|
|
|
date: Wed, 09 Jan 2008 11:55:19 +0000,
group: microsoft.public.exchange.clustering
back
Re: Exchange 2007 SP1 SCR standby continuous replication
The SCR box won't be part of the CCR cluster.
If you're using CCR at your production site, implementing a standby cluster
at the remote site would probably make the most sense since, if you run from
the standby site for any length of time, I presume you'd want similar high
availability for the duration. So for the standby cluster, you install only
the passive clustered mailbox server role on each node and then enable SCR
to one of them from the source CCR. When the time comes, you recover the
CMS onto the node you enabled for SCR and also seed to the other passive
node to bring up a new CCR environment at the standby site. This also
assumes you have AD, DNS, Hub Transport, CAS, etc, etc at the standby site.
--
Neil Hobson
Exchange MVP
http://www.msexchange.org/Neil_Hobson/
http://www.msexchangeblog.com
"Nick" wrote in message
news:qgd9o31tgdrs5co4s9oi9fnonbo4ch1aas@4ax.com...
> Hi
>
> We currently have an Exchange 2007 enviroment with 2 x CCR mailbox (
> on Windows 2003 R2 Sp2) clusters in production. We have just
> upgraded to Exchange 2007 SP1 and are looking to utilise SCR at our
> D.R site over a 1gig connection.
>
> I have read through the Microsoft documentation, however I am unclear
> on a few details.
>
> Does the server that we use for SCR need to be a member of the CCR
> cluster, or do you install only the "mailbox role" on the server. I
> know the same OS, drive letters and installation locations are
> required.
>
> I presume it can be on a physically seperate subnet or Vlan.
>
> We would like to have the capability to use the SCR server for
> contingency and disaster recovery in the event we lose our local CCR
> cluster.
>
> Any advise would be appreciated.
>
> Nick Thompson
date: Wed, 9 Jan 2008 13:50:24 -0000
author: Neil Hobson [MVP]
Re: Exchange 2007 SP1 SCR standby continuous replication
Hi Neil
Thanks for the prompt reply. I'm still a bit confused. To confirm,
with our current setup below:
2 x Two node CCR clusters running on Windows 2003 R2 with SP2 and
Exchange SP1. Both active nodes are on san disks, and both passive
nodes are on local raid arrays. If we have a san failure Exchange
2007 will still be avaialble on the passive nodes. Both clusters are
running using a majority node set quorum with file share witness on
the Hub Transport server.
We are happy that this setup will give us maximum service and data
availability.
Relistically we would only go to our D.R site to perform test recovery
scenarios, and in the unlikely event of a major issue with our primary
production site, we would then utilise our D.R site.
We wouldnt have an issue running in a non high availability enviroment
at D.R for some time, as most of other other services at D.R are the
same.
I believe it is possible to activate SCR using the following 3
methods.
1. Database Portability
2. Recovery CMS ( for clusters )
3. Recovery Server
What one is the most transpart to use from a Outlook perspective?
We use a third party application( EMC EmailXtender) which uses email
journaling on Exchange for email archiving and retention. This service
effectively expects to talk to a specific Exchange server and
database.
Im assuing using the "Recover CMS" or "Recovery Server" option to
restore Exchange would result in a different server name, ip address.
If so this would potentially prevent our third party application from
working.
Let me know if you need any further information .
Thanks again.
Nick Thompson
On Wed, 9 Jan 2008 13:50:24 -0000, "Neil Hobson [MVP]"
wrote:
>The SCR box won't be part of the CCR cluster.
>
>If you're using CCR at your production site, implementing a standby cluster
>at the remote site would probably make the most sense since, if you run from
>the standby site for any length of time, I presume you'd want similar high
>availability for the duration. So for the standby cluster, you install only
>the passive clustered mailbox server role on each node and then enable SCR
>to one of them from the source CCR. When the time comes, you recover the
>CMS onto the node you enabled for SCR and also seed to the other passive
>node to bring up a new CCR environment at the standby site. This also
>assumes you have AD, DNS, Hub Transport, CAS, etc, etc at the standby site.
date: Wed, 09 Jan 2008 16:51:12 +0000
author: Nick
Re: Exchange 2007 SP1 SCR standby continuous replication
If you lost your primary site, you'd likely go down the "recover CMS" route
as this brings up the CMS at the DR site. The CMS name will stay the same
(the name which Outlook connects to) but during the CMS recovery process
you'd give the CMS a new IP address since, most likely, the DR site is in a
different subnet. Once DNS sorts itself out all should be ok. So whilst
the IP address may change, the CMS name won't.
--
Neil Hobson
Exchange MVP
http://www.msexchange.org/Neil_Hobson/
http://www.msexchangeblog.com
"Nick" wrote in message
news:avs9o3pnjov4t0s6hc83vv6lnftqovua81@4ax.com...
> Hi Neil
>
> Thanks for the prompt reply. I'm still a bit confused. To confirm,
> with our current setup below:
>
> 2 x Two node CCR clusters running on Windows 2003 R2 with SP2 and
> Exchange SP1. Both active nodes are on san disks, and both passive
> nodes are on local raid arrays. If we have a san failure Exchange
> 2007 will still be avaialble on the passive nodes. Both clusters are
> running using a majority node set quorum with file share witness on
> the Hub Transport server.
>
> We are happy that this setup will give us maximum service and data
> availability.
>
> Relistically we would only go to our D.R site to perform test recovery
> scenarios, and in the unlikely event of a major issue with our primary
> production site, we would then utilise our D.R site.
>
> We wouldnt have an issue running in a non high availability enviroment
> at D.R for some time, as most of other other services at D.R are the
> same.
>
> I believe it is possible to activate SCR using the following 3
> methods.
>
> 1. Database Portability
> 2. Recovery CMS ( for clusters )
> 3. Recovery Server
>
> What one is the most transpart to use from a Outlook perspective?
>
> We use a third party application( EMC EmailXtender) which uses email
> journaling on Exchange for email archiving and retention. This service
> effectively expects to talk to a specific Exchange server and
> database.
>
> Im assuing using the "Recover CMS" or "Recovery Server" option to
> restore Exchange would result in a different server name, ip address.
> If so this would potentially prevent our third party application from
> working.
>
> Let me know if you need any further information .
>
> Thanks again.
>
> Nick Thompson
>
>
>
>
>
>
>
>
> On Wed, 9 Jan 2008 13:50:24 -0000, "Neil Hobson [MVP]"
> wrote:
>
>>The SCR box won't be part of the CCR cluster.
>>
>>If you're using CCR at your production site, implementing a standby
>>cluster
>>at the remote site would probably make the most sense since, if you run
>>from
>>the standby site for any length of time, I presume you'd want similar high
>>availability for the duration. So for the standby cluster, you install
>>only
>>the passive clustered mailbox server role on each node and then enable SCR
>>to one of them from the source CCR. When the time comes, you recover the
>>CMS onto the node you enabled for SCR and also seed to the other passive
>>node to bring up a new CCR environment at the standby site. This also
>>assumes you have AD, DNS, Hub Transport, CAS, etc, etc at the standby
>>site.
date: Fri, 11 Jan 2008 08:38:32 -0000
author: Neil Hobson [MVP]
Re: Exchange 2007 SP1 SCR standby continuous replication
Another possiblity, one that works well if you have two active sites that
play the role of DR site for each other, is to leverage database portability
and do a restore to alternate server.
"Neil Hobson [MVP]" wrote in message
news:u4qma1CVIHA.3400@TK2MSFTNGP03.phx.gbl...
> If you lost your primary site, you'd likely go down the "recover CMS"
> route as this brings up the CMS at the DR site. The CMS name will stay
> the same (the name which Outlook connects to) but during the CMS recovery
> process you'd give the CMS a new IP address since, most likely, the DR
> site is in a different subnet. Once DNS sorts itself out all should be
> ok. So whilst the IP address may change, the CMS name won't.
>
> --
> Neil Hobson
> Exchange MVP
> http://www.msexchange.org/Neil_Hobson/
> http://www.msexchangeblog.com
>
>
> "Nick" wrote in message
> news:avs9o3pnjov4t0s6hc83vv6lnftqovua81@4ax.com...
>> Hi Neil
>>
>> Thanks for the prompt reply. I'm still a bit confused. To confirm,
>> with our current setup below:
>>
>> 2 x Two node CCR clusters running on Windows 2003 R2 with SP2 and
>> Exchange SP1. Both active nodes are on san disks, and both passive
>> nodes are on local raid arrays. If we have a san failure Exchange
>> 2007 will still be avaialble on the passive nodes. Both clusters are
>> running using a majority node set quorum with file share witness on
>> the Hub Transport server.
>>
>> We are happy that this setup will give us maximum service and data
>> availability.
>>
>> Relistically we would only go to our D.R site to perform test recovery
>> scenarios, and in the unlikely event of a major issue with our primary
>> production site, we would then utilise our D.R site.
>>
>> We wouldnt have an issue running in a non high availability enviroment
>> at D.R for some time, as most of other other services at D.R are the
>> same.
>>
>> I believe it is possible to activate SCR using the following 3
>> methods.
>>
>> 1. Database Portability
>> 2. Recovery CMS ( for clusters )
>> 3. Recovery Server
>>
>> What one is the most transpart to use from a Outlook perspective?
>>
>> We use a third party application( EMC EmailXtender) which uses email
>> journaling on Exchange for email archiving and retention. This service
>> effectively expects to talk to a specific Exchange server and
>> database.
>>
>> Im assuing using the "Recover CMS" or "Recovery Server" option to
>> restore Exchange would result in a different server name, ip address.
>> If so this would potentially prevent our third party application from
>> working.
>>
>> Let me know if you need any further information .
>>
>> Thanks again.
>>
>> Nick Thompson
>>
>>
>>
>>
>>
>>
>>
>>
>> On Wed, 9 Jan 2008 13:50:24 -0000, "Neil Hobson [MVP]"
>> wrote:
>>
>>>The SCR box won't be part of the CCR cluster.
>>>
>>>If you're using CCR at your production site, implementing a standby
>>>cluster
>>>at the remote site would probably make the most sense since, if you run
>>>from
>>>the standby site for any length of time, I presume you'd want similar
>>>high
>>>availability for the duration. So for the standby cluster, you install
>>>only
>>>the passive clustered mailbox server role on each node and then enable
>>>SCR
>>>to one of them from the source CCR. When the time comes, you recover the
>>>CMS onto the node you enabled for SCR and also seed to the other passive
>>>node to bring up a new CCR environment at the standby site. This also
>>>assumes you have AD, DNS, Hub Transport, CAS, etc, etc at the standby
>>>site.
>
>
date: Tue, 15 Jan 2008 12:51:54 -0800
author: John Fullbright fjohn@donotspamnetappdotcom
|
|
