DBSPI: Finding the HARG precisely (191 Views)
Reply
Honored Contributor
Sujith Kumar C.S
Posts: 868
Registered: ‎04-07-2010
Message 1 of 2 (191 Views)

DBSPI: Finding the HARG precisely

Hello,

 

I need help in finding the HARG for the cluster on SQL DB servers. I am creating the virtual nodes as the nodes are clustered with single instance. I am using the command ovclusterinfo -a/ovcluster.exe resource to find the HARG ( as this is required to create virtual node for servers in cluster configuration). I was told by the HP engineer to consider the line #Group as HARG but this does not seem to work properly because whenever I deploy DB policies (measurement threshold ones) on the virtual node, its disabling these policies even on the active node (where instance is active). But actually these policies should be enabled on the active node.

 

So I am really confused here - how to find the correct HARG in a SQL cluster. If there were multiple instances I would have used apminfo.xml concept. Any advice/tip to create virtual node. DBA is also unaware of this.

 

Here is the sample output of ovclusterinfo -a

 

C:\Users\abcd>ovclusterinfo -a
#Cluster
type         Microsoft Cluster Server (MSCS)
name         abcd
status       Up
nodes        server1 server2
groups       Cluster Group Available Storage SQL Server (XXXXXXXXXXX)
#Node server1
state        Online
id           1
#Node server2
state        Online
id           2
#Group Cluster Group
state        Online
nodes        server1 server2
local state  Online
virtual IP   1.1.1.1
active node  server1
#Group Available Storage
state        Online
nodes
local state  Online
virtual IP   None
active node
#Group SQL Server (XXXXXXXXX)
state        Online
nodes        server1 server2
local state  Online
virtual IP   2.2.2.2

 

Regards,

Sujith

 

Close the thread if your issue is resolved to make the thread meaningful and complete. Do show gratitude to people who helped you by assigning points.
Valued Contributor
NioByte
Posts: 67
Registered: ‎05-16-2011
Message 2 of 2 (154 Views)

Re: DBSPI: Finding the HARG precisely

[ Edited ]

Hi Sujith,

 

I had this exact same issue with any kind of HARG (High Availaibility Resource Group) Monitoring using the HP Operations Agent.

 

What you will need to to do is create all of the Virtual Node(s) on the HP Operations Manager Management Server which are based on the HARG names.

 

Create a Virual Node for each High Availability Resource Group [HARG] in OM e.g.

 

Name: HARG1

FQDN: <the HARG1 FQDN Domain Name>

IP: <the HARG1 IP Address>

Virtual: <manually assign each managed node that the HARG can be assigned to [in your case server1 and server2] - check the Microsoft Cluster settings for this>.

 

Key to the CLAW [CLuster AWareness] monitoring is: Make sure you have assigned the appropriate managed node[s] to the virtual node[s], have you done this?

 

I have done experiments surrounding the above with multiple agent versions (each one did experience slightly different behaviour, although once assigning the correct managed node[s] to virtual node[s] it seems to work [OA 8.53 or higher]).

 

Once you have performed this then assign the DBSPI policies to Virtual Node(s) [strangely in the older versions of the operations agent I had to assign it to both the physical and virtual node(s) for this to work. This should be ok if using HP Operations Agent 11.0x or above].

 

Then deploy accordingly. Did you assign the policies to the managed node or the virtual node?

 

You will then find the polcies will now be enabled on the HARG node that is currently active with the appropriate resource.

 

Performing an ovpolicy -list on the node should show the appropriate policies enabled [ovcluster -a should show the current active node - log onto that active node and check].

 

Note: Confirm through Microsoft Cluster Failover Manager (assuming you are using Windows Server 2008 or higher) that the node you're checking is currently 'active' with the appropriate cluster resource [ovclusterinfo -a should match with this].

 


Based on your output [server 1 is currently the active node for the HARG group]:

 

 

C:\Users\abcd>ovclusterinfo -a
#Cluster
type         Microsoft Cluster Server (MSCS)
name         abcd
status       Up
nodes        server1 server2
groups       Cluster Group Available Storage SQL Server (XXXXXXXXXXX)
#Node server1
state        Online
id           1
#Node server2
state        Online
id           2
#Group Cluster Group
state        Online
nodes        server1 server2 [This is a HARG Group - as it contains 2 or more nodes]
local state  Online
virtual IP   1.1.1.1
active node  server1


#Group Available Storage
state        Online
nodes
local state  Online
virtual IP   None [this one is not active as it is local and does not have an IP]
active node [this one is usually NULL]

 

Note you have missed the important 'active node' line from the end of the output below:


#Group SQL Server (XXXXXXXXX)
state        Online
nodes        server1 server2 [This is also a HARG group - as it contains 2 or more nodes]
local state  Online
virtual IP   2.2.2.2

The opinions expressed above are the personal opinions of the authors, not of HP. By using this site, you accept the Terms of Use and Rules of Participation.