Physical Disk goes offline when cluster node reboots

I have a 2 node Windows 2003 SP1 EE cluster connected to an MSA1000 SAN
via integrated FC hub. My SAN is single-path since this is our Dev/QA
environment.

When I reboot any node in the cluster all physical disk resources go
offline while the rebooted server goes through POST. I get Delayed
Write Failed errors in the event log of the node that is still running.
Once the rebooted node is up and running the cluster returns to
normal.

I'm worried that our production cluster may exhibit the same issues
when it goes live even though it is built in a more robust fashion.

I'm open for suggestions.

The servers are HP DL145's, using Emulex FC2243 cards. If I simply
failover a cluster group everything works great.

Thanks.
-DK

.

Relevant Pages

  • Re: Physical Disk goes offline when cluster node reboots
    ... will also require that you apply a STORport hotfix, ... When I reboot any node in the cluster all physical disk resources go ... offline while the rebooted server goes through POST. ...
    (microsoft.public.windows.server.clustering)
  • Re: Problem Windows 2003 Cluster
    ... I think that the latency during reboot is not linked to a service that the ... to reboot.(now I upgrade this cluster to Biproc, more than 1 GB of RAM, after ... When I talk about the reboot time I mean how many time windows 2003 take to ... > resources and when you fail a resource group to another node, ...
    (microsoft.public.windows.server.clustering)
  • Re: Exchange Services not running in cluster?
    ... you shouldn't be starting any of the services from the services control panel, you should have cluster resources for all of the services and start them from Cluster services manager. ... reboot node2. ... Our A/C failed yesterday in our server room and it got ... After things cooled down I brought excluster1 back up and move ...
    (microsoft.public.exchange.admin)
  • Can a satellite survive the reboot of its boot server?
    ... Imagine a 3-node cluster with each node having 1 vote. ... 0-vote satellite into the cluster from one of the nodes. ... to this satellite if its boot server reboots? ... I also like to avoid a cluster reboot. ...
    (comp.os.vms)
  • Re: Reboot? How often?
    ... That is what I am getting at, if you patch monthly then you will reboot ... Rod ... > compelling reason given the stability of Windows server these days. ... >>> to the inactive side of the cluster then rebooting the passive node. ...
    (microsoft.public.sqlserver.clustering)
Loading