Cluster Fails when lose network...

Hello. We have a SQL Server 2000 Cluster on Windows Server 2003 that is
running great. We had a network outage last week, however, and the
whole cluster went down. We noticed that the Cluster Group "IP Address"
resource went down and that then took the SQL Server services offline.
We reproduced this behavior this morning by unplugging the WAN network
card [while leaving the "heartbeat" connection] and the Cluster Group
failed and took the SQL Cluster group offline as a result. This is a
major flaw. What needs to be configured to work around this? A
stand-alone SQL Server does not fail when unplugged from the network, it
just keeps on running until connectivity is restored.
This isn't a flaw, it is a feature. The cluster service monitors the
service and its underlying resources. When a resource is unavailable, the
cluster tries to find a node where the resource is available. If you unplug
one host from the WAN, the SQL server group should fail over to the other
node as should any other cluster groups. Read a little more about clusters
and you will see why this is happening.
Geoff N. Hiten
Microsoft SQL Server MVP
Senior Database Administrator
Careerbuilder.com
I support the Professional Association for SQL Server
www.sqlpass.org
"michael [multnomah]" <wewill@.reclaimourfuture.net> wrote in message
news:g7mdnZS4QJZQaQ3cRVn-iA@.speakeasy.net...
> Hello. We have a SQL Server 2000 Cluster on Windows Server 2003 that is
> running great. We had a network outage last week, however, and the
> whole cluster went down. We noticed that the Cluster Group "IP Address"
> resource went down and that then took the SQL Server services offline.
> We reproduced this behavior this morning by unplugging the WAN network
> card [while leaving the "heartbeat" connection] and the Cluster Group
> failed and took the SQL Cluster group offline as a result. This is a
> major flaw. What needs to be configured to work around this? A
> stand-alone SQL Server does not fail when unplugged from the network, it
> just keeps on running until connectivity is restored.
|||This is correct. The cluster service periodically connects to SQL Server to
verify that it is usable. If it cannot connect it will fail the resource
and attempt to bring it back on line.
Rand
This posting is provided "as is" with no warranties and confers no rights.