Transaction log backup failure on clustered MS SQL

[JClementi]

Hi I've run into a rather particular error that I've been able to replicate but have no solution for.
We have one SQL Cluster where the Instance name is MSSQLServer. It appears the way Veeam does the application aware backups is by enumerating the instance names running on a host and using the (local) variable to target the instances. For the other explicitly named instances this seems to work fine with the sqloledb driver using this method, for example we have one named SQL-Util-Cstr\UtilDB and succeeds using (local)\UTILDB. On the other hand our production data uses SQL-Prod-Cstr and tries to target it by using first (local) and later the Virtual Machine name SQVM1-VSRV both of which fail. I did write a small .NET app to test SQLOLEDB connections strings and both of those did return the same 0x80004005 error while using the Clustered Name SQL-Prod-Cstr succeeded. Is there any hotfix available to have it try to use the proper Cluster Names to connect to an instance when the instance is Clustered?

[rreed]

We've been getting the same w/ an existing local (non-cluster) SQL server here for a while. Was about to open a ticket but figured I'd give a quick search here real quick.

Code: Select all

5/22/2016 7:14:32 PM :: Unable to truncate Microsoft SQL Server transaction logs. Details:  Error code: 0x80004005
Failed to invoke func [TruncateSqlLogs]: Unspecified error. Failed to process 'TruncateSQLLog' command.
Failed to truncate SQL server transaction logs for instances:  . See guest helper log. . 
 Error code: 0x80004005
Failed to invoke func [TruncateSqlLogs]: Unspecified error. Failed to process 'TruncateSQLLog' command.

Failed to truncate SQL server transaction logs for instances: . See guest helper log. .

[rreed]

At face value, this appears to apply, but we're using the latest v9. https://www.veeam.com/kb1746 I'll step through this workaround and see if that fixes ours.

[JClementi]

Ya in my circumstance I did verify that it was not a permissions issue, any comment from the Veeam staff on a potential fix or workaround ?

[Vitaliy S.]

Hi Joseph, I have just consulted with our SQL Server expert, and it's hard to say what's going on without seeing full debug log files. Do you have a support ticket opened with our team already?

[foggy]

@rreed, your issue seems to be different, please review the following threads, could be useful:
Truncating SQL transaction logs - SUDDENLY trouble
Unable to truncate transaction logs.

[rreed]

Thanks Foggy. I don't recall when the issue for this particular server popped up, it could have always been doing it. Upon inspection, this SQL server has been running built-in SQL backups setup by the third-party app it handles since it was built long ago so we simply disabled application processing for it and went on. Veeam gets the file system/OS which is storing its SQL backups which we discerned is good enough for us in this particular case.

[rreed]

If it helps, the third-party app that installed/configured SQL does have it locked down tightly - not even domain admins have access to SQL on it so that is likely why were getting the NT authentication/access denied when Veeam was trying to process the log files.

[JClementi]

Hi Joseph, I have just consulted with our SQL Server expert, and it's hard to say what's going on without seeing full debug log files. Do you have a support ticket opened with our team already?

Thanks for the reply, I had been on vacation. I have opened a case with the Support Team #01816413

[JClementi]

Just an update, after working with Tier 3 support we were able to come up with a workaround which was to add a SQL Server alias for the clustered instance with the fully qualified Host name it was running under, this allowed the fallback attempt from (local) to the host to establish the connection. Figured I would post the resolution incase anyone else ran across it.

Example Solution:

Clustered Instance Name: Company-SQL-CSTR

Alias Configuration where Host.Company.local is the Node the cluster is running on:

[UweKnoepfle]

Did this realy resolved your problem?
We have an case open with the support team and say always want test the connection to the local node and not the cluster node over port 1433.
Where did you made the alias configuration?