App_cluster does not follow active cluster node

[GulzarShaikhAUS]

Hi All,

Its Netbackup 7.6.0.3 on master and media servers.

We have a SQL instance installed on windows cluster and we have configured app_cluster for the backup. Both the nodes of cluster are configured as SAN media server.

Recently the SQL DB instanced failed over to another node. As per the app_cluster function it should use the active node for the backup. Its still using the passive node for the backup.

I have tried to delete and recreate the app_cluser but still same issue.

Restarted Netbackup services on master server.

Even if I stop device manager on inactive node, it still goes to the same node as media server and fails….

Any pointer on this?

Can someone explain how Netbackup determines which media server to use when the storage unit is configured with the app_cluster as media server. what processes to look for for the issue.... which processes determines the media server to use?

Regards,

[RamNagalla]

when the cluster fail over to other node , does the IP referring the app_cluster is also fail over to the other node or its still sitting on the old node...?

you can confirm this by checking the ipconfig command... or try loggin in to the server using the app_cluster name/IP and see the hostname of the server that you are in.

when the cluster got failover the IP associated with the app_cluster should also failover... i have seen in many cases it does not happen and backup stil points to passive node.

[GulzarShaikhAUS]

The IP has also failed over. Thats the first thing I checked..

[GulzarShaikhAUS]

Is there any one who would like to help me on this?

[Marianne]

Is the hostname used for app_cluster identical to hostname and IP used for SQL instance/cluster group?

We have seen users associating the app_cluster with the MS Cluster instead of the application (SQL or File server, etc) Cluster Group.

If all of this checks out, you may be a 'victim' of NBU host cache.
But then the issue should not persist for longer than one hour.

[GulzarShaikhAUS · Accepted solution]

Dear Marianne

It has been working for quite some time now. We came to know recently when the resources failed over. This is the first time after we upgraded to version 7.6.0.3

I opened case with support but the response was sluggish as usual.

I finally did following and it resolved the issue - 

1. Deleted and recreated the app_cluster with all possible information I can put in the nbemm command.

2. restarted NBU services on both nodes and Netbackup master server.

3. Found that there is no pointer available in DNS associated with cluster instance which we are trying to backup. (nslookup IP fails)

4. Created hosts file entry for instance name.

5. Changed the option -reverse_name_lookup in bpclntcmd to RESTRICTED on active node.(We later created the pointer and I reverted the setting back to defult ALLOWED)

6. Reatarted the services on media servers and the node was populated in nbemmcmd -listhosts -verbose as Activenodename for the instance.