Problem with MS Cluster Servers

[Guest]

I have a problem with my 2 server active/passive windows 2000 adv. server cluster. When I move the cluster group from the active member to the passive member, the smtp virtual server fails to move. It eventually times out and then becomes active on the previously passive member but, it takes a failure for this to occur (and an extra 7 or 8 minutes).

I've checked the event log and it yields no clues

One other point of interest, when I shutdown the power of the active cluster member (to a simulate an active member server outage), I expect the passive member to take control of the cluster services. This fails also. It too, fails on the smtp virtual server cluster service. I notice when examining the services that are started on the remaining cluster member that the smtp service does not start. Of course this need to start first before the smtp virtual service can

So, my 2 server cluster is not at all fault tolerant and does not fail over like I think it should

Any help or suggestions would be greatly appreciated

Thanks
Matt K.

 

[John Toner]

Matt,

First, you shouldn't expect fault tolerance from a cluster server. Windows
clustering is "high availability" not fault tolerance. Important
distinction, yet you are not seeing high availability either...

Check the cluster.log for clues. This can be found in the
%systemroot%\cluster folder.

How are your groups setup? Sounds like you have two groups...the Cluster
Group and an "SMTP" group...or are all of your resources grouped into the
cluster group?

You state that the smtp virtual server fails to move...which leads me to
believe that you have a this in a group separate from the Cluster group.
Which resource in this group is failing? Can you successfully move the
Cluster Group between nodes?

Regards,
John

I have a problem with my 2 server active/passive windows 2000 adv. server

cluster. When I move the cluster group from the active member to the passive
member, the smtp virtual server fails to move. It eventually times out and
then becomes active on the previously passive member but, it takes a failure
for this to occur (and an extra 7 or 8 minutes).

I've checked the event log and it yields no clues.

One other point of interest, when I shutdown the power of the active

cluster member (to a simulate an active member server outage), I expect the
passive member to take control of the cluster services. This fails also. It
too, fails on the smtp virtual server cluster service. I notice when
examining the services that are started on the remaining cluster member that
the smtp service does not start. Of course this need to start first before
the smtp virtual service can.