question about failover clustering service

here is how to setup a cluster. You can check your setip to see if it cose.

http://blogs.msdn.com/clustering/archive/2008/01/18/7151154.aspx

CT

i've already set it up ... question is different here !

What are you using to run your cluster???

Is your cluster configured for failover or processing???

I used the Failover cluster manager for failover ... So its not NLB

Each of the server in the cluster have 1 NIC ... when I unplug the cable from there, thats when I cant ping from the client to the Virtual I.P of the cluster ....

Thats correct because the Virtual IP is still bound through the physical NIC. Communications are still sent and recieved throught the adapter.

I generally would suggest when configuring a Cluster for failover would have 2x NIC's.

1. Standard Network Traffic.
2. Private IP for internal comms between the cluster.

So i guess the answer to your question is that if the system is unplugged from the network and using a Virtual NIC (dependant on the system being on for the routing to occur) then it cannot act as part of the cluster.

hence the reason for the second NIC. as a turned off server can be sent a message from the headnode to turn on to enact the failover.

Ok, am not using the 2nd NIC at the moment ... but even If i was using the second NIC, lets say I switch off the second box altogether, then that second NIC wont do good either because there wont be any communication between the 2 servers in the cluster ? Right ?

So this means clustering only works if one of the services on that server is down; but if the whole server goes down, then the cluster fails ?

Hi,

I have not  clustered windows 2008 yet but in the cluster manager you should have the single NIC configured for both public access and the private heartbeat. If its not make sure that it is. When a node in a failover cluster goes offline due to any type of failure i.e. switch issue, hardware, OS failure then the node that's still online will recognise the primary node has failed and initiate a failover to bring the cluster group and associated services online.

Are you seeing any errors in the even logs when the cluster group is trying to failover? Can you initiate a failover betweeen the nodes via the cluster manager with both servers online? (You should be able to failover the cluster group from one to the other without turning anything off, etc).

Many times this type of thing is due to some dependant service that's not coming online on the secondary node first.

In short you don't have to have a second NIC but I would definately recommend it.

Ok I added the 2nd NIC as well and configured it with Public I.P .... What I've done so far is just added 2 nodes to the cluster and havent configured any service or application ... I just test using the ping command and unplug the cables from the 2nd server (both public NIC and private NIC to simulate the fact that the server has gone down rather than trigerring the usual failover option) .... Do I need to add any service to the cluster first and then test or should this ping to the virtual I.P would be successfull if i take down the second server !

Am a new comer to microsoft, so if anything is unclear, let me know and i shall send u the screenshots ...

typo above in first line .... I configured the 2nd NIC with a prvt. I.P, not public ...

Ok..Cluster gave warnings against storage test ... and in the warnings, it said that disk type is not supported for Clustering ...