RPC errors when attempting to add a node

[BHuckfield]

I have 2 servers loaded with 2003 Enterprise. They are both connected to an Compaq MSA1000 Disk Array. Selective Storage has been correctly configured so that both servers can access the SAN one at a time.
I created a cluster on one of the servers, and joined the second server to the cluster. All was well until I shutdown one of the servers to add 2 SCSI cards for an LTO changer. After coming back up, the server no longer was able to access the cluster. After trying a lot, I eventually evicted the server from the cluster, and subsequently I am unable to rejoin the cluster with this machine.

The servers connectivity is 100% as both can see one another via the public and heartbeat network. When I attempt to add the node to the cluster, it gets to the part where is says "Synchronizing the cluster database with the sponsor cluster" and fails...
In the cluster log it says "The Remote Procedure call failed"
I have checked that RPC is alive and well on both machines and this is the case. Short of trashing the cluster and starting again, I am kind of stuck.

Anyone seen this kinda behaviour before?

thanks in advance.
Brian

 

[xmsre]

Sounds a lot like this. Try manually removing with /force before rejoining.

http://support.microsoft.com/default.aspx?scid=kb;en-us;297881

 

[BHuckfield]

thanks for the response. You are correct, it looks a lot like this problem...however, I have already run the /forcecleanup on the node name that was removed. Also, I did do the eviction process on the server that does not want to join again....and an attempt to join the cluster brings the same results...

 

[BHuckfield]

Just in case anyone is interested, I fixed the problem by recreating the cluster. The strange thing is that I deleted the cluster, restarted the server that was last one in the cluster and then recreated it. After this, I attempted to join the machine that was giving me problems in the first place and got the same problem...

So, I deleted the cluster again, and recreated the cluster on the machine that was giving me uphill. Worked 100%.
The machine that was originally used to create the cluster was then joined to the cluster and all is well..

I have no idea how this is possible, I am still looking for answers, so if anyone has any ideas, please let me know..

thanks
Brian