proper shutdown order for MSCS SQL cluster?

Hi,
We're having a SAN outage & I've noticed some **flakiness with regard to a
'clean' shutdown of our SQL 2000(active/active) & 2005(active/passive)
clusters.
**In the past when we've disconnected storage and had to cold boot all the
cluster nodes, we've had trouble starting the cluster back up. ( network
name not found in the cluster resource group and/or cluster services
wouldn't start)
Is there a recommended order to shut down & restart cluster resource groups
(including SQL+DTC) when system maintenanence need to occur?
Thanks,
Mike Bonvie
Shutdown:
Take the SQL resources offline and set the startup type to DISABLED on all
nodes
Stop the Cluster Service on all nodes and set the startup type to disabled.
Perform hardwarework as necessary.
Only boot one node at a time. If all nodes are down, us a staggered startup
sequence.
Enable and start cluster services on each node. Test moving the cluster
group.
Set SQL Service startup type to MANUAL.
Using the cluster admin tool, bring the diskm IP address, and network name
resources online. Test moving them across nodes.
Bring the SQL Services online using the cluster tool. Test moving groups.
Geoff N. Hiten
Senior Database Administrator
Microsoft SQL Server MVP
"Mike Bonvie" <m@.d.net> wrote in message
news:e7WcCzMYHHA.1388@.TK2MSFTNGP05.phx.gbl...
> Hi,
> We're having a SAN outage & I've noticed some **flakiness with regard to a
> 'clean' shutdown of our SQL 2000(active/active) & 2005(active/passive)
> clusters.
> **In the past when we've disconnected storage and had to cold boot all the
> cluster nodes, we've had trouble starting the cluster back up. ( network
> name not found in the cluster resource group and/or cluster services
> wouldn't start)
> Is there a recommended order to shut down & restart cluster resource
> groups (including SQL+DTC) when system maintenanence need to occur?
> Thanks,
> Mike Bonvie
|||Thank You Geoff,
Great information. Is this based on Microsoft documentation, or a combo of
experience/documentation?
If there's a KB or other resource, where can I get it?
Thanks again,
Mike Bonvie
"Geoff N. Hiten" <SQLCraftsman@.gmail.com> wrote in message
news:u$vRT9MYHHA.4440@.TK2MSFTNGP03.phx.gbl...
> Shutdown:
> Take the SQL resources offline and set the startup type to DISABLED on all
> nodes
> Stop the Cluster Service on all nodes and set the startup type to
> disabled.
> Perform hardwarework as necessary.
> Only boot one node at a time. If all nodes are down, us a staggered
> startup sequence.
> Enable and start cluster services on each node. Test moving the cluster
> group.
> Set SQL Service startup type to MANUAL.
> Using the cluster admin tool, bring the diskm IP address, and network name
> resources online. Test moving them across nodes.
> Bring the SQL Services online using the cluster tool. Test moving groups.
>
> --
> Geoff N. Hiten
> Senior Database Administrator
> Microsoft SQL Server MVP
>
>
> "Mike Bonvie" <m@.d.net> wrote in message
> news:e7WcCzMYHHA.1388@.TK2MSFTNGP05.phx.gbl...
>
|||This is a mix of best practices guidelines and experience. The staggered
startup is documented somewhere, but the other steps are stuff I have worked
out over the years. The DISABLED service ettings just makes sure the
resource won't attempt to come online accidentally.
Geoff N. Hiten
Senior Database Administrator
Microsoft SQL Server MVP
"Mike Bonvie" <m@.d.net> wrote in message
news:Oqu9AHNYHHA.4552@.TK2MSFTNGP05.phx.gbl...
> Thank You Geoff,
> Great information. Is this based on Microsoft documentation, or a combo of
> experience/documentation?
> If there's a KB or other resource, where can I get it?
> Thanks again,
> Mike Bonvie
>
> "Geoff N. Hiten" <SQLCraftsman@.gmail.com> wrote in message
> news:u$vRT9MYHHA.4440@.TK2MSFTNGP03.phx.gbl...
>
sql