I have a two node cluster. One node has crashed and I need to know the best approach to rebuild the second node. In my case the node crashed from a hardware failure. This client did not have a good backup of the system state for this node. But I have a working node up and still running. We have evicted the failed node from the cluster. What steps should I perform to rebuild the failed node if the system state backup is not available? I taken a look at a white paper and some other docs that discribed methods to get back up but was interested in what the expert community advised. Also since the disk signature is lost can I get help with step that include this type of failure.

Look in BOL for step by step instructions on maintaining a failover SQL
cluster, including how to remove and replace a failed node. You should be
able to do this without disrupting the working node.

--
Geoff N. Hiten
Microsoft SQL Server MVP
Senior Database Administrator
Careerbuilder.com

I support the Professional Association for SQL Server
www.sqlpass.org

"billet" <anonymous (AT) discussions (DOT) microsoft.com> wrote

hardware failure. This client did not have a good backup of the system state
for this node. But I have a working node up and still running. We have
evicted the failed node from the cluster. What steps should I perform to
rebuild the failed node if the system state backup is not available? I taken
a look at a white paper and some other docs that discribed methods to get
back up but was interested in what the expert community advised. Also since
the disk signature is lost can I get help with step that include this type
of failure.

Hi
Build the second node as a stand-alone server. Add you
clustering services (MSCS) to it (as a secondary node to
your active node). Then start the SQL setup and say you
want to add this machine as a failover for the exisiting
SQL virtual server name. Works a treat.
-J

A couple of things I missed. First we have a Active/Passive cluster. Second the rebuilt node is comcust.exe run on that node? right? Ok. Should I run SP4 for the OS W2K next or join the cluster? Then run SQL install on the currenly good node
Will I need a system state backup to fix the new node or will new node act as if it was a newly built node that just joined the cluster

Thanks,

Thanks for that information. A friend told me that following your last suggestion I should delete SQL information execpt the binaries and the second node would do a discovery. Is this true or will the second node sync up with the active node when I fail over to the new second node?

On the existing node, run SQL install and remove the old node. Evict the
node From the cluster.

For the new node, build the OS and OS service pack first. Join to the
existing cluster. Reapply service pack. to the new node only. Verify the
cluster works (you should be able to move the quorum group at this time).

From the old node, run the SQL install and add the new node to SQL. Reboot
the new node. From the new node, apply the service pack (it will run a
binary update only). REboot the new node again.

Your cluster should be fully restored at this time,


--
Geoff N. Hiten
Microsoft SQL Server MVP
Senior Database Administrator
Careerbuilder.com

I support the Professional Association for SQL Server
www.sqlpass.org

"billet" <anonymous (AT) discussions (DOT) microsoft.com> wrote

run SP4 for the OS W2K next or join the cluster? Then run SQL install on the
currenly good node.

I just want to say thanks a bunch for the heads up information the newsgroup provided. In my lab I was able to reproduce the rebuild process. This was great information to feedback to the client and estimate down time. My finding is that it will only take a Max of 30 minutes to reapply W2K SP4, reinstall cluster services join the node back and run SQL wizard, apply SP3a and a hotfix. Not bad for a complete rebuild of a failed node

Thanks again to the newsgroups experts

Bill

You are welcome.

If you find yourself in the Atlanta area, you can buy me a beer.

--
Geoff N. Hiten
Microsoft SQL Server MVP
Senior Database Administrator
Careerbuilder.com

I support the Professional Association for SQL Server
www.sqlpass.org

"billet" <anonymous (AT) discussions (DOT) microsoft.com> wrote

This was great information to feedback to the client and estimate down time.
My finding is that it will only take a Max of 30 minutes to reapply W2K SP4,
reinstall cluster services join the node back and run SQL wizard, apply SP3a
and a hotfix. Not bad for a complete rebuild of a failed node.