Home | Contact Us | FAQ | Search & Site Map | Link to Us
Sign In | Join | Other 45 Sites in Network
Home
Discussion GroupsWindows Server 2003Windows 2000Windows NTSmall Business ServerVirtual ServerExchange ServerIISHost Integration ServerISA ServerSMSWSUSMOMWindows Media ServerSecurityCertification
Related Topics
SQL ServerMS WindowsMS OfficePC HardwareMore Topics ...

Windows Server Forum / Exchange Server / Design / November 2006

Tip: Looking for answers? Try searching our database.

Geo Clustering - Longish Post

Thread view: 
Enable EMail Alerts  Start New Thread
Thread rating: 
AndyJ - 03 Nov 2006 20:07 GMT
Hello All

Here is what I want to achieve:

I want to have a 2 Node Exchange 2003 cluster (A/P) stretched across 2
locations, 100 metres between the buildings connected via fibre.
Latency is not an issue.

The cluster nodes will be SAN attached with each node having a
connection to the primary SAN. In the DR location I want to have a DR
SAN which will be an exact copy of the primary SAN from a hardware and
software perspective.

SAN replication will replicate exchange data including the logs and the
Quorum disk in real-time from the primary SAN to the DR SAN.

Can someone tell me with this configuration if we lose the primary site
and we are in a DR situation could we simply present the DR SAN to the
remaining node in the DR location?

I was hoping that upon boot up the cluster node would be able to see
the quorum resource and therefore the EVS should start correctly and
clients will be able to connect

Ideally I would like both nodes to be connected to their local SAN and
for any failover to happen automatically but our storage vendor says
this is not possible. I have this question out with them as the guy we
spoke with was not an Exchange expert.

The vendor actually suggested that we used Replistore to replicate to a
DR Exchange server in our DR location and not to stretch the cluster
between the locations. I like this solution as it does provide
automatic failover but is also going to cost us quite a bit more money.

He also said that you cannot present a server that is a part of a
cluster to the DR SAN without first removing it from the cluster,
anyone know if this is true?

Thanks

AJ

PS we already have the SAN hardware allocated to the project and we
can't wait for Exchange 2007  :)
Mark Arnold [MVP] - 03 Nov 2006 20:48 GMT
>Hello All
>
[quoted text clipped - 40 lines]
>PS we already have the SAN hardware allocated to the project and we
>can't wait for Exchange 2007  :)

Stop, just stop.
Go and talk to the SAN team/vendor.
They will do this solution, in their sleep, without you going creating
completely unsupported solutions with server clustering.
Essentially you're looking at booting the server from the SAN and the
having spare tin at the other site (as you want). This will allow you
to keep the server solution sensible (read simple) and get the server
to boot and run Exchange is about the same time as a failover would
occur.
AndyJ - 03 Nov 2006 20:55 GMT
> >Hello All
> >
[quoted text clipped - 50 lines]
> to boot and run Exchange is about the same time as a failover would
> occur.
AndyJ - 03 Nov 2006 21:08 GMT
I am not creating unsupported solutions and we are talking to the
vendor if you read the post.

I am simply asking a question to see if what we want to do is possible,
its another idea we have had.

The SAN replication methods (RSME/VSS Clones/SAN copy and Replistore)
are all SUPPORTED replication solutions. We are not looking to boot
from SAN and the storage vendor (EMC) never mentioned this either as a
possible solution.

My question was simple I thought. Will the node in the DR location come
up and service clients or not?

Can you explain how your boot from SAN scenario is beneficial to us in
the scenario I talk about below? then maybe it will make sense.

Thanks

AJ

It was simply a question, do you have an aswer for it or not?

> >Hello All
> >
[quoted text clipped - 50 lines]
> to boot and run Exchange is about the same time as a failover would
> occur.
Mark Arnold [MVP] - 03 Nov 2006 22:32 GMT
>I am not creating unsupported solutions and we are talking to the
>vendor if you read the post.
[quoted text clipped - 16 lines]
>
>AJ

Whilst it can be done and I was going to give you some NetApp slanted
information you'll have seen the post from JF already who is a NetApp
employee, saving me the typing and adding more much credibility than
I'd have given you.

I'd still avoid this as a solution though. Why? Because your interest
in Exchange 2007 will mean that you have less work to do if you
maintain a couple of high availability boxes that aren't clustered.

So, EMC didn't talk to you about boot from SAN? Not even to tell you
why you didn't want it or why they were excluding it from their
design? Wow, if my company (we're the biggest NetApp SI  in the UK by
a long margin) had done (or not done to be more accurate) that then
there would have been questions asked.
AndyJ - 06 Nov 2006 09:37 GMT
Thanks Mark. In fairness to EMC the guy we had the initial meeting with
was not an Exchange guru, more a generalist. I will carry on talks with
EMC. I have already tried to convince the company in question regarding
Exchange 2007 but they may not move to this for sometime.  Too bleedin
edge for them which is understandable at this stage......

AJ

> >I am not creating unsupported solutions and we are talking to the
> >vendor if you read the post.
[quoted text clipped - 31 lines]
> a long margin) had done (or not done to be more accurate) that then
> there would have been questions asked.
AndyJ - 06 Nov 2006 10:05 GMT
Thanks Mark. In fairness to EMC the guy we had the initial meeting with
was not an Exchange guru, more a generalist. I will carry on talks with
EMC. I have already tried to convince the company in question regarding
Exchange 2007 but they may not move to this for sometime.  Too bleedin
edge for them which is understandable at this stage......

AJ

> >I am not creating unsupported solutions and we are talking to the
> >vendor if you read the post.
[quoted text clipped - 31 lines]
> a long margin) had done (or not done to be more accurate) that then
> there would have been questions asked.
John Fullbright [MVP] - 03 Nov 2006 21:06 GMT
I did a design very similar to this for Arizona Sate University.   I did
however use an MNS cluster and synchronous Snapmirror .

http://www.netapp.com/library/cs/asu.pdf?xCountry=US&xLanguage=EN

> Hello All
>
[quoted text clipped - 40 lines]
> PS we already have the SAN hardware allocated to the project and we
> can't wait for Exchange 2007  :)
AndyJ - 03 Nov 2006 21:16 GMT
Thanks John, very interesting.

AJ

> I did a design very similar to this for Arizona Sate University.   I did
> however use an MNS cluster and synchronous Snapmirror .
[quoted text clipped - 45 lines]
> > PS we already have the SAN hardware allocated to the project and we
> > can't wait for Exchange 2007  :)
 
Sign In
Join
My Latest Posts
My Monitored Threads
My Blog
My Photo Gallery
My Profile
My Homepage

Start New Thread
Enable EMail Alerts
Rate this Thread



©2009 Advenet LLC   Privacy Policy - Terms of Use
This website includes both content owned or controlled by Advenet as well as content owned or controlled by third parties.