The SQL cluster is not failover over to another node. Can anyone shed some
light in to this. I have checked the config on the cluster administrator side
and it all seems correct to me. The event id appeared is 1153 event source:
clussvc.
In theory the cluster is not working if a disaster happens. A manual reboot
is required to bring everything up to normal. Please advise.

Are you able to move any group from node to node?

If so, then which group will not move?

Also, please advise which versions of SQL and Windows you are using.

--
Geoff N. Hiten
Principal SQL Infrastructure Consultant
Microsoft SQL Server MVP

"Praful Varsani" <PrafulVarsani (AT) discussions (DOT) microsoft.com> wrote

Hvent done that. Need to do this during the maintenance window.

To test the failover this was carried out

Attempted to fail over from node 2 to node 1. Failover is unsuccessful.

Both Nodes were shutdown. Node 2 started first. The node 1 was started. SQL
Databases is now le on node 2.

SQL Server ENT V9.00 4035.00
Windows server 2003 R2 ENT

Well there was a issue with the switch that was conncected to the NIC cards
and should have failed over to another node.



"Geoff N. Hiten" wrote:

If a node is in shutdown, it will not accept new groups.

If a network switch fails and the cluster is segmented, no failover occurs
since (a) the cluster cannont communicate the failover requests and (b)
there is no node that can accept support the IP address resource(s).

In short, it appears the cluster is behaving normally.

You definitely need to test your failover scenarios before putting the
system into production so you know what failures are capable of being
handled by clustering.

--
Geoff N. Hiten
Principal SQL Infrastructure Consultant
Microsoft SQL Server MVP


"Praful Varsani" <PrafulVarsani (AT) discussions (DOT) microsoft.com> wrote

what do you recommend for failover scenarios

"Geoff N. Hiten" wrote:

Test every likely scenario to see what happens. NIC failures, HBA failures,
unplug a server, kill the switch. I usually do this with the SQL services
offline and just a dummy disk/IPaddress/Name resource group.

Se what triggers a failover and what does not.

Clustering alone will not give you High Availability. You need People and
Process to complement the Technology.

--
Geoff N. Hiten
Principal SQL Infrastructure Consultant
Microsoft SQL Server MVP


"Praful Varsani" <PrafulVarsani (AT) discussions (DOT) microsoft.com> wrote

Are there other options to check failover other than manual failovers//

"Geoff N. Hiten" wrote:

I usually move all disk resources to all nodes before installing SQL, just
to make sure everything works smoothly.

--
Geoff N. Hiten
Principal SQL Infrastructure Consultant
Microsoft SQL Server MVP

"Praful Varsani" <PrafulVarsani (AT) discussions (DOT) microsoft.com> wrote

sure...Can you elaborate more onhow would you move all the disk resources
example, go to resources, select the disks etc etc

"Geoff N. Hiten" wrote:

Um. Right click on the group - "Move Group"

--
Geoff N. Hiten
Principal SQL Infrastructure Consultant
Microsoft SQL Server MVP


"Praful Varsani" <PrafulVarsani (AT) discussions (DOT) microsoft.com> wrote