Monitoring AlwaysOn HA Groups

jtrumbul

I am currently getting an error when setting up monitor for our clustered environment. The group consists of 2 servers dbA and dbB. DBA is the primary. When I add the group to monitor, I am using clustername.foo.bar\* SQL monitor seems to be reading the registry just fine, the server adds, the nodes add, then when it comes to connecting to the (local) db on the nodes i get the following error:

16 Aug 2012 12:33 PM
File
Open: C:\SQLDATA1\MSSQL11.MSSQLSERVER\MSSQL\Log\ERRORLOG
Cannot connect
IOException
The network name cannot be found.

16 Aug 2012 12:32 PM
File
List Files: C:\SQLDATA1\MSSQL11.MSSQLSERVER\MSSQL\LOG
Cannot connect
IOException
The network name cannot be found

I have run through all the connection tests connecting as the windows account i am using, everything checks out. I have navigated to that share as that user and had no permission issues. I am fresh out of ideas, any help would be greatly appreciated. The only thing I can mention not sure what effect it would or wouldn't have is the drive is mounted as a folder (which obviously cannot change).

Any help is much appreciated,
John

chriskelly

Are you using a SQL Server 2012 High Availablity Group?

jtrumbul

Yes Chris, I sure am. I was just told by support this isn't fully supported yet and a ticket was opened. I guess for now i will just have to deal?

chriskelly

Unfortunately, SQL Monitor does not yet support High Availability groups in SQL Server 2012. Whomever you spoke to is correct. Adding support for a number of new SQL 2012 features is in the backlog and the developers will hopefully look at it soon.

johnmcp

Is it still the case that 2012 HA Groups are not supported in SQL Monitor?

chriskelly

SQL Monitor does now support SQL Server 2012 and can monitor AlwaysOn Availability Groups.

johnmcp

Hi Chris,

Although I can add the Availability group, I get backup overdue messages for the members that are not performing backups. Is this expected behavior for SQL Monitor?

chriskelly

I should think so. The backup overdue alert is based upon the result of a query on the msdb.dbo.backupset table. Where the entry for a particular database is older than the alert threshold, or there isn't an entry then the alert will be raised. To avoid this happening you can disable the alert for any database where you do not want to be informed that backups have not been taken.

johnmcp

I would say that its not ideal to have it this way given that a backup can be taken on any member of an availability group depending on the backup preferences. I do not want to disable the alert on servers that are not performing backups currently because they may change roles at some point and become the preferred backup server.
I would suggest that this alert first check the function sys.fn_hadr_backup_is_preferred_replica to check that the server is specified as a backup server for the replica.