Re: How would you configure an N+1 cluster using MS Cluster Servic

Richard,

You could try to create a script that all your application resources depend
upon that tests whether or not any other apps are present on the node. If
found, you could leave the group in an offline state. Exchange has something
like this builtin for 3+ node clusters.

Otherwise, you cannot really prevent multiple nodes from coming online on
the same node and still maintain automatic failover. You can establish
"preferred owners" and "group AntiAffinity" to help keep groups seperate,
but in a worst case scenerio, the groups would still be allowed to come
online on the same node.

You can read more about the AntiAffinityClassNames in the following KB:

How to configure Windows clustering groups for hot spare support
http://support.microsoft.com/kb/296799

Regards,
John


"Richard Glanville" <fredfredfred@xxxxxxxxxxxxx> wrote in message
news:7EE1F88B-B02E-4CBD-9DAB-DC925C979625@xxxxxxxxxxxxxxxx
Thank you for the reply.

Please correct me if I'm wrong, but I can't see how I can allow two active
nodes to fail over to the single passive node in my situation.

The application consists of a number of services; the services are being
added as "Generic Services" in Cluster Administrator. These services get
configuration information from a key under HKLM, so that key is setup to
replicate in the generic service properties. When the first node fails
over
to the passive node then the registry would replicate OK, but when the
second
node fails over to the passive node its registry configuration data would
overwrite and interfere with the data for the first node that failed over.
This is why I am looking to prevent more than one node from failing over
to
the passive node.

With regard to option 2, I could have one passive node per active node to
eliminate the possibility of multiple failover but it was the cost of
these
additional passive nodes that we want to save by accepting that only one
node
failure could be handled at any one time.

Thanks,
Richard.


"Ryan Sokolowski (MVP)" wrote:

You can define your Failover and Failback preferences, but I'd also
suggest
that if you're concerned about 2 active nodes failing over
simultaneously,
do one of two things.

1. Scope out the cluster configuration resource-wise so that 2 active
nodes
could legimately fail their resources to your single passive node, or,
2. Build the cluster with multiple passive nodes to accomodate this
potentiality (i.e. - A/A/A/A/P/P)

Hope this helps,
--Ryan

--
--
Ryan Sokolowski
MVP - Windows Server - Clustering
MCSE, CCNA, CCDA, BCFP

"A troubleshooter's best tool is the Event Viewer and understanding the
events and messages contained therein."

This posting is provided "AS IS" with no warranties, and confers no
rights.

"Richard Glanville" <fredfredfred@xxxxxxxxxxxxx> wrote in message
news:AF14741A-AA94-4422-A733-C7AEAB8C7B90@xxxxxxxxxxxxxxxx
I have an application that I want to cluster. It consists of a number
of
services that need to be dependent on MSMQ, Physical Disk, Network
Name &
IP
Address resources.

I have successfully tested an Active/Passive cluster configuration
using a
single group. What I need to be able to do is get a configuration
where
there
are N active nodes each with their own Network Name etc. and 1 passive
node
to which any of the active nodes could failover. If a second active
node
needs to failover then that should be prevented.

Can this be done?

Thanks,
Richard

Can this be done?

Thanks,
Richard






.

Relevant Pages

  • Re: How would you configure an N+1 cluster using MS Cluster Servic
    ... hence prevent multiple failovers to a passive node in an N+1 configuration. ... You could try to create a script that all your application resources depend ... you cannot really prevent multiple nodes from coming online on ... Scope out the cluster configuration resource-wise so that 2 active ...
    (microsoft.public.windows.server.clustering)
  • Re: Changing Node & Virtual IPs for Print Server Cluster
    ... It's a clustered print server. ... Cluster Node 1 Name: cluster1a.domain.com ... Move all resources to Node A. ... IP addresses for all virtual servers including the cluster itself. ...
    (microsoft.public.windows.server.clustering)
  • Re: Exch 2003 SP2 - applied on one node, but cant move resources
    ... resources to Node2, the failover did not complete because 'system attendant' ... Virtual Exchange server and failover occurred normally again upon taking ... cluster resources oline. ...
    (microsoft.public.exchange.admin)
  • Re: Failed cluster node confusion!
    ... Blue exclamation marks usually means that the cluster service has terminated ... If this fails then the heartbeat will go over the teamed NIC ... the second node did NOT failover the resources. ... working node when one node has completely died (blue screen, ...
    (microsoft.public.windows.server.clustering)
  • RE: Cluster migrations
    ... Our file clusters have one virtual name per cluster group, ... individual file share resources within that group. ... Then if you need to move it you just present the LUN ... run mountvol /e at the command on the Win2k3 node before you present the LUN ...
    (microsoft.public.windows.server.clustering)