|
|
|
date: Tue, 4 Dec 2007 00:41:12 -0700,
group: microsoft.public.exchange.clustering
back
Re: SCC with SCR or CCR Clusters
You have some options here, but there are some restrictions and requirements
to be aware of when designing your solution. I assume that because you're
using SCC and you want to introduce SCR that you have a business need for
high availability with both service and data redundancy.
With just SCC (no SCR yet) you can achieve in-datacenter service
availability, or add on third-party replication and achieve site resilience
by building a stretch SCC. In SP1, now that we have introduced SCR, you
have more flexible deployment options.
However, you'll find that CCR is preferable to SCC for several reasons:
1. It is less complex to setup and manage than SCC.
2. It is less expensive than SCC because with CCR you don't have to buy
certified cluster solutions (or geographically dispersed cluster solutions).
You can use any hardware on the Windows Server HCL, even when stretching the
cluster between two data centers.
3. It provides you with both service redundancy and data redundancy. SCC
provides you only with service redundancy, but not data redundancy. In
fact, we specifically call this model a 'single copy cluster' to emphasize
the fact that you only have a single copy of your data within the cluster.
With CCR you have two copies of your data in the cluster.
To get data redundancy with SCC, you can implement a very small data loss
solution (SCR) or a zero data loss solution (third-party replication) on top
of SCC.
Moving forward you'll want to plan accordingly. You cannot mix Windows 2003
and Windows 2008 in the same cluster, and when using SCR, the source OS and
target OSes must match. You cannot, for example, replication from Windows
2003 to Windows 2008.
Hope this helps.
--
Regards,
Scott Schnoll
Microsoft Corporation
This posting is provided "AS IS" with no warranties, and confers no
rights. Please do not send email directly to this alias. This alias is for
newsgroup purposes only.
"Del" wrote in message
news:OG1lWkkNIHA.4272@TK2MSFTNGP05.phx.gbl...
> Customer now has X7 SCC Clusters:
>
> Customer now wants with (SP1) SCR targets with Replication to Remote Data
> Center.....(2,000 Miles)
>
> Would it better to deploy CCR clusters and eventually with Win2008 have
> CCR replication to the Remote DataCenter?
> Can W2008 X7SP1 join a W2003 X7SP1 CCR Cluster?
>
> Thanks in advance
>
date: Tue, 4 Dec 2007 08:23:41 -0800
author: Scott Schnoll [MSFT]
Re: SCC with SCR or CCR Clusters
"John Fullbright" <fjohn@donotspamnetappdotcom> wrote in message
news:eJnDHspNIHA.4712@TK2MSFTNGP04.phx.gbl...
> If you do CCR within a given site only, then yes. If you do CCR only
> across sites, then you only cover DR, not HA.
I disagree, but we may just be talking semantics. In a CCR enironment, the
key is that the email is available on the other node in the event of a
single node failure and that it is available pretty quickly. That is what HA
is all about.
But the issue with doing CCR across physical sites is taht the nodes must be
members of the same AD site, so that would mean an odd mapping of subnets to
sites. Also, you have to be really concerned with the placement of the FSW
so you can maintain majority for the event where you are attempting to
mitigate against the risk of a failure.
In any event, I would recommend CCR in the production environment with the
CCR being the source for SCR in the DR environment. Why? Because it makes
sense to me, and I think I have had Scott drilling it into my brain long
enough that it is sticking. <G>
--
Russ Kaufmann
MVP - Windows Server - Clustering
ClusterHelp.com, a Microsoft Certified Gold Partner
Web http://www.clusterhelp.com
Blog http://msmvps.com/clusterhelp
The next ClusterHelp classes are:
Dec 10 - 13 in Denver
Jan 18 - 31 in Denver
date: Tue, 4 Dec 2007 11:09:25 -0700
author: Russ Kaufmann [MVP]
Re: SCC with SCR or CCR Clusters
"Scott Schnoll [MSFT]" wrote in message
news:OId0KIpNIHA.1212@TK2MSFTNGP05.phx.gbl...
> However, you'll find that CCR is preferable to SCC for several reasons:
>
> 1. It is less complex to setup and manage than SCC.
> 2. It is less expensive than SCC because with CCR you don't have to buy
> certified cluster solutions (or geographically dispersed cluster
> solutions). You can use any hardware on the Windows Server HCL, even when
> stretching the cluster between two data centers.
> 3. It provides you with both service redundancy and data redundancy. SCC
> provides you only with service redundancy, but not data redundancy. In
> fact, we specifically call this model a 'single copy cluster' to emphasize
> the fact that you only have a single copy of your data within the cluster.
> With CCR you have two copies of your data in the cluster.
I think the key is point number 3. Because CCR is doing copies of
transactions (and it is not block level copying), if there are any problems
with the production database, the passive copy can be brought up to full
currency and it will not likely share the same issue as the production
database since it is not a block level copy.
> To get data redundancy with SCC, you can implement a very small data loss
> solution (SCR) or a zero data loss solution (third-party replication) on
> top of SCC.
Yep, and we can also get really close to the same level by using the
transport dumpster feature with CCR.
> Moving forward you'll want to plan accordingly. You cannot mix Windows
> 2003 and Windows 2008 in the same cluster, and when using SCR, the source
> OS and target OSes must match. You cannot, for example, replication from
> Windows 2003 to Windows 2008.
Hopefully the cluster group will have a good migration tool to ease this
move from 2003 to 2008 in the near future.
--
Russ Kaufmann
MVP - Windows Server - Clustering
ClusterHelp.com, a Microsoft Certified Gold Partner
Web http://www.clusterhelp.com
Blog http://msmvps.com/clusterhelp
The next ClusterHelp classes are:
Dec 10 - 13 in Denver
Jan 18 - 31 in Denver
date: Tue, 4 Dec 2007 11:13:58 -0700
author: Russ Kaufmann [MVP]
|
|
