Hi all

I have a 2 node NW 65sp6 cluster with no shared storage. This cluster is
running a DHCP resource and is having a lot of issues since DHCP went live.
Here is the scenario I'm seeing:

1. Server2 will say it has lost communication with Server2
2. Server1 will think everything is fine - no communication errors showing.
3. The DHCP resource for some reason will migrate from Server2 to Server1.
4. A few hours later Server2 thinks the resource loaded on it again even
though it didn't . Server2 then has communication errors to all other
5. Server1 has now has communication errors to Server2
6. Server2 now thinks it is up and running the resource and that Server1 has
6. Actually - Server1 is fine and running the resource and Server2 has lost
all communication including the ability to connect via freecon.

Here's a little more info also -

I've not really done much on a cluster with no shared storage so have some

A cluster with no shared storage has no SBD partition correct? - so
wouldn't the inconsistent cluster views be caused by not having an SBD?

I'm really not sure where to look on the communication problems. It's not a
network problem. Both servers network connectivity is fine - it's only one
server that loses it's communication and it could be either one. The other
server will still be communicating with all servers except for it's cluster

We are getting ready to migrate these servers to new hardware and I think
that a big part of this issue relates to the lack of the SBD partition. I
would appreciate any input anyone can give since I've always run clusters
with shared storage in a production environment.