Exchange Cluster - NO SMTP after failover

Nicky · Joined: 06 Aug 2007 Posts: 7

Question to all the Exchange gurus out there. I have an Active/Passive
Exchange cluster. Exchange 2000 on Windows 2000. After a failover to
the passive node, (planned or not planned) SMTP mailflow does not work.
Mail builds in the queues. Outlook is accessible but mail does not
flow. If I fail Exchange back over to the previous node SMTP works
fine. Any thoughts on this?

Archived from group: microsoft>public>exchange2000>clustering

Nicky · Joined: 06 Aug 2007 Posts: 7

Wanted to add the following, The SMTP virtual server does show the
Cluster shared IP not the physical IP of either server. Also in the
cluster administrator, the shared IP is used.

~~~~~~~~~~~~~~~~~~~~~~~~
Nicky wrote:
> Question to all the Exchange gurus out there. I have an Active/Passive
> Exchange cluster. Exchange 2000 on Windows 2000. After a failover to
> the passive node, (planned or not planned) SMTP mailflow does not work.
> Mail builds in the queues. Outlook is accessible but mail does not
> flow. If I fail Exchange back over to the previous node SMTP works
> fine. Any thoughts on this?

Leif Pedersen [MVP] · Joined: 05 Aug 2007 Posts: 193

Hi,

When a SMTP resource in a cluster sends emails it uses the physical IP
address of the node. You haven't by any chance blocked this address in your
fírewall?
If that is not the problem you can try to run smtpdiag on the node -
http://www.microsoft.com/downloads/details.aspx?familyid=bc1881c7-925d-4a29-bd42-71e8563c80a9&displaylang=en

Leif

"Nicky" wrote in message @m73g2000cwd.googlegroups.com...
> Wanted to add the following, The SMTP virtual server does show the
> Cluster shared IP not the physical IP of either server. Also in the
> cluster administrator, the shared IP is used.
>
> ~~~~~~~~~~~~~~~~~~~~~~~~
> Nicky wrote:
>> Question to all the Exchange gurus out there. I have an Active/Passive
>> Exchange cluster. Exchange 2000 on Windows 2000. After a failover to
>> the passive node, (planned or not planned) SMTP mailflow does not work.
>> Mail builds in the queues. Outlook is accessible but mail does not
>> flow. If I fail Exchange back over to the previous node SMTP works
>> fine. Any thoughts on this?
>

Nicky · Joined: 06 Aug 2007 Posts: 7

I will do that along with some other things, (AD communication, etc.)

More info:

When I fail over from the Active node to the passive node the following
happens:
No Internal mail flow from mailbox to mailbox on the same server
No internet mail flow coming or going

Messages queing in the awaiting directory lookup queue
It seems like there is not complete communication between Exchange and
AD on the 2nd node (former passive)

I have been getting the following errors in the logs:
9152. But from my research I have located the following article. The
only problem I have with this is that I am getting the error during
times that I am NOT failing over.

http://support.microsoft.com/kb/287610/en-us

Checked DNS records - good
Security on DNS records - good
Computer account-good
Computer account permissions - good

More to come..
If anyone has any ideas or suggestions please let me know. This is a
real bugger

Nicky · Joined: 06 Aug 2007 Posts: 7

I have ran an ensemble of diagnostics and testing on the cluster to no
avail. Everything appears to be normal

Things I have done so far:
Netdiag from both nodes to test network functionality
DCDiag on each DC to ensure a healthy AD
ExBPA Analyzer to detect any connectivy issues from Exchange to AD
SMTPDiag to watch the route of a manual message
DNS records look good as does the security ACL
Computer accounts look good as well as the security ACL
Telnetting works fine through Port 25 even when the node is failed over

Winroute looks good
I disabled ORF and Macafee Groupshield to make sure that there were no
conflicts and neither of those products seem to be the source of the
issue
Turned up Diagnostics logging for Transporting and DSAccess (AD
communication) and no errors are being spit out that are useful.

When Exchange is failed over to the passive node, All messages
immediately fail in the awaiting directory lookup queue that indicates
an issue talking to AD, but from all the other testing the server seems
to be communicating just fine and can find a GC and DC. So thats what
makes this even more perplexing.

Nicky · Joined: 06 Aug 2007 Posts: 7

No replies and my post has been here for a LONG time. Guess it will
help the next person who has this problem

My resolution to this:
Rebuilt the node that was having problems. Luckily this can be done in
the day so that I didnt have to work my nights doing it. Even after
uninstalling SMTP and Exchange and reinstalling it did not work. My
onlyl solution was rebuilding (OS and all) After the rebuild I failed
over and Walaa everything worked.

Exchange POP3 Server resource does'nt failover to Node B I have 2 nodes clustering setup for our Exchange environment, Node A and Node B. I created a new POP3 server resource and made both nodes the owner of it, also added System Attendant as the dependency. It works fine on Node a, its online and all, But when

Exchange Cluster SAN Move We currently have 2 exchange servers in an cluster connected to an EMC SAN. Within the next week we will be migrating to a new SAN solution. I am wondering if anyone knows the best way to go about doing this migration. I know there will

Capacity of active/passive Exchange 2000 cluster Are there any good documents on how many mailboxes and how big the information stores can get on an Exchange 2000 cluster before you get into performance issues ? I have heard that the practical limit is approx 1,500 mailboxes but are thes

Exchange 2000 Post SP3 update rollup on cluster. I have problems with the Exchange Management Service on 1 node of a W2K Cluster. This is the node on which the post sp3 update rollup is installed. Can I uninstall the Exchange 2000 Post SP3 update rollup fix without creating any more trouble? Or can I e.

Restoring exchange cluster virtual server to exchange enterp Hi, I've been searching for quite some time now for a document that can tell me the possibility of restoring an exchange cluster to an exchange enterprise server without the clustering part. At the moment we have 2 nodes, Let's name these s