Re: Microsoft Clustering on Exchange 2003

Tech-Archive recommends: Fix windows errors by optimizing your registry

10 minute failover? Unexpected failovers? You really need go back over your
hardware and software configuration, as well as your operational procedures.

The two key factors that impact failover in an MSCS exchange cluster are the
time to flush data to disk so the disk can be taken offline, and the time to
commit/replay transactions. You have a lot of control over these depending
on how you configure/operate your cluster. I've done MNS geographically
dispersed 4 node clusters with over 10,000 mailboxes, and the the observed
failover time was less than 60 seconds. The eunuchs clusters I deal with
most run on BSD variants. The observed failover time on these has been
similar, 60-90 seconds.

If you have problems configuring/operating an MSCS cluster with any degree
of reliability, how do you expect changing the clustering software to solve
your problems?



"Mike" <mjj2u2@xxxxxxxxx> wrote in message
news:1138722834.980019.145790@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
> We have been running two MS Clusters on Exchange 2003 for about a year
> now and we have never benefited from the redundancy. To the contrary we
> have actually had a significant amount of down time due to the cluster
> failing over when it shouldn't or becoming unstable. (The configuration
> is one active node and one standby server in the cluster.) The real
> kicker here is the convergence time. I was familiar with Linux clusters
> that take a few seconds to fail over, so you can imagine my surprise
> when the Exchange cluster can take up to 10 minutes to fail over! The
> MS consultant who was helping configure this cluster told us that it
> may take up to 15 minutes. In my mind this isn't high availability, it
> is rapid recovery.
>
> At this point we are looking for alternatives to the MS clustering. Has
> anyone had experience with other clustering or high availability
> solutions for Exchange 2003 that have work better then this?
> Specifically has anyone used Marathon clustering or Polyserve?
>
> The best insights come from co-engineers who actually have experience
> working with these systems and I would very much appreciate your
> insights from any experiences you have had with this.
>
> Thanks!
>
> Mike
>


.

Relevant Pages

  • Re: Why Cluster in a test Environment?
    ... If your apps are not failover aware, ... SQL Server 2000 then takes anywhere from 30 seconds to minutes before ... a cluster but not appear in a standalone server. ... A non-clustered environment would not allow accurate stress testing do ...
    (microsoft.public.sqlserver.clustering)
  • Re: Is there a best practise to implement a warm standby-solution
    ... Doing all this on a failover cluster is just adding complexity. ... I suspected that some of what I need is not provided by MSCS. ... ApplicationX is cluster aware. ...
    (microsoft.public.windows.server.clustering)
  • Re: Configure failover in seconds
    ... When a failover starts, it starts. ... add this script as a resource to your cluster ... Is this done in CLuster Administrator? ... before failover occurs from one node to another in Windows Server ...
    (microsoft.public.windows.server.clustering)
  • Re: Virtual Server and Cluster
    ... Manual shutdown of a host is a special case for failover, and there is a section in that document which describes a simple batch file which runs on shutdown that essentially stops the cluster service. ... Virtual Server per MS documentation and it is working fine. ...
    (microsoft.public.windows.server.clustering)
  • Re: Configure failover in seconds
    ... When a failover starts, it starts. ... "pause" because of your san disks who need to failover. ... add this script as a resource to your cluster ... Is this done in CLuster Administrator? ...
    (microsoft.public.windows.server.clustering)