burnsynêç³b±Ë¬²*²hœ®‹(~×(

Environment consists of 2 front end servers, 2 bridgeheads, and 3 backends on
a 4 node cluster.
Began receiving MAPI logon failures and Mail Flow Receiver scripts failures
last week. Ran the Exchange MP config Wizard to remove this server from all
Mailbox Availability/ Mail Flow monitoring, and still receive the alerts.
Tried uninstalling/reinstalling the agent, still get the errors. Funny thing
is, even with the agent uninstalled, the Wizard still finds a way to contact
the non-existent agent when you select the "verify connectivity" flag....

Confused.....any help is appreciated.

blake mengotto
•©1éà¢Ûhv+ºË¢{&‰Ê貇ír‰

Where are you getting these alerts? On the cluster?

I don't have any 4way clusters...However, check out your mom admin console.
Go to all agent managed machines and verify the nodes for the one cluster are
not listed. Go to Windows Server Cluster Computers and check that the
virtual exchange name is not being managed.

After you do all this, and knowing the agent has been uninstalled...go to
the registry on each node of that cluster and blow away HKLM\Software\Mission
Critical (and all keys under it), then go to HKLM\Software\Microsoft\ and
look for:

Exchange MOM
ExMPLS

Delete these two keys.

Remove Mission Critical temp folder from: C:\windowsnt\temp

Remove Mission Critical folder from: all users\application
data\microsoft\microsoft operations manager (would be just all
users\application data\mission critical if mom 2000).

Make sure the other back end servers are not set up to try and deliver mail
to this cluster that no longer has an agent. You can run the exconfigwizard
again to verify the mailflow matrix.

burnsynêç³b±Ë¬²*²hœ®‹(~×(

OK...here's the thing...I still need to monitor this server, just not mail
flow as it is a public folder server.
I uninstalled the 4 physical agents, stopped managing and deleted all
clusters, wacked the registry keys and folders as specified in the previous
post. I then ran the config wizard to make sure that MOM could not contact
the server in question. Once that was confirmed, I reinstalled all the
agents, ran attribute discovery, and MOM picked up the the clusters, Selected
"start managing". Received the "please run the Exchange MP Wizard" message
for 2 of the 3 backends, but not for the server in question. Before I even
ran the wizard, I started receiving the MAPI logon errors again.
Frustration is setting in....

blake mengotto
•©1éà¢Ûhv+ºË¢{&‰Ê貇ír‰

Can you send me the following to: mengotto@(nospam)hotmail.com (remove the
(nospam).

Registry export of Mission Critical from the Exchange servers which are
causing you an issue (each node if clustered).

The roles they play within Exchange
(Front-end/Back-end/Mailbox/PubFolder/Etc).

If clustered, the name of the virtual exchange server, and the names of the
cluster physical nodes.

If possible Export the following event views to CSV and send them as well:

MS Exchange Server>Exchange 2003>Exchange Health Monitoring>Mapi Logon
Performance>MAPI logon warning events

MS Exchange Server>Exchange 2003>Server Configuration and
Security>Configuration Warning Events>Mailbox Stores with no agent mailbox

If the second view displays nothing, like mine does, that's because it is
looking for event id 9953, but the alert generated with this view name is
9024 and 9023. So if you can search for those alerts or events, then send
that list as well.

Please also send the mailbox access account and it's membership (should be
Exchange Org View admin at the least). Then send me some screenshots of a
mom mailbox account (using advanced view in ADUC (dsa.msc)) showing me the
mailbox access configuration (self associated with external account,
momaccess account full access to the mailbox, and on the security tab the
momaccess account should have send as and recieve as rights). Also do a
search in ADUC for the server that has no mailboxes, but is throwing up a
mapi logon error. Screen shot the results. Then go to that server, open
each store and look for an orphaned momtest mailbox. Will be a red X
ServernameMOM, ServernameMOM1, etc.

If you find the orphaned mailboxes, I would execute a purge on them and get
them off your store, that might stop the failed mapi logons. Anyway, send
this information and I can give you more suggestions in private.

burnsynêç³b±Ë¬²*²hœ®‹(~×(

Thanks for your efforts Blake.......how's this for an update:
Exchange Admin added disk to the cluster last night, brought the cluster
resource offline and then online with a failover and fail back......at that
point the MAPI logon magically begins succeeding again.......started seeing
the issue last time there was a failover, wondering if MOM didn't handle it
very well......I am continuing to investigate......

exchange2003!jxÛm7v+ºË¢{&‰Ê貇ír‰

Yesterday, we ran the updated tool against none clustered exchange servers.
We pre-created the <servername>MOM mailboxes.

Everything was running successfully.

Today, we ran the Configuration Wizard again all the servers (including the
Exchange 2003 Cluster) and all the <servername>MOM mailboxes associated with
the cluster were deleted.

Does this tool have a problem associating pre-created mailboxes with the
Exchange 2003 Virtual servernames?

Thank you,
Exchange2003