfigSQLServerSystemDatabases(EffectiveProperties properties, Boolean isConfiguringTemplateDBs, Boolean useInstallInputs) (01) 2016-10-10 17:25:48 Slp:         at Microsoft.SqlServer.Configuration.SqlEngine.SqlEngineDBStartConfig.DoCommonDBStartConfig(ConfigActionTiming timing) (01) 2016-10-10 17:25:48 Slp:         at Microsoft.SqlServer.Configuration.SqlEngine.SqlEngineDBStartConfig.Install(ConfigActionTiming timing, Dictionary`2 actionData, PublicConfigurationBase spcb) (01) 2016-10-10 17:25:48 Slp:         at Microsoft.SqlServer.Configuration.SqlConfigBase.PrivateConfigurationBase.Execute(ConfigActionScenario scenario, ConfigActionTiming timing, ConfigBaseAction action, Dictionary`2 actionData, PublicConfigurationBase spcbCurrent) (01) 2016-10-10 17:25:48 Slp:         at Microsoft.SqlServer.Configuration.SqlConfigBase.SqlFeatureConfigBase.Execute(ConfigActionScenario scenario, ConfigActionTiming timing, ConfigBaseAction action, Dictionary`2 actionData, PublicConfigurationBase spcbCurrent) (01) 2016-10-10 17:25:48 Slp:         at Microsoft.SqlServer.Configuration.SqlConfigBase.SlpConfigAction.ExecuteAction(String actionId) (01) 2016-10-10 17:25:48 Slp:         at Microsoft.SqlServer.Configuration.SqlConfigBase.SlpConfigAction.Execute(String actionId, TextWriter errorStream) (01) 2016-10-10 17:25:48 Slp: Watson Bucket 1 Original Parameter Values (01) 2016-10-10 17:25:48 Slp: Parameter 0 : SQL Server 2016@RTM@ (01) 2016-10-10 17:25:48 Slp: Parameter 1 : Microsoft.SqlServer.Configuration.Sco.Service.StartService (01) 2016-10-10 17:25:48 Slp: Parameter 2 : Microsoft.SqlServer.Configuration.Sco.Service.StartService (01) 2016-10-10 17:25:48 Slp: Parameter 3 : System.ComponentModel.Win32Exception@-2147467259 (01) 2016-10-10 17:25:48 Slp: Parameter 4 : System.ComponentModel.Win32Exception@-2147467259 (01) 2016-10-10 17:25:48 Slp: Parameter 5 : SqlEngineDBStartConfigAction_install_configrc (01) 2016-10-10 17:25:48 Slp: Parameter 6 : INSTALL@CONFIGRC@SQL_ENGINE_CORE_INST (01) 2016-10-10 17:25:48 Slp: Parameter 7 : 1053 (01) 2016-10-10 17:25:48 Slp: Parameter 8 : 1053 (01) 2016-10-10 17:25:26 SQLEngine: : Checking Engine checkpoint 'StartEngineService' (01) 2016-10-10 17:25:26 SQLEngine: --SqlServerServiceSCM: Starting SQL via SCM (-mSqlSetup -Q -qLatin1_General_CI_AS -T4022 -T4010 -T3659 -T3610 -T8015)... (01) 2016-10-10 17:25:26 Slp: Sco: Attempting to start service MSSQL$SQL2016 (01) 2016-10-10 17:25:26 Slp: Sco: Attempting to open SC Manager (01) 2016-10-10 17:25:26 Slp: Sco: Attempting to open service handle for service MSSQL$SQL2016 (01) 2016-10-10 17:25:44 Slp: Sco: Attempting to close service handle for service MSSQL$SQL2016 (01) 2016-10-10 17:25:44 Slp: Sco: Attempting to close SC Manager (01) 2016-10-10 17:25:44 SQLEngine: --SqlServerServiceSCM: Exception generated: The service did not respond to the start or control request in a timely fashion.   So we can cleary see that SQL SERVER Service failed to start and hitting an exception. We looked into the Watson dump signature and can clearly see a APPCRASH and faulting module was MSVCR120.dll with EXCEPTION_ILLEGAL_INSTRUCTION exception.   Version=1 EventType=APPCRASH EventTime=131194442869667947 ReportType=2 Consent=1 ReportIdentifier=d1637b61-8499-11e6-80b6-aeb7d004036e IntegratorReportIdentifier=d1637b60-8499-11e6-80b6-aeb7d004036e NsAppName=sqlservr.exe Response.type=4 Sig[0].Name=Application Name Sig[0].Value=sqlservr.exe Sig[1].Name=Application Version Sig[1].Value=2015.130.1601.5 Sig[2].Name=Application Timestamp Sig[2].Value=5724ae45 Sig[3].Name=Fault Module Name Sig[3].Value=MSVCR120.dll Sig[4].Name=Fault Module Version Sig[4].Value=12.0.40649.5 Sig[5].Name=Fault Module Timestamp Sig[5].Value=56bc00d3 Sig[6].Name=Exception Code Sig[6].Value=c000001d                 =============> This means EXCEPTION_ILLEGAL_INSTRUCTION Sig[7].Name=Exception Offset Sig[7].Value=00000000000931f3 DynamicSig[1].Name=OS Version DynamicSig[1].Value=6.3.9600.2.0.0.272.7 DynamicSig[2].Name=Locale ID DynamicSig[2].Value=2057 DynamicSig[22].Name=Additional Information 1 DynamicSig[22].Value=9d85 DynamicSig[23].Name=Additional Information 2 DynamicSig[23].Value=9d85f85e0140d2c5e78ed7595acfc4cb DynamicSig[24].Name=Additional Information 3 DynamicSig[24].Value=4062 DynamicSig[25].Name=Additional Information 4 DynamicSig[25].Value=406218417176eedb9b6d8b2c6586febb UI[2]=C:\Program Files\Microsoft SQL Server\MSSQL13.MSSQLSERVER\MSSQL\Binn\sqlservr.exe UI[5]=Check online for a solution (recommended) UI[6]=Check for a solution later (recommended) UI[7]=Close UI[8]=SQL Server Windows NT - 64 Bit stopped working and was closed UI[9]=A problem caused the application to stop working correctly. Windows will notify you if a solution is available. UI[10]=&Close     Solution We found that customer already installed Critical Update for Microsoft VC++ 2013 Runtime but issue was still happening. https://support.microsoft.com/en-us/kb/3164398 We have done further research on this and later asked customer to apply following fix and after that SQL SERVER 2016 installation completed successfully. FIX: Programs that are built in Visual C++ 2013 crash with "Illegal Instruction" exception https://support.microsoft.com/en-in/kb/3174417   I hope this will help.   Vikas Rana (@vikasrana_dba) Support Escalation Engineer Microsoft India GTSC]]> SSMS behaviour if you have Cluster Shared Volume and C drive repeating while choosing location for backup\restore\database attach https://docs.microsoft.com/archive/blogs/vikas_rana/ssms-behavior-if-you-have-cluster-shared-volume-and-c-drive-repeating-while-choosing-location-for-backuprestoredatabase-attach Wed, 23 Nov 2016 15:27:44 GMT https://blogs.msdn.microsoft.com/vikas_rana/?p=36 ssms_gui_location   So this behavior may raise questions like, is it by design behavior? Will this cause any issue? Which drive to choose? So this typically happened if you are using Cluster Shared Volume in your environment and have more than one CSV located under C drive. In such scenario you may see this behavior while doing any of above database activity using SSMS GUI, so you can choose any drive and this will not cause any issue as all are representing C drive only. Also worth mentioning that when you have multiple CSV on Server then following DMV basically used in background to pull the drives information which basically return two rows for C drives which contain complete path detail and that’s the reason you may see multiple C drives in SSMS GUI if you have CSV. select * from sys.dm_io_cluster_valid_path_names    
    path_name cluster_owner_node is_cluster_shared_volume
    E:\ 45536SQLNODE1 0
    C:\ClusterStorage\Volume1\ 45536SQLNODE1 1
    C:\ClusterStorage\Volume2\ 45536SQLNODE1 1
      Just in case if you are using CSV and getting following error message while using SQL SERVER 2014 SSMS and unable to access the path on Server: Cannot access the specified path or file on the server. Verify that you have the necessary security privileges and that the path or file exists. If you know that the service account can access a specific file, type in the full path for the file in the File Name control in the Locate dialog box.   Fix for this issue already released for SQL SERVER 2014: FIX: "Cannot access the specified path or file on the server" when you use CSV in SSMS to back up or restore database https://support.microsoft.com/en-in/kb/3048627   I hope this will help.   Vikas Rana (@vikasrana_dba) Support Escalation Engineer Microsoft India GTSC]]>