| GNocent 2006-01-10, 9:24 am |
| I agree with Mike, stopping cluster service or shutdowning the node is the
worst way of doing a failover, as it forces it instead of doing it "quietly".
The major difference is that if you have any problem during the failover, it
will be much more difficult to failback quickly ...
For the steps, no matter what order you chose, but it is generally
preferable to move DTC resource first, and then SQLSever instances.
Last, remember to avoid moving a group when quorum is moving as writes to
quorum disk are done when moving any group.
Guillaume.
==============
"Michael Hotek" wrote:
> Yes, it is gracefully shutdown before failing over. What happens is that
> the cluster service goes offline. The heartbeat fails. So, the cluster
> service on the other node initiates failover. This process issues a stop to
> the SQL Server service. Once shutdown, it changes ownership of the
> underlying physical resources and then starts the service on the other node.
>
> I definitely do NOT recommend doing administrative failovers that way. It's
> basically the same and walking over to the machine and yanking the power
> plug. It takes just as much effort to right click and select Move Group as
> it does to yank the cluster service out from underneath running resources.
>
> --
> Mike
> Mentor
> Solid Quality Learning
> http://www. solidqualitylearning
.com
>
>
> "Sean" <spowell71@hotmail.com> wrote in message
> news:1136735619.646895.101960@g49g2000cwa.googlegroups.com...
>
>
>
|