Sorry for the OT, but I haven't received any answers on the VMware forums.

I need to set a 2 node Win2008 R2/SQL 2008 R2 cluster on ESXi 5.5 U3b.
This will be only for testing; I have a production cluster on 2
physical nodes already, but what I need to do is test making some
changes to the configuration, so I thought I would test the procedure
on a couple VMs, rather than barrel ahead on the production
machine.What I need to test is first removing one of the SQL
instances, then removing one of the nodes completely.

I know this cluster is supported on my ESXi; what I don't have is a
specific step-by-step HOWTO. My production cluster is 2 nodes SQL 2008
R2 cluster, and it has 2 SQL instances. What I need to test is first
removing one of the instances, then removing one of the nodes
completely.

>From what I've read, this should be one of those "cluster-in-a-box"
configurations.

I know I need 2 NICs, one for the heartbeat, one for the public IP of
the node. I'm pretty sure I know how to handle the heartbeat: a
virtual switch that has no physical NICs assigned to it, and have the
heartbeat NIC use it; that way, the only thing it can reach is only on
that private vswitch, i.e., the other node. I've done similar in the
past.

I only need like a 2G shared disk (I'm not really going to have any
data on it, I just need it to exist as a cluster shared storage), but
I know it has to be shared between both VMs. It doesn't have to be
RDM, right? I can do this with a regular virtual disk? Do I need a
special SCSI adapter for each VM? And just tell node #2 to reuse an
existing virtual disk? (I've found an article that says that, but I
don't know if that's still required?).

I'm searching, but getting confused at what I am finding. Anyone know
of a good step-by-step article to follow, along with GOTCHAs? As I
said, what I need to do is test the process of removing a SQL instance
from a SQL cluster (leaving the other, named instance alone, as that's
still in production), and then removing the SQL node from the SQL
cluster, and then removing that same node from the Windows cluster.
I'm pretty sure I know how to do the last 2, it's the remove an
instance without screwing up the remaining instance I am leery about
...

 Thanks. As I said, a bit OT, but this list seems to have people who
have done every sort of imaginable configuration before, so I'm hoping
someone has ... :-)


Reply via email to