Re: Windows Multi-node RAC (VMware)

"Dan Yusuph" <oradba_at_mail.kz> wrote in message news:150627cb.0212222225.24387366_at_posting.google.com...
> Hello Group,
>
> I have successfully installed single-node multi-instance 9.2.0.2.1 RAC on
> W2K using "how-to" document from oaktable.net.
> Now I'm trying next how-to: multi-node RAC. There is a requirement that
> virtual nodes must use "vmware host-only networking", hence both nodes
> should be configured using DHCP. But it makes the cluster accessible
> only from "vmware host os", and inaccessible from any other machine in
> real network.

Yup. It's a limitation of the paper, for sure. But there's no way around it.

> Have anybody configured vmware multinode RAC either with static IP
> addressing, or by configuring virtual switches and gateways, or by any
> other way? The goal is to access virtual RAC database not from vmware
> host os, but from client workstation in real network.

Nope, don't even try it. The paper makes its aims perfectly clear: this is for testing a two-node RAC implementation on a domestic single PC. It's there to experiment with from the DBAs perspective. Use the host machine to make connection to the RAC and test out connectivity and so forth. That's as far as it will go.

Out of interest, the limitation being dealt with in that paper's recommendation to use host only networking is that the bridged networking doesn't work. It's got nothing to do with static or dynamic IP addresses. Your two node RAC will work fine with static IP addresses, but won't work with bridged networking.

All the author was doing was saying that since host-only networking happens to include a virtual DHCP server, you might as well not bother setting IP addresses of your own, but get the DHCP server to do it for you. Static or dynamic, that's not the problem.

> Thank you.
>

Why don't you write to the author direct, since he includes his email address at the end of the paper?

Regards
HJR
> Dan Yusuph
Received on Mon Dec 23 2002 - 06:01:57 CST