I have installed SnapCenter 4.2.0 agents in two node SQL Server cluster with active-active configuration. But SnapCenter only can access to the instances running in the node which own cluster name resource. Does anybody know how can I work with the instances on the secod cluster node?
Interesting query, Is that not a normal behavior ?
SQL Cluster supports a Failover Clustering only where only one node is active at a certain point of time and next node takes up the work post a failover happens automatically/manually triggered.
Considering that , SnapCenter agent will only see the resources for the Node that is active.
Does SQL Server support 'true' Active/Active Clustering?
Could someone from SnapCreator team confirm this ?
When I say active/active I mean that each node runs different instances of SQL Server. For example:
- Node 1 executes instances A and B
- Node 2 executes instances C and D
SnapCenter can only access instances that run on the node that has the cluster name. If the cluster name is on node 1, SnapCenter only accesses instances A and B. If the cluster name is on node 2, SnapCenter only accesses instances C and D.
Thanks for clearing that for me, that makes sense when you say 'active/active'.
I guess, the question now is : Is it supported with SnapCenter ?
I will try to find this out.
There is a reference of installing plug-in on SQL cluster.
Adding hosts and installing SnapCenter Plug-in for Microsoft Windows:
Installing plug-ins on a cluster: (I don't know if AlwaysOn means Active/Active)
If you install plug-ins on a cluster (WSFC, Oracle RAC, SQL Always OnAvailability Groups, or Exchange DAG), they are installed on all of the nodes of the cluster.
Could you ensure this:
SnapCenter depends on the proper configuration of the DNS. Therefore, the best practice is to enter the fully qualified domain name (FQDN). If you are adding a host using SnapCenter and it is part of a subdomain, you must provide the FQDN.
It is possible that the problem lies in the resolution names. We have a disjoint Windows namespace: the domain FQDN is "usc.es" but the servers FQDN is "server1.rpso.usc.es" being "rpso.usc.es", not a Windows subdomain , but a DNS zone.
SnapCenter doesn't understand that and insists on always resolving "server1.usc.es" so we have to fool it with a "hosts" file establishiong alias like "server1.usc.es".
But after installing the agents we find this problem.
Thanks for the update. Yes, it could well be.
I think looking at the version of the SnapCenter, you should be able to raise a ticket with NetApp. Have you already raised one regarding this issue ? If not, I suggest please do so, chances are your issue will get resolved sooner.
Also, I come across this KB, talking about DNS issue. I guess you have acces to KBs. Please take a look.
SnapCenter Resource "Availability Group":
This problem has nothing to do with the disjoint namespace. In order to work with all instances the nodes have to be able to access the SQL Browser service, which we had disabled. By enabling it all instances are accessible by Snapcenter.