MAPI and Test-Mailflow Rules Running on Inactive Mailbox Server Cluster Node?

[Megan Kielman]

I am running the latest Exchange 2007 MP and suddenly we started
receiving alerts from the inactive mailbox server node because some of
the MAPI and Test-Mailbox commands were failing. This particular servers
has been the inactive node for over a week. Suggestions on how I should
resolve this?

[Åke Pettersson [MSFT]]

Does the issue disappear if the cluster fails over again to the other node?

I am running the latest Exchange 2007 MP and suddenly we started receiving
alerts from the inactive mailbox server node because some of the MAPI and
Test-Mailbox commands were failing. This particular servers has been the
inactive node for over a week. Suggestions on how I should resolve this?

[Steven Halsey [MSFT]]

Which version of the Exchange 2007 MP are you running? In version 6278.12
there were fixes to the discoveries for Cluster servers, so if you are not
running this version you should upgrade then follow the instructions in the
MP Guide to clean up your environment (involves running the
remove-disabledmonitoringobject commandlet on your RMS).

If you look in the Operations Manager Console under discovered inventory and
check for the object types for all 4 of the types below, in which ones do
you see the passive physical node?
1. "Ex. Mailbox" - Appear as Monitored, physical node should not be of
this type
2. "Microsoft Exchange 2007 Mailbox Servers Installation" - Appear as
UnMonitored, physical node will appear as this type
3. "Ex. Common" - Appear as Monitored, physical node should be of this type
4. "Microsoft Exchange 2007 All Servers Installation" - Appear as
UnMonitored, physical node should be of this type
--
Steven Halsey [MSFT]

This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

I am running the latest Exchange 2007 MP and suddenly we started receiving
alerts from the inactive mailbox server node because some of the MAPI and
Test-Mailbox commands were failing. This particular servers has been the
inactive node for over a week. Suggestions on how I should resolve this?

[Megan Kielman]

Thanks for the reply. The Exchange MP I am running is 6.0.6461.0.

Here is what I have found in Discovered Inventory:

Ex. Mailbox - includes both physical nodes (one of which is passive) AND the
cluster name. Everything is monitored.

Microsoft Exchange 2007 Mailbox Servers Installation - Cluster name appears
as unmonitored. No physical nodes exist.

Ex. Common - All exchange servers (including cluster name) exist and are
monitored

Microsoft Exchange 2007 All Servers Installation - All exchange servers
(including cluster name) appear but are unmonitored


do I still need to run the cleanup you referred to?

Post by Steven Halsey [MSFT]
Which version of the Exchange 2007 MP are you running? In version 6278.12
there were fixes to the discoveries for Cluster servers, so if you are not
running this version you should upgrade then follow the instructions in the
MP Guide to clean up your environment (involves running the
remove-disabledmonitoringobject commandlet on your RMS).
If you look in the Operations Manager Console under discovered inventory and
check for the object types for all 4 of the types below, in which ones do
you see the passive physical node?
1. "Ex. Mailbox" - Appear as Monitored, physical node should not be of
this type
2. "Microsoft Exchange 2007 Mailbox Servers Installation" - Appear as
UnMonitored, physical node will appear as this type
3. "Ex. Common" - Appear as Monitored, physical node should be of this type
4. "Microsoft Exchange 2007 All Servers Installation" - Appear as
UnMonitored, physical node should be of this type
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

I am running the latest Exchange 2007 MP and suddenly we started receiving
alerts from the inactive mailbox server node because some of the MAPI and
Test-Mailbox commands were failing. This particular servers has been the
inactive node for over a week. Suggestions on how I should resolve this?

[Steven Halsey [MSFT]]

No you shouldn't need to run the commandlet. If you are still seeing this
behavior can you answer a couple of more questions?

If you look at the agentless managed under Administration, which physical
node is responsible for monitoring the virtual server? Is it the passive
node for which you are getting alerts or the active node?

If you look at the performance counters for Mapi logon under the monitoring
pane Exchange 2007, is the virtual server logging mapi connectivity latency
currently or is the data not being logged? Is the Ex ailbox instance of the
virtual server healthy or unhealthy?
--
Steven Halsey [MSFT]

This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included script samples are subject to terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
Thanks for the reply. The Exchange MP I am running is 6.0.6461.0.
Ex. Mailbox - includes both physical nodes (one of which is passive) AND the
cluster name. Everything is monitored.
Microsoft Exchange 2007 Mailbox Servers Installation - Cluster name appears
as unmonitored. No physical nodes exist.
Ex. Common - All exchange servers (including cluster name) exist and are
monitored
Microsoft Exchange 2007 All Servers Installation - All exchange servers
(including cluster name) appear but are unmonitored
do I still need to run the cleanup you referred to?

Post by Steven Halsey [MSFT]
Which version of the Exchange 2007 MP are you running? In version 6278.12
there were fixes to the discoveries for Cluster servers, so if you are not
running this version you should upgrade then follow the instructions in the
MP Guide to clean up your environment (involves running the
remove-disabledmonitoringobject commandlet on your RMS).
If you look in the Operations Manager Console under discovered inventory and
check for the object types for all 4 of the types below, in which ones do
you see the passive physical node?
1. "Ex. Mailbox" - Appear as Monitored, physical node should not be of
this type
2. "Microsoft Exchange 2007 Mailbox Servers Installation" - Appear as
UnMonitored, physical node will appear as this type
3. "Ex. Common" - Appear as Monitored, physical node should be of this type
4. "Microsoft Exchange 2007 All Servers Installation" - Appear as
UnMonitored, physical node should be of this type
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

I am running the latest Exchange 2007 MP and suddenly we started receiving
alerts from the inactive mailbox server node because some of the MAPI and
Test-Mailbox commands were failing. This particular servers has been the
inactive node for over a week. Suggestions on how I should resolve this?

[Megan Kielman]

The physical node monitoring the Exchange virtual server is the passive node!

The virtual server is listed in the Test-MAPIConnectivity Logon Latency
performance view, but there is no information being collected.

In the MOnitoring Pane->Exchange SErver 2007->State View, the Virtual Server
is healthy but the passive node has a warning below the Ex. Mailbox column.

Post by Steven Halsey [MSFT]
No you shouldn't need to run the commandlet. If you are still seeing this
behavior can you answer a couple of more questions?
If you look at the agentless managed under Administration, which physical
node is responsible for monitoring the virtual server? Is it the passive
node for which you are getting alerts or the active node?
If you look at the performance counters for Mapi logon under the monitoring
pane Exchange 2007, is the virtual server logging mapi connectivity latency
currently or is the data not being logged? Is the Ex ailbox instance of the
virtual server healthy or unhealthy?
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included script samples are subject to terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
Thanks for the reply. The Exchange MP I am running is 6.0.6461.0.
Ex. Mailbox - includes both physical nodes (one of which is passive) AND the
cluster name. Everything is monitored.
Microsoft Exchange 2007 Mailbox Servers Installation - Cluster name appears
as unmonitored. No physical nodes exist.
Ex. Common - All exchange servers (including cluster name) exist and are
monitored
Microsoft Exchange 2007 All Servers Installation - All exchange servers
(including cluster name) appear but are unmonitored
do I still need to run the cleanup you referred to?

Post by Steven Halsey [MSFT]
Which version of the Exchange 2007 MP are you running? In version 6278.12
there were fixes to the discoveries for Cluster servers, so if you are not
running this version you should upgrade then follow the instructions in the
MP Guide to clean up your environment (involves running the
remove-disabledmonitoringobject commandlet on your RMS).
If you look in the Operations Manager Console under discovered inventory and
check for the object types for all 4 of the types below, in which ones do
you see the passive physical node?
1. "Ex. Mailbox" - Appear as Monitored, physical node should not be of
this type
2. "Microsoft Exchange 2007 Mailbox Servers Installation" - Appear as
UnMonitored, physical node will appear as this type
3. "Ex. Common" - Appear as Monitored, physical node should be of this type
4. "Microsoft Exchange 2007 All Servers Installation" - Appear as
UnMonitored, physical node should be of this type
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

I am running the latest Exchange 2007 MP and suddenly we started receiving
alerts from the inactive mailbox server node because some of the MAPI and
Test-Mailbox commands were failing. This particular servers has been the
inactive node for over a week. Suggestions on how I should resolve this?

[Steven Halsey [MSFT]]

If you use Operations Manager Console, and have the Agent Responsible for
monitoring the Virtual Server switched to the active node, do the errors go
away and do the performance counters for mapi-connectivity appear for the
virtual server?
--
Steven Halsey [MSFT]

This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
The physical node monitoring the Exchange virtual server is the passive node!
The virtual server is listed in the Test-MAPIConnectivity Logon Latency
performance view, but there is no information being collected.
In the MOnitoring Pane->Exchange SErver 2007->State View, the Virtual Server
is healthy but the passive node has a warning below the Ex. Mailbox column.

Post by Steven Halsey [MSFT]
No you shouldn't need to run the commandlet. If you are still seeing this
behavior can you answer a couple of more questions?
If you look at the agentless managed under Administration, which physical
node is responsible for monitoring the virtual server? Is it the passive
node for which you are getting alerts or the active node?
If you look at the performance counters for Mapi logon under the monitoring
pane Exchange 2007, is the virtual server logging mapi connectivity latency
currently or is the data not being logged? Is the Ex ailbox instance of the
virtual server healthy or unhealthy?
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included script samples are subject to terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
Thanks for the reply. The Exchange MP I am running is 6.0.6461.0.
Ex. Mailbox - includes both physical nodes (one of which is passive)
AND
the
cluster name. Everything is monitored.
Microsoft Exchange 2007 Mailbox Servers Installation - Cluster name appears
as unmonitored. No physical nodes exist.
Ex. Common - All exchange servers (including cluster name) exist and are
monitored
Microsoft Exchange 2007 All Servers Installation - All exchange servers
(including cluster name) appear but are unmonitored
do I still need to run the cleanup you referred to?

Post by Steven Halsey [MSFT]
Which version of the Exchange 2007 MP are you running? In version 6278.12
there were fixes to the discoveries for Cluster servers, so if you are not
running this version you should upgrade then follow the instructions
in
the
MP Guide to clean up your environment (involves running the
remove-disabledmonitoringobject commandlet on your RMS).
If you look in the Operations Manager Console under discovered
inventory
and
check for the object types for all 4 of the types below, in which ones do
you see the passive physical node?
1. "Ex. Mailbox" - Appear as Monitored, physical node should not be of
this type
2. "Microsoft Exchange 2007 Mailbox Servers Installation" - Appear as
UnMonitored, physical node will appear as this type
3. "Ex. Common" - Appear as Monitored, physical node should be of
this
type
4. "Microsoft Exchange 2007 All Servers Installation" - Appear as
UnMonitored, physical node should be of this type
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

I am running the latest Exchange 2007 MP and suddenly we started receiving
alerts from the inactive mailbox server node because some of the MAPI and
Test-Mailbox commands were failing. This particular servers has been the
inactive node for over a week. Suggestions on how I should resolve this?

[Steven Halsey [MSFT]]

Are you running an SCC (shared storage) or a CCR (continuous replication)
Exchange Cluster?
What version of the Operating System is Exchange running on including the
service pack?

The passive node can appear as the host for the agentless monitoring, this
is a normal situation. There is code in the agent that should run those
commandlets from the Active node of the cluster, but for some reason this
does not seem to be happening on your server.

As a work around you can override the commandlets that the Synthetic
Transactions are running so that they will run successfully.

1. Go to Authoring->Managmenet Pack Objects->Rules
2. Set the Scope as "Microsoft Exchange 2007 Mailbox Servers Installation"
3. Search for the rule "Execute: Test-MAPIConectivity diagnostic Commandlet.
(Report Collection)"
4. Override the rule for the specific instance of "Microsoft Exchange 2007
Mailbox Servers Installation" that matches your Exchange Virtual server
5. Find the cmdletCommandParameter and replace [Virtual Server Netbios Name]
and then use the following line instead of the default:
"Test-MAPIConnectivity -server:[Virtual Server Netbios
Name] -MonitoringContext:$true -PerConnectionTimeout:60 -AllConnectionsTimeout:90"
6. Click OK to save the changes
7. Search for the rule "Execute: Test-Mailflow (Remote) diagnostic cmdlet.
(Report Collection)"
8. Override the rule for the specific instance of "Microsoft Exchange 2007
Mailbox Servers Installation" that matches your Exchange Virtual server
9. Find the cmdletCommandParameter and replace [Virtual Server Netbios Name]
and then use the following line instead of the default:
"Test-Mailflow -identity:[Virtual Server Netbios
Name] -AutoDiscoverTargetMailboxServer:$true -ErrorLatency:180 -MonitoringContext:$true"
10. Click OK to save the changes
11. Search for the rule "Execute: Test-Mailflow (Local) diagnostic cmdlet.
(Report Collection)"
12. Override the rule for the specific instance of "Microsoft Exchange 2007
Mailbox Servers Installation" that matches your Exchange Virtual server
13. Find the cmdletCommandParameter and replace [Virtual Server Netbios
Name] and then use the following line instead of the default:
"Test-Mailflow -identity:[Virtual Server Netbios
Name] -ErrorLatency:60 -MonitoringContext:$true"
14. Click OK to save the changes
--
Steven Halsey [MSFT]

This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Steven Halsey [MSFT]
If you use Operations Manager Console, and have the Agent Responsible for
monitoring the Virtual Server switched to the active node, do the errors
go away and do the performance counters for mapi-connectivity appear for
the virtual server?
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
The physical node monitoring the Exchange virtual server is the passive node!
The virtual server is listed in the Test-MAPIConnectivity Logon Latency
performance view, but there is no information being collected.
In the MOnitoring Pane->Exchange SErver 2007->State View, the Virtual Server
is healthy but the passive node has a warning below the Ex. Mailbox column.

Post by Steven Halsey [MSFT]
No you shouldn't need to run the commandlet. If you are still seeing this
behavior can you answer a couple of more questions?
If you look at the agentless managed under Administration, which physical
node is responsible for monitoring the virtual server? Is it the passive
node for which you are getting alerts or the active node?
If you look at the performance counters for Mapi logon under the monitoring
pane Exchange 2007, is the virtual server logging mapi connectivity latency
currently or is the data not being logged? Is the Ex ailbox instance of the
virtual server healthy or unhealthy?
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included script samples are subject to terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
Thanks for the reply. The Exchange MP I am running is 6.0.6461.0.
Ex. Mailbox - includes both physical nodes (one of which is passive)
AND
the
cluster name. Everything is monitored.
Microsoft Exchange 2007 Mailbox Servers Installation - Cluster name appears
as unmonitored. No physical nodes exist.
Ex. Common - All exchange servers (including cluster name) exist and are
monitored
Microsoft Exchange 2007 All Servers Installation - All exchange servers
(including cluster name) appear but are unmonitored
do I still need to run the cleanup you referred to?

Post by Steven Halsey [MSFT]
Which version of the Exchange 2007 MP are you running? In version 6278.12
there were fixes to the discoveries for Cluster servers, so if you
are
not
running this version you should upgrade then follow the instructions
in
the
MP Guide to clean up your environment (involves running the
remove-disabledmonitoringobject commandlet on your RMS).
If you look in the Operations Manager Console under discovered
inventory
and
check for the object types for all 4 of the types below, in which ones do
you see the passive physical node?
1. "Ex. Mailbox" - Appear as Monitored, physical node should not be of
this type
2. "Microsoft Exchange 2007 Mailbox Servers Installation" - Appear as
UnMonitored, physical node will appear as this type
3. "Ex. Common" - Appear as Monitored, physical node should be of
this
type
4. "Microsoft Exchange 2007 All Servers Installation" - Appear as
UnMonitored, physical node should be of this type
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

I am running the latest Exchange 2007 MP and suddenly we started receiving
alerts from the inactive mailbox server node because some of the
MAPI
and
Test-Mailbox commands were failing. This particular servers has been the
inactive node for over a week. Suggestions on how I should resolve this?

[Megan Kielman]

We are running SCC and the OS is Windows 2003 x64 SP2 R2 Enterprise.

I have made the below changes. Can you explain why, if they target the
virtual node anyway, they weren't working?

Post by Steven Halsey [MSFT]
Are you running an SCC (shared storage) or a CCR (continuous replication)
Exchange Cluster?
What version of the Operating System is Exchange running on including the
service pack?
The passive node can appear as the host for the agentless monitoring, this
is a normal situation. There is code in the agent that should run those
commandlets from the Active node of the cluster, but for some reason this
does not seem to be happening on your server.
As a work around you can override the commandlets that the Synthetic
Transactions are running so that they will run successfully.
1. Go to Authoring->Managmenet Pack Objects->Rules
2. Set the Scope as "Microsoft Exchange 2007 Mailbox Servers Installation"
3. Search for the rule "Execute: Test-MAPIConectivity diagnostic Commandlet.
(Report Collection)"
4. Override the rule for the specific instance of "Microsoft Exchange 2007
Mailbox Servers Installation" that matches your Exchange Virtual server
5. Find the cmdletCommandParameter and replace [Virtual Server Netbios Name]
"Test-MAPIConnectivity -server:[Virtual Server Netbios
Name] -MonitoringContext:$true -PerConnectionTimeout:60 -AllConnectionsTimeout:90"
6. Click OK to save the changes
7. Search for the rule "Execute: Test-Mailflow (Remote) diagnostic cmdlet.
(Report Collection)"
8. Override the rule for the specific instance of "Microsoft Exchange 2007
Mailbox Servers Installation" that matches your Exchange Virtual server
9. Find the cmdletCommandParameter and replace [Virtual Server Netbios Name]
"Test-Mailflow -identity:[Virtual Server Netbios
Name] -AutoDiscoverTargetMailboxServer:$true -ErrorLatency:180 -MonitoringContext:$true"
10. Click OK to save the changes
11. Search for the rule "Execute: Test-Mailflow (Local) diagnostic cmdlet.
(Report Collection)"
12. Override the rule for the specific instance of "Microsoft Exchange 2007
Mailbox Servers Installation" that matches your Exchange Virtual server
13. Find the cmdletCommandParameter and replace [Virtual Server Netbios
"Test-Mailflow -identity:[Virtual Server Netbios
Name] -ErrorLatency:60 -MonitoringContext:$true"
14. Click OK to save the changes
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Steven Halsey [MSFT]
If you use Operations Manager Console, and have the Agent Responsible for
monitoring the Virtual Server switched to the active node, do the errors
go away and do the performance counters for mapi-connectivity appear for
the virtual server?
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
The physical node monitoring the Exchange virtual server is the passive node!
The virtual server is listed in the Test-MAPIConnectivity Logon Latency
performance view, but there is no information being collected.
In the MOnitoring Pane->Exchange SErver 2007->State View, the Virtual Server
is healthy but the passive node has a warning below the Ex. Mailbox column.

Post by Steven Halsey [MSFT]
No you shouldn't need to run the commandlet. If you are still seeing this
behavior can you answer a couple of more questions?
If you look at the agentless managed under Administration, which physical
node is responsible for monitoring the virtual server? Is it the passive
node for which you are getting alerts or the active node?
If you look at the performance counters for Mapi logon under the monitoring
pane Exchange 2007, is the virtual server logging mapi connectivity latency
currently or is the data not being logged? Is the Ex ailbox instance of the
virtual server healthy or unhealthy?
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included script samples are subject to terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
Thanks for the reply. The Exchange MP I am running is 6.0.6461.0.
Ex. Mailbox - includes both physical nodes (one of which is passive)
AND
the
cluster name. Everything is monitored.
Microsoft Exchange 2007 Mailbox Servers Installation - Cluster name appears
as unmonitored. No physical nodes exist.
Ex. Common - All exchange servers (including cluster name) exist and are
monitored
Microsoft Exchange 2007 All Servers Installation - All exchange servers
(including cluster name) appear but are unmonitored
do I still need to run the cleanup you referred to?

Post by Steven Halsey [MSFT]
Which version of the Exchange 2007 MP are you running? In version 6278.12
there were fixes to the discoveries for Cluster servers, so if you
are
not
running this version you should upgrade then follow the instructions
in
the
MP Guide to clean up your environment (involves running the
remove-disabledmonitoringobject commandlet on your RMS).
If you look in the Operations Manager Console under discovered
inventory
and
check for the object types for all 4 of the types below, in which ones do
you see the passive physical node?
1. "Ex. Mailbox" - Appear as Monitored, physical node should not be of
this type
2. "Microsoft Exchange 2007 Mailbox Servers Installation" - Appear as
UnMonitored, physical node will appear as this type
3. "Ex. Common" - Appear as Monitored, physical node should be of
this
type
4. "Microsoft Exchange 2007 All Servers Installation" - Appear as
UnMonitored, physical node should be of this type
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

I am running the latest Exchange 2007 MP and suddenly we started receiving
alerts from the inactive mailbox server node because some of the
MAPI
and
Test-Mailbox commands were failing. This particular servers has been the
inactive node for over a week. Suggestions on how I should resolve this?

[Steven Halsey [MSFT]]

The reason the workflows are not working correctly has to do with how
agentless monitoring has to be written. Basically the work flow is executed
on an agent machine, and happens from that machine. To run correctly the
workflow needs to be directed to the correct remote machine, similar to
doing an eventvwr.msc \\RemoteMachine.

These workflows in the Exchange MP are not getting that pointer to the
remote machine. Essentially what is happening is Exchange Management shell
is opened on the passive node and then run test-mapiconnectivity, the
commandlet attempts to determine what virtual machine to run against and
can't so you see those errors. The workaround I gave below is simply
pointing the Exchange commandlets which machines to execute against.

As I explained in another post of this thread, there is logic in the agents
of a cluster physical nodes to recognize the virtual server and run from the
correct physical node. We are looking into what might be going on with
Exchange servers around this area.
--
Steven Halsey [MSFT]

This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
We are running SCC and the OS is Windows 2003 x64 SP2 R2 Enterprise.
I have made the below changes. Can you explain why, if they target the
virtual node anyway, they weren't working?

Post by Steven Halsey [MSFT]
Are you running an SCC (shared storage) or a CCR (continuous replication)
Exchange Cluster?
What version of the Operating System is Exchange running on including the
service pack?
The passive node can appear as the host for the agentless monitoring, this
is a normal situation. There is code in the agent that should run those
commandlets from the Active node of the cluster, but for some reason this
does not seem to be happening on your server.
As a work around you can override the commandlets that the Synthetic
Transactions are running so that they will run successfully.
1. Go to Authoring->Managmenet Pack Objects->Rules
2. Set the Scope as "Microsoft Exchange 2007 Mailbox Servers
Installation"
3. Search for the rule "Execute: Test-MAPIConectivity diagnostic Commandlet.
(Report Collection)"
4. Override the rule for the specific instance of "Microsoft Exchange 2007
Mailbox Servers Installation" that matches your Exchange Virtual server
5. Find the cmdletCommandParameter and replace [Virtual Server Netbios Name]
"Test-MAPIConnectivity -server:[Virtual Server Netbios
Name] -MonitoringContext:$true -PerConnectionTimeout:60 -AllConnectionsTimeout:90"
6. Click OK to save the changes
7. Search for the rule "Execute: Test-Mailflow (Remote) diagnostic cmdlet.
(Report Collection)"
8. Override the rule for the specific instance of "Microsoft Exchange 2007
Mailbox Servers Installation" that matches your Exchange Virtual server
9. Find the cmdletCommandParameter and replace [Virtual Server Netbios Name]
"Test-Mailflow -identity:[Virtual Server Netbios
Name] -AutoDiscoverTargetMailboxServer:$true -ErrorLatency:180 -MonitoringContext:$true"
10. Click OK to save the changes
11. Search for the rule "Execute: Test-Mailflow (Local) diagnostic cmdlet.
(Report Collection)"
12. Override the rule for the specific instance of "Microsoft Exchange 2007
Mailbox Servers Installation" that matches your Exchange Virtual server
13. Find the cmdletCommandParameter and replace [Virtual Server Netbios
"Test-Mailflow -identity:[Virtual Server Netbios
Name] -ErrorLatency:60 -MonitoringContext:$true"
14. Click OK to save the changes
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Steven Halsey [MSFT]
If you use Operations Manager Console, and have the Agent Responsible for
monitoring the Virtual Server switched to the active node, do the errors
go away and do the performance counters for mapi-connectivity appear for
the virtual server?
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
The physical node monitoring the Exchange virtual server is the
passive
node!
The virtual server is listed in the Test-MAPIConnectivity Logon Latency
performance view, but there is no information being collected.
In the MOnitoring Pane->Exchange SErver 2007->State View, the Virtual Server
is healthy but the passive node has a warning below the Ex. Mailbox column.

Post by Steven Halsey [MSFT]
No you shouldn't need to run the commandlet. If you are still seeing this
behavior can you answer a couple of more questions?
If you look at the agentless managed under Administration, which physical
node is responsible for monitoring the virtual server? Is it the passive
node for which you are getting alerts or the active node?
If you look at the performance counters for Mapi logon under the monitoring
pane Exchange 2007, is the virtual server logging mapi connectivity latency
currently or is the data not being logged? Is the Ex ailbox instance
of
the
virtual server healthy or unhealthy?
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included script samples are subject to terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
Thanks for the reply. The Exchange MP I am running is 6.0.6461.0.
Ex. Mailbox - includes both physical nodes (one of which is passive)
AND
the
cluster name. Everything is monitored.
Microsoft Exchange 2007 Mailbox Servers Installation - Cluster name appears
as unmonitored. No physical nodes exist.
Ex. Common - All exchange servers (including cluster name) exist
and
are
monitored
Microsoft Exchange 2007 All Servers Installation - All exchange servers
(including cluster name) appear but are unmonitored
do I still need to run the cleanup you referred to?

Post by Steven Halsey [MSFT]
Which version of the Exchange 2007 MP are you running? In version 6278.12
there were fixes to the discoveries for Cluster servers, so if you
are
not
running this version you should upgrade then follow the instructions
in
the
MP Guide to clean up your environment (involves running the
remove-disabledmonitoringobject commandlet on your RMS).
If you look in the Operations Manager Console under discovered
inventory
and
check for the object types for all 4 of the types below, in which ones do
you see the passive physical node?
1. "Ex. Mailbox" - Appear as Monitored, physical node should not
be
of
this type
2. "Microsoft Exchange 2007 Mailbox Servers Installation" -
Appear
as
UnMonitored, physical node will appear as this type
3. "Ex. Common" - Appear as Monitored, physical node should be of
this
type
4. "Microsoft Exchange 2007 All Servers Installation" - Appear as
UnMonitored, physical node should be of this type
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers
no
rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
I am running the latest Exchange 2007 MP and suddenly we started
receiving
alerts from the inactive mailbox server node because some of the
MAPI
and
Test-Mailbox commands were failing. This particular servers has
been
the
inactive node for over a week. Suggestions on how I should
resolve
this?

[Megan Kielman]

HOw do I switch which host is in charge? How do I prevent this from happening
again?

Post by Steven Halsey [MSFT]
If you use Operations Manager Console, and have the Agent Responsible for
monitoring the Virtual Server switched to the active node, do the errors go
away and do the performance counters for mapi-connectivity appear for the
virtual server?
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
The physical node monitoring the Exchange virtual server is the passive node!
The virtual server is listed in the Test-MAPIConnectivity Logon Latency
performance view, but there is no information being collected.
In the MOnitoring Pane->Exchange SErver 2007->State View, the Virtual Server
is healthy but the passive node has a warning below the Ex. Mailbox column.

Post by Steven Halsey [MSFT]
No you shouldn't need to run the commandlet. If you are still seeing this
behavior can you answer a couple of more questions?
If you look at the agentless managed under Administration, which physical
node is responsible for monitoring the virtual server? Is it the passive
node for which you are getting alerts or the active node?
If you look at the performance counters for Mapi logon under the monitoring
pane Exchange 2007, is the virtual server logging mapi connectivity latency
currently or is the data not being logged? Is the Ex ailbox instance of the
virtual server healthy or unhealthy?
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included script samples are subject to terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
Thanks for the reply. The Exchange MP I am running is 6.0.6461.0.
Ex. Mailbox - includes both physical nodes (one of which is passive)
AND
the
cluster name. Everything is monitored.
Microsoft Exchange 2007 Mailbox Servers Installation - Cluster name appears
as unmonitored. No physical nodes exist.
Ex. Common - All exchange servers (including cluster name) exist and are
monitored
Microsoft Exchange 2007 All Servers Installation - All exchange servers
(including cluster name) appear but are unmonitored
do I still need to run the cleanup you referred to?

Post by Steven Halsey [MSFT]
Which version of the Exchange 2007 MP are you running? In version 6278.12
there were fixes to the discoveries for Cluster servers, so if you are not
running this version you should upgrade then follow the instructions
in
the
MP Guide to clean up your environment (involves running the
remove-disabledmonitoringobject commandlet on your RMS).
If you look in the Operations Manager Console under discovered
inventory
and
check for the object types for all 4 of the types below, in which ones do
you see the passive physical node?
1. "Ex. Mailbox" - Appear as Monitored, physical node should not be of
this type
2. "Microsoft Exchange 2007 Mailbox Servers Installation" - Appear as
UnMonitored, physical node will appear as this type
3. "Ex. Common" - Appear as Monitored, physical node should be of
this
type
4. "Microsoft Exchange 2007 All Servers Installation" - Appear as
UnMonitored, physical node should be of this type
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

I am running the latest Exchange 2007 MP and suddenly we started receiving
alerts from the inactive mailbox server node because some of the MAPI and
Test-Mailbox commands were failing. This particular servers has been the
inactive node for over a week. Suggestions on how I should resolve this?

[Steven Halsey [MSFT]]

You can switch the responsible agent in Operations Manager Console.
1. Go to administration Pane->Device Management->Agentless Managed
2. Right click on the virtual server name and select Change Proxy Agent
3. Select the new proxy agent and hit OK

It is normal for the cluster virtual server to be listed as agentless
monitoring as the passive node, the node that first discovers the cluster
virtual server becomes the host agent always. The way Operations Manager
works is only 1 agent can be responsible for the virtual server. What
happens is the agents on the physical nodes recognize that they are part of
a cluster and know which virtual servers they are running, and then they run
the appropriate monitoring. For some reason, on your agents or inside of
the Exchange 2007 Converted MP this logic isn't happening correctly.

If you could tell me the version of the Operating System on the cluster
(including SP), the type of Exchange cluster (CCR or SCC), and which of the
following Management Packs and the version are imported to the Management
Group it might help to figure out where the problem lies (Exchange 2007
Converted MP, Exchange 2007 Native MP Beta, Microsoft Windows Server Cluster
MP, Windows 2003 MP, and Windows 2008 MP).

I added another post where you can set the workflows that are currently
failing so they can be successful, and you won't see alerts about the
failing mapi logon or mail flow.
--
Steven Halsey [MSFT]

This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
HOw do I switch which host is in charge? How do I prevent this from happening
again?

Post by Steven Halsey [MSFT]
If you use Operations Manager Console, and have the Agent Responsible for
monitoring the Virtual Server switched to the active node, do the errors go
away and do the performance counters for mapi-connectivity appear for the
virtual server?
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
The physical node monitoring the Exchange virtual server is the passive node!
The virtual server is listed in the Test-MAPIConnectivity Logon Latency
performance view, but there is no information being collected.
In the MOnitoring Pane->Exchange SErver 2007->State View, the Virtual Server
is healthy but the passive node has a warning below the Ex. Mailbox column.

Post by Steven Halsey [MSFT]
No you shouldn't need to run the commandlet. If you are still seeing this
behavior can you answer a couple of more questions?
If you look at the agentless managed under Administration, which physical
node is responsible for monitoring the virtual server? Is it the passive
node for which you are getting alerts or the active node?
If you look at the performance counters for Mapi logon under the monitoring
pane Exchange 2007, is the virtual server logging mapi connectivity latency
currently or is the data not being logged? Is the Ex ailbox instance
of
the
virtual server healthy or unhealthy?
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included script samples are subject to terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
Thanks for the reply. The Exchange MP I am running is 6.0.6461.0.
Ex. Mailbox - includes both physical nodes (one of which is passive)
AND
the
cluster name. Everything is monitored.
Microsoft Exchange 2007 Mailbox Servers Installation - Cluster name appears
as unmonitored. No physical nodes exist.
Ex. Common - All exchange servers (including cluster name) exist and are
monitored
Microsoft Exchange 2007 All Servers Installation - All exchange servers
(including cluster name) appear but are unmonitored
do I still need to run the cleanup you referred to?

Post by Steven Halsey [MSFT]
Which version of the Exchange 2007 MP are you running? In version 6278.12
there were fixes to the discoveries for Cluster servers, so if you
are
not
running this version you should upgrade then follow the
instructions
in
the
MP Guide to clean up your environment (involves running the
remove-disabledmonitoringobject commandlet on your RMS).
If you look in the Operations Manager Console under discovered
inventory
and
check for the object types for all 4 of the types below, in which
ones
do
you see the passive physical node?
1. "Ex. Mailbox" - Appear as Monitored, physical node should not
be
of
this type
2. "Microsoft Exchange 2007 Mailbox Servers Installation" - Appear as
UnMonitored, physical node will appear as this type
3. "Ex. Common" - Appear as Monitored, physical node should be of
this
type
4. "Microsoft Exchange 2007 All Servers Installation" - Appear as
UnMonitored, physical node should be of this type
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

I am running the latest Exchange 2007 MP and suddenly we started receiving
alerts from the inactive mailbox server node because some of the
MAPI
and
Test-Mailbox commands were failing. This particular servers has
been
the
inactive node for over a week. Suggestions on how I should resolve this?

[Megan Kielman]

Thanks for the information! I provided you with OS details and here are MP
details:

Microsoft.Exchange.2007 (Converted) v6.0.6461

Windows 2003 Cluster Management Monitoring v6.0.6277.0
Windows Cluster Management Monitoring 6.0.6277.0
Windows 2003 cluster Management Library 6.0.6277.0
Windows Cluster Management Library 6.0.6277.0
Windows cluster Library 6.0.6278.0

Please note I realized we are very behind in cluster MPs, however,
apparently you cannot perform an in-place upgrade from the version we are
running to the current version. Due to dependencies on custom MPs, I am
hesistant to perform the steps needed to upgrade the cluster MP. Is there a
way to download an v 6.0.6277.1 so I can then upgrade to the latest version?

Post by Steven Halsey [MSFT]
You can switch the responsible agent in Operations Manager Console.
1. Go to administration Pane->Device Management->Agentless Managed
2. Right click on the virtual server name and select Change Proxy Agent
3. Select the new proxy agent and hit OK
It is normal for the cluster virtual server to be listed as agentless
monitoring as the passive node, the node that first discovers the cluster
virtual server becomes the host agent always. The way Operations Manager
works is only 1 agent can be responsible for the virtual server. What
happens is the agents on the physical nodes recognize that they are part of
a cluster and know which virtual servers they are running, and then they run
the appropriate monitoring. For some reason, on your agents or inside of
the Exchange 2007 Converted MP this logic isn't happening correctly.
If you could tell me the version of the Operating System on the cluster
(including SP), the type of Exchange cluster (CCR or SCC), and which of the
following Management Packs and the version are imported to the Management
Group it might help to figure out where the problem lies (Exchange 2007
Converted MP, Exchange 2007 Native MP Beta, Microsoft Windows Server Cluster
MP, Windows 2003 MP, and Windows 2008 MP).
I added another post where you can set the workflows that are currently
failing so they can be successful, and you won't see alerts about the
failing mapi logon or mail flow.
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
HOw do I switch which host is in charge? How do I prevent this from happening
again?

Post by Steven Halsey [MSFT]
If you use Operations Manager Console, and have the Agent Responsible for
monitoring the Virtual Server switched to the active node, do the errors go
away and do the performance counters for mapi-connectivity appear for the
virtual server?
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
The physical node monitoring the Exchange virtual server is the passive node!
The virtual server is listed in the Test-MAPIConnectivity Logon Latency
performance view, but there is no information being collected.
In the MOnitoring Pane->Exchange SErver 2007->State View, the Virtual Server
is healthy but the passive node has a warning below the Ex. Mailbox column.

Post by Steven Halsey [MSFT]
No you shouldn't need to run the commandlet. If you are still seeing this
behavior can you answer a couple of more questions?
If you look at the agentless managed under Administration, which physical
node is responsible for monitoring the virtual server? Is it the passive
node for which you are getting alerts or the active node?
If you look at the performance counters for Mapi logon under the monitoring
pane Exchange 2007, is the virtual server logging mapi connectivity latency
currently or is the data not being logged? Is the Ex ailbox instance
of
the
virtual server healthy or unhealthy?
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included script samples are subject to terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
Thanks for the reply. The Exchange MP I am running is 6.0.6461.0.
Ex. Mailbox - includes both physical nodes (one of which is passive)
AND
the
cluster name. Everything is monitored.
Microsoft Exchange 2007 Mailbox Servers Installation - Cluster name appears
as unmonitored. No physical nodes exist.
Ex. Common - All exchange servers (including cluster name) exist and are
monitored
Microsoft Exchange 2007 All Servers Installation - All exchange servers
(including cluster name) appear but are unmonitored
do I still need to run the cleanup you referred to?

Post by Steven Halsey [MSFT]
Which version of the Exchange 2007 MP are you running? In version
6278.12
there were fixes to the discoveries for Cluster servers, so if you
are
not
running this version you should upgrade then follow the instructions
in
the
MP Guide to clean up your environment (involves running the
remove-disabledmonitoringobject commandlet on your RMS).
If you look in the Operations Manager Console under discovered
inventory
and
check for the object types for all 4 of the types below, in which
ones
do
you see the passive physical node?
1. "Ex. Mailbox" - Appear as Monitored, physical node should not
be
of
this type
2. "Microsoft Exchange 2007 Mailbox Servers Installation" - Appear as
UnMonitored, physical node will appear as this type
3. "Ex. Common" - Appear as Monitored, physical node should be of
this
type
4. "Microsoft Exchange 2007 All Servers Installation" - Appear as
UnMonitored, physical node should be of this type
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no
rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
I am running the latest Exchange 2007 MP and suddenly we started
receiving
alerts from the inactive mailbox server node because some of the
MAPI
and
Test-Mailbox commands were failing. This particular servers has
been
the
inactive node for over a week. Suggestions on how I should resolve
this?

[Steven Halsey [MSFT]]

Two more questions: How many Exchange Virtual Servers are you running in the
cluster? How many Network resource names does the cluster have in all
including the Exchange, and MSDTC resources, cluster administration names,
etc?

On the upgrade of the MPs, there isn't anything that can be done with the
current versions of Operations Manager. Once a feature in an MP is made
public there is no way to remove it. So MP Authors either have to keep the
feature and working around the issues the feature might create, or break
upgrade to go to a cleaner monitoring state. This behavior is under
examination for future versions of Operations Manager. But unfortunately
for you it means either keeping your current cluster MP, or doing the work
to migrate your custom MP work to use the newer MP.
--
Steven Halsey [MSFT]

This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
Thanks for the information! I provided you with OS details and here are MP
Microsoft.Exchange.2007 (Converted) v6.0.6461
Windows 2003 Cluster Management Monitoring v6.0.6277.0
Windows Cluster Management Monitoring 6.0.6277.0
Windows 2003 cluster Management Library 6.0.6277.0
Windows Cluster Management Library 6.0.6277.0
Windows cluster Library 6.0.6278.0
Please note I realized we are very behind in cluster MPs, however,
apparently you cannot perform an in-place upgrade from the version we are
running to the current version. Due to dependencies on custom MPs, I am
hesistant to perform the steps needed to upgrade the cluster MP. Is there a
way to download an v 6.0.6277.1 so I can then upgrade to the latest version?

Post by Steven Halsey [MSFT]
You can switch the responsible agent in Operations Manager Console.
1. Go to administration Pane->Device Management->Agentless Managed
2. Right click on the virtual server name and select Change Proxy Agent
3. Select the new proxy agent and hit OK
It is normal for the cluster virtual server to be listed as agentless
monitoring as the passive node, the node that first discovers the cluster
virtual server becomes the host agent always. The way Operations Manager
works is only 1 agent can be responsible for the virtual server. What
happens is the agents on the physical nodes recognize that they are part of
a cluster and know which virtual servers they are running, and then they run
the appropriate monitoring. For some reason, on your agents or inside of
the Exchange 2007 Converted MP this logic isn't happening correctly.
If you could tell me the version of the Operating System on the cluster
(including SP), the type of Exchange cluster (CCR or SCC), and which of the
following Management Packs and the version are imported to the Management
Group it might help to figure out where the problem lies (Exchange 2007
Converted MP, Exchange 2007 Native MP Beta, Microsoft Windows Server Cluster
MP, Windows 2003 MP, and Windows 2008 MP).
I added another post where you can set the workflows that are currently
failing so they can be successful, and you won't see alerts about the
failing mapi logon or mail flow.
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
HOw do I switch which host is in charge? How do I prevent this from happening
again?

Post by Steven Halsey [MSFT]
If you use Operations Manager Console, and have the Agent Responsible for
monitoring the Virtual Server switched to the active node, do the
errors
go
away and do the performance counters for mapi-connectivity appear for the
virtual server?
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
The physical node monitoring the Exchange virtual server is the
passive
node!
The virtual server is listed in the Test-MAPIConnectivity Logon Latency
performance view, but there is no information being collected.
In the MOnitoring Pane->Exchange SErver 2007->State View, the
Virtual
Server
is healthy but the passive node has a warning below the Ex. Mailbox column.

Post by Steven Halsey [MSFT]
No you shouldn't need to run the commandlet. If you are still
seeing
this
behavior can you answer a couple of more questions?
If you look at the agentless managed under Administration, which physical
node is responsible for monitoring the virtual server? Is it the passive
node for which you are getting alerts or the active node?
If you look at the performance counters for Mapi logon under the monitoring
pane Exchange 2007, is the virtual server logging mapi connectivity latency
currently or is the data not being logged? Is the Ex ailbox instance
of
the
virtual server healthy or unhealthy?
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included script samples are subject to terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
Thanks for the reply. The Exchange MP I am running is 6.0.6461.0.
Ex. Mailbox - includes both physical nodes (one of which is passive)
AND
the
cluster name. Everything is monitored.
Microsoft Exchange 2007 Mailbox Servers Installation - Cluster
name
appears
as unmonitored. No physical nodes exist.
Ex. Common - All exchange servers (including cluster name) exist
and
are
monitored
Microsoft Exchange 2007 All Servers Installation - All exchange servers
(including cluster name) appear but are unmonitored
do I still need to run the cleanup you referred to?

Post by Steven Halsey [MSFT]
Which version of the Exchange 2007 MP are you running? In version
6278.12
there were fixes to the discoveries for Cluster servers, so if you
are
not
running this version you should upgrade then follow the instructions
in
the
MP Guide to clean up your environment (involves running the
remove-disabledmonitoringobject commandlet on your RMS).
If you look in the Operations Manager Console under discovered
inventory
and
check for the object types for all 4 of the types below, in which
ones
do
you see the passive physical node?
1. "Ex. Mailbox" - Appear as Monitored, physical node should not
be
of
this type
2. "Microsoft Exchange 2007 Mailbox Servers Installation" -
Appear
as
UnMonitored, physical node will appear as this type
3. "Ex. Common" - Appear as Monitored, physical node should be of
this
type
4. "Microsoft Exchange 2007 All Servers Installation" - Appear as
UnMonitored, physical node should be of this type
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no
rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
I am running the latest Exchange 2007 MP and suddenly we started
receiving
alerts from the inactive mailbox server node because some of the
MAPI
and
Test-Mailbox commands were failing. This particular servers has
been
the
inactive node for over a week. Suggestions on how I should resolve
this?

[Megan Kielman]

1 Exchange virtual server and 3 total network resource names.

Thanks!

Post by Steven Halsey [MSFT]
Two more questions: How many Exchange Virtual Servers are you running in the
cluster? How many Network resource names does the cluster have in all
including the Exchange, and MSDTC resources, cluster administration names,
etc?
On the upgrade of the MPs, there isn't anything that can be done with the
current versions of Operations Manager. Once a feature in an MP is made
public there is no way to remove it. So MP Authors either have to keep the
feature and working around the issues the feature might create, or break
upgrade to go to a cleaner monitoring state. This behavior is under
examination for future versions of Operations Manager. But unfortunately
for you it means either keeping your current cluster MP, or doing the work
to migrate your custom MP work to use the newer MP.
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
Thanks for the information! I provided you with OS details and here are MP
Microsoft.Exchange.2007 (Converted) v6.0.6461
Windows 2003 Cluster Management Monitoring v6.0.6277.0
Windows Cluster Management Monitoring 6.0.6277.0
Windows 2003 cluster Management Library 6.0.6277.0
Windows Cluster Management Library 6.0.6277.0
Windows cluster Library 6.0.6278.0
Please note I realized we are very behind in cluster MPs, however,
apparently you cannot perform an in-place upgrade from the version we are
running to the current version. Due to dependencies on custom MPs, I am
hesistant to perform the steps needed to upgrade the cluster MP. Is there a
way to download an v 6.0.6277.1 so I can then upgrade to the latest version?

Post by Steven Halsey [MSFT]
You can switch the responsible agent in Operations Manager Console.
1. Go to administration Pane->Device Management->Agentless Managed
2. Right click on the virtual server name and select Change Proxy Agent
3. Select the new proxy agent and hit OK
It is normal for the cluster virtual server to be listed as agentless
monitoring as the passive node, the node that first discovers the cluster
virtual server becomes the host agent always. The way Operations Manager
works is only 1 agent can be responsible for the virtual server. What
happens is the agents on the physical nodes recognize that they are part of
a cluster and know which virtual servers they are running, and then they run
the appropriate monitoring. For some reason, on your agents or inside of
the Exchange 2007 Converted MP this logic isn't happening correctly.
If you could tell me the version of the Operating System on the cluster
(including SP), the type of Exchange cluster (CCR or SCC), and which of the
following Management Packs and the version are imported to the Management
Group it might help to figure out where the problem lies (Exchange 2007
Converted MP, Exchange 2007 Native MP Beta, Microsoft Windows Server Cluster
MP, Windows 2003 MP, and Windows 2008 MP).
I added another post where you can set the workflows that are currently
failing so they can be successful, and you won't see alerts about the
failing mapi logon or mail flow.
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
HOw do I switch which host is in charge? How do I prevent this from happening
again?

Post by Steven Halsey [MSFT]
If you use Operations Manager Console, and have the Agent Responsible for
monitoring the Virtual Server switched to the active node, do the
errors
go
away and do the performance counters for mapi-connectivity appear for the
virtual server?
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
The physical node monitoring the Exchange virtual server is the
passive
node!
The virtual server is listed in the Test-MAPIConnectivity Logon Latency
performance view, but there is no information being collected.
In the MOnitoring Pane->Exchange SErver 2007->State View, the
Virtual
Server
is healthy but the passive node has a warning below the Ex. Mailbox column.

Post by Steven Halsey [MSFT]
No you shouldn't need to run the commandlet. If you are still
seeing
this
behavior can you answer a couple of more questions?
If you look at the agentless managed under Administration, which physical
node is responsible for monitoring the virtual server? Is it the passive
node for which you are getting alerts or the active node?
If you look at the performance counters for Mapi logon under the
monitoring
pane Exchange 2007, is the virtual server logging mapi connectivity
latency
currently or is the data not being logged? Is the Ex ailbox instance
of
the
virtual server healthy or unhealthy?
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no
rights.
Use of included script samples are subject to terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
Thanks for the reply. The Exchange MP I am running is 6.0.6461.0.
Ex. Mailbox - includes both physical nodes (one of which is passive)
AND
the
cluster name. Everything is monitored.
Microsoft Exchange 2007 Mailbox Servers Installation - Cluster
name
appears
as unmonitored. No physical nodes exist.
Ex. Common - All exchange servers (including cluster name) exist
and
are
monitored
Microsoft Exchange 2007 All Servers Installation - All exchange
servers
(including cluster name) appear but are unmonitored
do I still need to run the cleanup you referred to?

Post by Steven Halsey [MSFT]
Which version of the Exchange 2007 MP are you running? In version
6278.12
there were fixes to the discoveries for Cluster servers, so if you
are
not
running this version you should upgrade then follow the
instructions
in
the
MP Guide to clean up your environment (involves running the
remove-disabledmonitoringobject commandlet on your RMS).
If you look in the Operations Manager Console under discovered
inventory
and
check for the object types for all 4 of the types below, in which
ones
do
you see the passive physical node?
1. "Ex. Mailbox" - Appear as Monitored, physical node should not
be
of
this type
2. "Microsoft Exchange 2007 Mailbox Servers Installation" -
Appear
as
UnMonitored, physical node will appear as this type
3. "Ex. Common" - Appear as Monitored, physical node should be of
this
type
4. "Microsoft Exchange 2007 All Servers Installation" - Appear as
UnMonitored, physical node should be of this type
--
Steven Halsey [MSFT]
This posting is provided "AS IS" with no warranties, and confers no
rights.
Use of included samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

Post by Megan Kielman
I am running the latest Exchange 2007 MP and suddenly we started
receiving
alerts from the inactive mailbox server node because some of the
MAPI
and
Test-Mailbox commands were failing. This particular servers has
been
the
inactive node for over a week. Suggestions on how I should
resolve
this?