Re: MSCS DR
- From: "John Fullbright [MVP]" <fjohn@donotspamnetappdotcom>
- Date: Wed, 15 Nov 2006 09:51:29 -0800
Hi Chuck, It's been a while since EMC class... 2002, 2003? I don't remember
the year.
Anyway the point is to have the vendor, whatever vendor you choose, clearly
define who is supporting what. The less ambiguity you have, the greater the
chance of success for your project.
John
"Chuck Timon [Microsoft]" <ctimon@xxxxxxxxxxxxxxxxxxxx> wrote in message
news:O3Zfya1BHHA.1224@xxxxxxxxxxxxxxxxxxxxxxx
Of course....that is their product.....why wouldn't they 'hype' it.
--
Chuck Timon, Jr.
Microsoft Corporation
Longhorn Readiness Team
This posting is provided "AS IS" with no
warranties, and confers no rights.
"AndyJ" <andyjones99@xxxxxxxxxxxxx> wrote in message
news:1163435140.608088.82820@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
One would hope so. They (EMC) have now pretty much stated to me that
they do not have a true geocluster solution that can be used with the
CX400 arrays so it looks as though we will be looking at Replistor
after all this.
Thanks
AJ
Chuck Timon [Microsoft] wrote:
Have they read this - http://support.microsoft.com/?kbid=895847
--
Chuck Timon, Jr.
Microsoft Corporation
Longhorn Readiness Team
This posting is provided "AS IS" with no
warranties, and confers no rights.
"AndyJ" <andyjones99@xxxxxxxxxxxxx> wrote in message
news:1163412198.802750.269300@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Hello John
Thanks for the further information. Regarding your question, EMC have
not actually advocated the proposed design that I talked about in my
original post. They are stating that we cannot utilise the DR
replicated SAN with the existing cluster. They are saying that we need
to have a standalone DR server in the DR site which is not a part of
the cluster. In this case the solution would not be a geocluster (Both
nodes would be in the primary location) and we would have to shell out
more money for a decent standalone DR Exchange box. From what I have
read I do not think this is true and this is why I posted the message.
I am getting lots of contradicting information and have now called MS
and also asked to get this escalated within EMC.
EMC are suggesting we purchase Replistore (Which is just like
DoubleTakes software) to do the whole shooting match to a DR server
or
Use Replication Manger/SE to manage the replicas (local clones and
remote copies) utilising SANCopy for remote copy to the DR site of the
daily full copy of DB + logs (copying from a clone) and utilising
replistore for log shipping during the day to allow for roll-forward
from the consistent copy.
Both of these solutions involve a DR server which is what we want to
avoid.
They also stated this:
"As I discussed in the meeting; Microsoft and EMC best practice
are that full and validated copies of the Exchange database are used
for
all Exchange DR recovery and synchronous replication is not supported
for Exchange in any way."
From what I understand now this is not a true statement so you canunderstand why I am little perplexed (unless they are reffering to EMC
supportability of this solution)
Thanks
AJ
John Fullbright [MVP] wrote:
Async replication of live data? fully supported? I guess you
haven't
read
KB 895847.
"Support policies for the different kinds of replication mechanism
Asynchronous replication
Any data that has been replicated asynchronously is supported only by
the
storage vendor. When you replicate Exchange data by using
asynchronous
replication, the vendor who provides your replication mechanism must
support
the Exchange data at the replication destination. We make no claims
as to
the reliability of or the performance of Exchange solutions that use
asynchronous replication. We use the following criteria to support
Exchange-related issues that you may experience in an environment
where
you
use asynchronous replication: "
Is that what you mean by "fully supported"?
The only way that I know of to use async replication in a fully
supported
solution is to use data that is not live - ie. snapshots. You can
see
this
here:
http://www.microsoft.com/technet/technetmag/issues/2006/10/FailoverClusters/default.aspx
Note the concluding comment: "We achieved our goal of building
failover
clusters for Exchange to provide a disaster recovery solution that
has a
relatively high RTO and RPO, is fully supported by Microsoft, and
avoids
the
complexity of full geoclustering."
"AndyJ" <andyjones99@xxxxxxxxxxxxx> wrote in message
news:1163149156.984465.72970@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Hello All
I am posting in this forum in hope that an MSCS expert can help me
out
here. I posted in the Exchange forum but didn't get the answers I
was
after so here goes:
Here is the scenario:
We have an Exchange 2003 cluster stretched across two locations in
an
Active/Passive configuration. Both cluster nodes talk to the
primary
SAN array. The primary SAN is mirrored to the DR location using
Aysnc
replication (logs, stores and Quorom) and this is fully supported
by MS
at this point.
Here is the question:
In the event that we loose the primary site (A node and SAN) will I
be
able to present the DR SAN disk array to the node in the DR
location
and bring the EVS up?
I am thinking that MSCS might see the DR disks as a new disk
resource
and will choke on this.
Any thoughts. I have been discussing this with the storage vendor
but
they cannot answer this question.
Thanks
AJ
.
- References:
- MSCS DR
- From: AndyJ
- Re: MSCS DR
- From: John Fullbright [MVP]
- Re: MSCS DR
- From: AndyJ
- Re: MSCS DR
- From: Chuck Timon [Microsoft]
- Re: MSCS DR
- From: AndyJ
- Re: MSCS DR
- From: Chuck Timon [Microsoft]
- MSCS DR
- Prev by Date: Re: windows 2003 clusters on different subnet
- Next by Date: Re: Disk Cluster Resource stays in online pending state
- Previous by thread: Re: MSCS DR
- Next by thread: Re: MSCS DR
- Index(es):
Relevant Pages
|
