cool and great!
At last, I can hear good news :-)
Sir Robert, how long has the cluster been running (without downtime)?
Did you have some difficulty? (You could share your experience off list
since this topic is too general, IMHO).
Thanks,
-botp
-----Original Message-----
From: Robert Ellis [mailto:[EMAIL PROTECTED]]
Sent: Thursday, September 06, 2001 3:11 PM
To: Exchange Discussions
Subject: RE: [Exchange2000] Exchange Fault Tolerance
we have a 2-node active/passive exchange 2000 cluster running fine. i can
fail the exchange group over, and downtime is generally around 60 seconds.
We are looking at possibly going with 2 x 4-node datacenter clusters to
support 20,000 exchange users.
Rob
-----Original Message-----
From: "Peņa, Botp" [mailto:[EMAIL PROTECTED]]
Sent: Thursday, September 06, 2001 5:07 AM
To: Exchange Discussions
Subject: RE: [Exchange2000] Exchange Fault Tolerance
Thanks, sir Andy.
I have seen firewalls clusters and web farms in action. And was quite very
impressed. None yet for Exchange though. In your case then, can we fairly
say that Exchange clusters are a flop (or "forget it" as some would say)?
You did not answer as to whether we should wait or "forget it -it is
impossible for exchange to attain 5 9s in a cluster, ever" (again, I am just
asking your opinion since you are well respected in this list).
Thanks,
-botp
-----Original Message-----
From: Webb, Andy [mailto:[EMAIL PROTECTED]]
Sent: Thursday, September 06, 2001 11:49 AM
To: '"Peņa, Botp"'
Subject: RE: [Exchange2000] Exchange Fault Tolerance
Definitely 1. 2 is a matter of your expertise and the ultimate roi of the
clustered solution in your environment.
I think clusters have their place, but it's more with applications which
behave more like the clusters do. People seem to be unsatisfied with their
clusters in my experience once they've seen them in action.
=======================================================
Andy Webb [EMAIL PROTECTED] www.swinc.com
Simpler-Webb, Inc. Austin, TX 512-322-0071
-- Way to go USPS Cycling Team and Lance Armstrong!! --
=======================================================
-----Original Message-----
From: "Peņa, Botp" [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, September 05, 2001 10:39 PM
To: 'Webb, Andy'
Subject: RE: [Exchange2000] Exchange Fault Tolerance
Thanks, Sir Andy. Very helpful and well said.
To summarize your point (in so far as I can comprehend):
1. clusters add complexity
2. clusters are risky
Do your statements apply to clusters in general, or is it only for Exchange
(or as to the current Exchange clustering stage)? I am asking since we might
abandon the plan totally or just wait for another year or years....
Thank you again,
-botp
-----Original Message-----
From: Webb, Andy [mailto:[EMAIL PROTECTED]]
Sent: Thursday, September 06, 2001 3:27 AM
To: '"Peņa, Botp" '
Cc: '[EMAIL PROTECTED]'
Subject: RE: [Exchange2000] Exchange Fault Tolerance
Knowing no more about your business than what is in that note, I'd say
you're needlessly adding complexity to your environment which will not
result in the net availability increase you forecast at the low cost you
expect.
There is downtime associated with a clean node swap in a clustered
environment. Is 2x that downtime significantly less than that required for
a service pack install? Not usually. Does it let you apply one or more
hotfixes with less impact to the users? Not usually.
Running a clustered environment also means you need to be running a
clustered test environment so that you can simulate all these rolling
upgrades. Having that AV upgrade hose an entire node of your cluster is not
significantly better than having it hose a standalone server, and you may
not know that things are bad until you've moved the active node back to the
already upgraded server.
As noted above, I don't know enough about your environment to make this a
firm recommendation. To spend the time to get that familiarity, I'd likely
want a consulting contract. Clustering is not for the faint of heart, and
it is not to be used without serious thought to the whys and whatfors. Do
you really have a business need for mail to every user to be available five
nines or more? You may, but you would be one of a very few. I'd rather
take that second server and use it for disaster recovery drills and testing
of all the software upgrades you mentioned. Then my actual downtime would
be minimized by having a known, tested, and understood upgrade process that
could proceed quickly. Most service packs and hotfixes are 10-15 minute
activities if planned and practiced.
$0.02
Andy
Andy
_________________________________________________________________
List posting FAQ: http://www.swinc.com/resource/exch_faq.htm
Archives: http://www.swynk.com/sitesearch/search.asp
To unsubscribe: mailto:[EMAIL PROTECTED]
Exchange List admin: [EMAIL PROTECTED]
The contents of this email and any attachments are sent for the personal
attention
of the addressee(s) only and may be confidential. If you are not the
intended
addressee, any use, disclosure or copying of this email and any attachments
is
unauthorised - please notify the sender by return and delete the message.
Any
representations or commitments expressed in this email are subject to
contract.
ntl Group Limited
_________________________________________________________________
List posting FAQ: http://www.swinc.com/resource/exch_faq.htm
Archives: http://www.swynk.com/sitesearch/search.asp
To unsubscribe: mailto:[EMAIL PROTECTED]
Exchange List admin: [EMAIL PROTECTED]
_________________________________________________________________
List posting FAQ: http://www.swinc.com/resource/exch_faq.htm
Archives: http://www.swynk.com/sitesearch/search.asp
To unsubscribe: mailto:[EMAIL PROTECTED]
Exchange List admin: [EMAIL PROTECTED]
