This article explains the process to manually edit a CA ARCserve RHA SQL scenario .XMC file to manually remove a directory as a work around when the following problem occurs.
In SQL 2012 and higher cluster instances, Microsoft allows for the temporary database directory location for the instance to be located on a non-shared physical disk resource as long as the volume exists on each node in the cluster. This can turn into an issue when trying to create a CA ARCserve RHA SQL scenario that uses our Auto-discovery logic to build the scenario root directories.
During auto-discovery of a SQL instance the RHA engine runs a SQL query on dbo to find where the location of all database files are and all of the configuration information for the system and user databases. Once this information is gathered the RHA manager scenario creation wizard or auto-discovery of Database FIles wizard from the root directories tab will display them.
A HA scenario type by default includes all System databases and creates an exclusion for the tempdb database files within their root directories. There is also a check box for 'Replicate new database files in the selected root directories' at the bottom of these wizards. When this is selected the auto-discovery wizard adds all root directories where databases (system which is auto checked for HA and checked user) exist to the scenario and add an 'Include' filter for '*.mdf *.ldf *.ndf *.dat' as they are the standard SQL server database file type extensions.
When the check box is checked and the scenario type is HA this means that the configuration in this article for Clustered SQL Server instance with the tempdb database on a local volume, will have the folder where the tempdb database is located added as a root directory. This configuration is not supported because of how the RHA file system driver works when replicating data from a Microsoft Cluster.
To address the above issue this article has the below procedure on how to manually edit the .XMC file to be able to start the scenario and protect the above configuration.
*** Please note that by unchecking the 'Replicate new databases in the selected root directories' box in the wizard RHA will automatically exclude the local volume in the root directories, but will not replicate any new databases created in those folders after the scenario is started ***
- Open the RHA Manager and begin by creating the SQL Scenario.
- During the Master root directory portion and select the user databases that will be protected as defined in the CA ARCserve RHA SQL Server Operations guide. Make sure to leave the check box for 'Replicate new databases ...' checked (unless you don't want new databases to be replicated)
- Complete the scenario creation wizard and at the end click 'Finish' to save it.
- Next right click on the scenario name and then click on 'Export'. Save the file to a location that you can easily find the file.
- Next open the file in notepad or a file editor such as textpad. You can use Wordpad but make sure you turn off the word wrap option.
- Perform a search for the root directory to be removed and then remove the entire 'Object' to '/Object' XML node text which will include the directory and also the exclusion configuration.
- Perform another search and remove the same 'Object' to '/Object' XML node information for the Replica as well.
- Save the file as a new name to be distinguished from the original.
- Then in the RHA Manager select the 'Group' you want to import the new scenario into and right-click then click on 'Import'.
- Select the newly named scenario file for import and choose to 'Save scenario with a different one', so as not to overwrite the previous scenario for comparison.
- Once the scenario is imported you should see that the root directory is no longer selected and the scenario can be started normally.