New Cluster Disk Resource Not Behaving Like Existing Resources 1

[Shaliko]

I am new to supporting MSCS.
Our MSCS consists of 2 x Dell 6350 Servers each with Raid 1 mirrored internal drives plus two Dell PowerVault 201s cabinets each containing a Raid 5 set of disks.
While this works OK, it was deemed necessary to install an additional pair of disks for a particular resource we want to establish on the cluster. The two new drives were installed in one powervault cabinet and configured using Dell’s (HP?) array manager to a raid 1 mirrored set, formatted to NTFS and drive letter allocated. This process was repeated for each of the servers.
We have found that while the new disk resource was easily added to the cluster (in cluster administrator), the offline node still has access to it. The offline node would normally not be able to access the cluster disks. On the offline node, cluster disks would normally be reported as “remote disks” but the new resource we have added is classed as a “local disk”. This is quite frustrating as I am not entirely sure if something is wrong or not but my gut feeling suggests the new disk resource should be seen in the same way as the existing resources from the offline node. Confused? I am!
Look forward to your comments. Tks.

 

[rmarston]

I had a simular problem, when we made the resource come 'online' the server blue screened!!!

However after a reboot of both nodes in the cluster it
worked fine. BUT be warned I found that the other node
in the cluster could not take control of the resource
group because the new resource wouldn't load!

Why I don't know

ROb

 

[rmarston]

Sorry thats misleading, should have said that it all worked OK, after both machines were rebooted. However when one machine was rebooted the other could not successfully pick up its resources {until it too was rebooted}

Rob

 

[Shaliko]

Tks for your input Rob. Much appreciated. Also good to know there are other cluster admins out there!
Interestingly, I attended a 2 day MSCS course with Dell Computers a couple of weeks ago. Quite interesting except not enough time was devoted to what really matters ie troubleshooting cluster.
We have lots of trouble with our Cluster Print server. I suspect its an RPC problem with selected HP Lasers (mainly series 5. We get hundreds of event 2004 errors and eventually the cluster hangs. I should say the Spooler hangs and takes out the other cluster resources 50% of the time. Other people I have talked to about Cluster printing have not given it a chance and moved to straight NT print serving.......I would like this one resolved. I have moved some of the troublesome printers to another server for the time being and will monitor with interest over the next few weeks to see if stability is achieved.
Keep in touch.
My email is Neil.Rogers@comalco.riotinto.com.au
Rgds

Neil

 

[rmarston]

We had a similar printing problem on a stadalong NT4
server {IP printing only}. It was a complete nightmare
and would just stop servicing printers at random.

In the end I think three things solved it...
1) Q243222
2) New printer drivers from HP
3) Static not DHCP Printer IP addresses
{who set them up as DHCP in the 1st place I will never know!}

Rob

 

[MarcS1]

When you added the additional disk's to your clusters did you shut down one of the node first ?. You should only ever have one node up and running when adding new storage.

 

[rmarston]

Due to the extra SCSI i/f we had to reboot both halves
of the cluster independently.

However when the h/w resources where added both halves
of the cluster servers were up and running...

Rob