{"id":1558,"date":"2018-02-24T14:01:41","date_gmt":"2018-02-24T14:01:41","guid":{"rendered":"http:\/\/www.theSQLReport.com\/?p=1558"},"modified":"2018-06-23T14:32:41","modified_gmt":"2018-06-23T14:32:41","slug":"problem-patching-a-sql-server-cluster","status":"publish","type":"post","link":"http:\/\/www.theSQLReport.com\/?p=1558","title":{"rendered":"Problem Patching a SQL Server Cluster"},"content":{"rendered":"

Problem:<\/strong><\/h3>\n

I was working with a SQL Server 2014 two node cluster in our test environment.\u00a0 After the cluster had been setup, I was asked to change the drive letters of the disks used by the database role in the cluster.\u00a0 I was unaware the issues that this would cause patching.<\/p>\n

Following the standard patching procedure, move all resources to the second node while patching the first node.\u00a0 After successfully patching the first node, I attempted to move the resources from the second node back to the first node, and failed with the following warnings and errors in the FailoverClustering log in the Event Viewer:<\/p>\n

Warning –
\n[RES] SQL Server <SQL Server (…)>: [sqsrvres] Worker Thread (…): ReAclDirectory : Failed to apply security to …\\MSSQL\\Data (…).<\/h6>\n
Warning –
\n[RES] SQL Server <SQL Server (…)>: [sqsrvres] Worker Thread (…): ReAclDirectory : Failed to apply security to …\\MSSQL\\Log (…)<\/h6>\n
Warning –
\n[RES] SQL Server <SQL Server (…)>: [sqsrvres] Worker Thread (…): ReAclDirectory : Failed to apply security to …\\MSSQL\\Backup (…)<\/h6>\n
Warning –
\n[RES] SQL Server <SQL Server (…)>: [sqsrvres] Worker Thread (…): DoREPLSharedDataUpgrade : Failed to create working directory.<\/h6>\n
Error –
\n[RES] SQL Server <SQL Server (…)>: [sqsrvres] SQL Cluster shared data upgrade failed with error 0 (worker retval = 3). Please contact customer support<\/h6>\n

Resolution:<\/strong><\/h3>\n

First Thank You to Pinal Dave for publishing:<\/p>\n

SQL SERVER \u2013 Unable to bring resource online \u2013 Error DoREPLSharedDataUpgrade : Failed to create working directory<\/a><\/h5>\n

Because SQL Server would not start up on the first node of the cluster, I had to patch the second node with SQL Server shutdown, which was a total outage of SQL Server.<\/p>\n

    \n
  1. On the second node of the cluster, I shutdown SQL Server and the SQL Agent, but brought all the disks for that role back to an on-line status.<\/li>\n
  2. Following Pinal Dave’s post listed above, I searched and changed every registry entries found with the old drive letters with in the folder:\n