Linchi,
No, I performed an Initiate Failure on the resource. All that happens is
the resource goes into a failed state, then back online to the same node.
Thanks,
Jim
> How did you do your failover test? Did you try to move the group?
>
[quoted text clipped - 7 lines]
> >
> > Jim
Geoff N. Hiten - 23 Jul 2008 22:47 GMT
This is normal and expected.
The Cluster "fails" the resource but since there are no underlying resources
failed (disks, network,etc. are all online), the Cluster restarts the
instance right where it was.
Move Group will move everything over. Physically failing a resource will
also cause a failover. I usually take SQL offline manually and test
failover using the cluster group. I do things like pull HBA cables as part
of my test suite.

Signature
Geoff N. Hiten
Principal SQL Infrastructure Consultant
Microsoft SQL Server MVP
> Linchi,
>
[quoted text clipped - 17 lines]
>> >
>> > Jim
John Toner [MVP] - 24 Jul 2008 15:28 GMT
Just adding to what Geoff wrote, the default behavior of a cluster resource
is to restart the resource if it fails. Cluster will do this 3 times (by
default in 2003) and if it is unsuccessful after the 3rd attempt, it will
failover the group owning the resource.You can see/adjust these settings by
looking at the properties of a resource and look at the Advanced tab.
So if you want to test using the Initiate Failure option, you will need to
initiate a failure 4 times before the group would failover.
Regards,
John
Visit my blog: http://msmvps.com/blogs/jtoner
> Linchi,
>
[quoted text clipped - 16 lines]
> > >
> > > Jim