Windows 2003 clustering issue with Exchange 2007 (0/1)   

Hi

We have a 2 node Windows 2003 cluster ( MNS quorum ) running Exchange
2007.  I have noticed that the "Cluster Group" failover times vary on
each node, one takes a few seconds, the other takes approx 1 minute
and generated a "windows write delay" error as below.

I have the MNS shared out successfully on the Hub Transport server.

Can anyone advise how to solve this.

Thanks

Nick Thompson

date: Sat, 19 Jan 2008 15:20:59 +0000   author:   Nick

Re: Windows 2003 clustering issue with Exchange 2007 (0/1)   

"Nick"  wrote in message
news:3v44p39v4uoip9udcsstoq9p9bdbonmoge@4ax.com...
> Hi
>
> We have a 2 node Windows 2003 cluster ( MNS quorum ) running Exchange
> 2007.  I have noticed that the "Cluster Group" failover times vary on
> each node, one takes a few seconds, the other takes approx 1 minute
> and generated a "windows write delay" error as below.
>

Please post the error, although there should not be a write delay error, as
you cluster group should not contain any disk resources
 rgds,
edwin.

date: Mon, 21 Jan 2008 10:15:28 -0000   author:   Edwin vMierlo [MVP]

Re: Windows 2003 clustering issue with Exchange 2007 (0/1)   

Do you only get the one write delay? Or is this continuous?

Oliver

date: Mon, 21 Jan 2008 13:35:05 -0000   author:   Oliver Moazzezi [MVP]

Re: Windows 2003 clustering issue with Exchange 2007 (0/1)   

Hi Oliver

The write delay error only seems to happen when one of the nodes is
rebooted, we havent had any issues with the database disks. Just want
to make sure the cluster and Exchange is good before migrating user.

Thanks

Nick



On Mon, 21 Jan 2008 13:35:05 -0000, "Oliver Moazzezi [MVP]"
 wrote:

>
>Do you only get the one write delay? Or is this continuous?
>
>Oliver 
>

date: Mon, 21 Jan 2008 17:46:27 +0000   author:   Nick

Re: Windows 2003 clustering issue with Exchange 2007 (0/1)   

Hi Edwin

I have already posted the errors. It is an event ID 50 error, the
error seems to appear when either of the nodes are rebooted.

I want to make sure that the server is good before migrating users.

Thanks

Nick


On Mon, 21 Jan 2008 10:15:28 -0000, "Edwin vMierlo [MVP]"
 wrote:

>
>"Nick"  wrote in message
>news:3v44p39v4uoip9udcsstoq9p9bdbonmoge@4ax.com...
>> Hi
>>
>> We have a 2 node Windows 2003 cluster ( MNS quorum ) running Exchange
>> 2007.  I have noticed that the "Cluster Group" failover times vary on
>> each node, one takes a few seconds, the other takes approx 1 minute
>> and generated a "windows write delay" error as below.
>>
>
>Please post the error, although there should not be a write delay error, as
>you cluster group should not contain any disk resources
> rgds,
>edwin.
>

date: Mon, 21 Jan 2008 17:49:33 +0000   author:   Nick

Re: Windows 2003 clustering issue with Exchange 2007 (0/1)   

The only time i've seen write delays has been in the lab. The cluster was a 
VM and it was a disk configuration issue within ESX.

I would call PSS, or revert your updates (did the version of storport.sys 
change?) and see if the issue is removed.

Oliver

date: Mon, 21 Jan 2008 18:35:59 -0000   author:   Oliver Moazzezi [MVP]

Re: Windows 2003 clustering issue with Exchange 2007 (0/1)   

Thanks for the reply. The issues appear to be with crs.log which I
presume is on the HUb Transport share for the MNS quorum. When a node
is rebooted you get a popup message saying write delay error to
crs.log.

The cluster does failover successfully and Exchange does work fine.

I havent had any issues with the data. I just want to be sure
everything is good.

We have had this issue for some time.

Nick

On Mon, 21 Jan 2008 18:35:59 -0000, "Oliver Moazzezi [MVP]"
 wrote:

>The only time i've seen write delays has been in the lab. The cluster was a 
>VM and it was a disk configuration issue within ESX.
>
>I would call PSS, or revert your updates (did the version of storport.sys 
>change?) and see if the issue is removed.
>
>Oliver
>

date: Mon, 21 Jan 2008 18:58:13 +0000   author:   Nick

Re: Windows 2003 clustering issue with Exchange 2007 (0/1)   

Nick,

didn't see those until now.

As you can see this is indeed something with your network/share, as the
Event ID 50 complains about \Device\LanmanReDirector. Sot this is a Delayed
Write Failed, but not on a local disk.

Can you check that in your cluster, for both (!) Private and Public
Networks, in TCP/IP you have enabled "Client for Microsoft Network"

thanks,
Edwin.


"Nick"  wrote in message
news:5lm9p3p3409tlov8n2jr7c0unljj77119q@4ax.com...
> Hi Edwin
>
> I have already posted the errors. It is an event ID 50 error, the
> error seems to appear when either of the nodes are rebooted.
>
> I want to make sure that the server is good before migrating users.
>
> Thanks
>
> Nick
>
>
> On Mon, 21 Jan 2008 10:15:28 -0000, "Edwin vMierlo [MVP]"
>  wrote:
>
> >
> >"Nick"  wrote in message
> >news:3v44p39v4uoip9udcsstoq9p9bdbonmoge@4ax.com...
> >> Hi
> >>
> >> We have a 2 node Windows 2003 cluster ( MNS quorum ) running Exchange
> >> 2007.  I have noticed that the "Cluster Group" failover times vary on
> >> each node, one takes a few seconds, the other takes approx 1 minute
> >> and generated a "windows write delay" error as below.
> >>
> >
> >Please post the error, although there should not be a write delay error,
as
> >you cluster group should not contain any disk resources
> > rgds,
> >edwin.
> >

date: Tue, 22 Jan 2008 09:12:33 -0000   author:   Edwin vMierlo [MVP]