Tuesday, November 4, 2014

Updated MP: Exchange Server 2013, Version 15.0.652.19

Update 2014-11-04: A good friend of mine, Oleg Kapustin, has written a posting about the Exchange MP as well. Since he’s a real good developer he has some additional insights in this MP. So for anyone using this MP, I highly recommend to read his posting as well, to be found here. Especially the part about ‘playing with the interval parameters’ is very interesting…

The first version of the Exchange Server 2013 MP…
We all know that the first version of the Exchange Server 2013 MP missed out on a lot of details. It reduced SCOM to nothing but a reporting tool since Exchange Server 2013 is expected to manage and monitor itself. So no additional monitoring by SCOM was required. It only needed to pick up the statuses and Alerts created by Exchange Server 2013 itself and that was it. And whenever something happened, Exchange would remedy itself…

In scenarios where organizations run many Exchange Servers (think about topologies with 100+ Exchange servers at least) and where are dedicated Exchange engineers available, an approach like this could work. Also self remediation could work in a scenario like this.

But how about smaller organizations, running less than 25 Exchange servers where the engineers have to keep a plethora of systems up and running, among them Exchange? So additional knowledge, like most of the SCOM Alerts contain, is very welcome.

In a situation like this the previous version of the Exchange Server 2013 MP didn’t deliver so Microsoft got a lot of criticism. Also from me

The new version is here…
At TechEd Europe 2014 Microsoft presented the newest version of the Exchange Server MP, version 15.0.652.19. When looking at the version numbers it looks like much hasn’t changed since only the last digits are changed (the previous version was on level 15.0.652.18).

However, MUCH has changed, so the version number is not telling the whole story here Smile. Some examples:

  1. Visualization (or better, eye candy)
    The status of the health of the Exchange Server 2013 components is visualized in an awesome way. Clearly Microsoft has decided to use the new features of the SCOM 2012 UR#2 Console (or later versions). It gives the presentation in the SCOM Console of the different Exchange Server 2013 components almost a sexy look & feel.

  2. Alerting
    When checking out this MP in MP Viewer you’ll see that there are only TWO Unit Monitors which will raise an Alert, and one Rule, which is disabled by default.
    image
    One thing to reckon with is that the Unit Monitor ‘>’ doesn’t relate to one of the default Aggregate Rollup Monitors (Availability, Configuration, Performance, Security) but has a place of it’s own:
    image
    So when building DA’s some additional tweaking is required. Use this posting of mine an be happy.

    However, there is no lack of amount of Alerts. This is because Exchange Server 2013 is still monitoring itself, and SCOM picks up the events which require attention.

    This way you get the best of two worlds: a single pane of glass (SCOM) and Exchange Server 2013 monitoring itself, without running those monitoring workloads twice. Future will show us whether this approach is viable. But the way it’s presented in SCOM 2012 R2 UR#2, it looks promising.

    Hopefully the website of Microsoft contains good information about how to solve the issues reported by Exchange Server 2013.

  3. Reports
    The first version of the Exchange Server 2013 MP contained NO reports. Period. However, the latest version of the Exchange Server 2013 MP contains MANY reports, also about the biggest mailboxes!

    In total 14 Reports are present:
    image

  4. Documentation
    For SCOM and many other Microsoft based technologies I live by the credo RTFM. Where the previous version of the Exchange Server 2013 MP didn’t contain any serious document at all, this version contains a REAL MP Guide, containing all the relevant information.

Verdict
Microsoft has shown to LISTEN to its customers. Even though this MP still uses Exchange Server 2013 to monitor itself, there is much to say about this approach. Why create a duplicate burden on those servers by running the same monitoring workloads twice (Exchange itself and the related MP)?

Point is that SCOM should be the single pane of glass for the organization using it to monitor the health of their IT systems and services, whether on-prem or cloud based (Azure, AWS and so on).

For what I’ve seen and heard so far, the latest version of this MP is a huge improvement compared to the previous version. So this is good news. I really hope the website Microsoft uses for providing additional information on the Alerts and how to solve them is just as good as this MP. Future will tell.

And don’t be afraid to share your personal experiences with this MP.

SCOM 2012 R2 UR#2 (and later) only!!!
For now this MP only runs in a SCOM 2012 R2 UR#2 (or later) environment. Later on a version for SCOM 2012 SP1 UR#x will be published, date unknown so far.

Where to download from?
MP can be downloaded from here. When interested in some screenshots of this new MP, visit this blog posting of Daniel Savage, Microsoft PM for the MPs.

Until now this version isn’t available from the MP Catalog available from the SCOM Console. It contains the old version:
image
I expect this to be fixed soon.

15 comments:

Richard Kinser said...

I had issues with this update. All of the notification emails for these alerts only contained {2} for the Subject and Alert Name. I let it sit for over 24 hours hoping it would resolve itself. It did not. So, I removed the MP and waited a day and then re-installed it. Now, the discovery is failing for all my exchange servers. So, be careful when installing...

Marnix Wolf said...

Hi Richard.

This is bad news. Can you tell me what error messages are shown in the OpsMgr eventlog of those Exchange servers? Perhaps I can help (or some of my contacts).

Cheers,
Marnix

Richard Kinser said...

Here is the error in the event log:

MicrosoftExchangeDiscovery.ps1 : (309): $orgCfg = Get-OrganizationConfig -DomainController $exchangeServer.OriginatingServer

System.InvalidOperationException: Couldn't find the Microsoft Exchange Recipient object.

I installed this updated Exchange 2013 MP (over the older one) about an hour after I updated my SCOM 2012 R2 environment to UR4. So, I am not sure if that contributed to the issue.

I have resolved at least the discovery issue by again removing the Exchange 2013 MP and then uninstalling the SCOM agent from each Exchange Server. Then I re-installed the SCOM agents and re-installed the Exchange 2013 MP. The discoveries have run, just waiting for the first alert to make sure the variable replacement is working in email notifications.

Richard Kinser said...

just got some alerts, and the variable replacement in the Alert emails is still not happening for the Alert Name. Both in the Subject Line and in the email body. Just has {2} instead of the actual alert name.

Marnix Wolf said...

Hi Richard.

I have passed the issue (parameter replacement in Notifications not happening) on to someone who knows a LOT more about this MP.

When I get feedback I'll let you know.

Cheers,
Marnix

Oleg Kapustin said...

1. Re discovery - this needs some investigation.
2. Re parameter replacement - see the last page of the guide:
---
Alerts generated by the previous version of the Management Pack for Exchange Server 2013 have unreadable name.
After upgrading the Management Pack from the version 15.0.620.18, alerts generated before the upgrade have the following name:
{2}
This happens because alert name has been parameterized, old alerts do not have the value for the required parameter. To regenerate alerts complete following steps:
1. Close all alerts generated before the upgrade.
2. Put all Health Set objects into Maintenance Mode for 30 minutes. All respective monitors will recheck their state and will generate new alerts after the Maintenance Mode interval is over.
---

Marnix Wolf said...

Hi Richard.

Can you sent me your e-mailaddress? This way Oleg can contact you directly and sort things out without using comments on my blog :).

You can mail me on MSWCAP GMAIL COM

Richard Kinser said...

Oleg,

I followed the steps that you listed, but after the objects came out of Maintenance Mode, the Alert Notification emails are still not showing the Alert Name, just {2} in both the subject line and in the message body.

I am sorry I didn't mention this before, but in the console, it shows the Alert Name, it's only in the Notification emails that it's showing {2}

meer said...

No support for SCOM 2012 SP1 :(

meer said...

NO support for SCOM 2012 SP1

rob1974 said...

Any idea if they fixed the exchange 2013 discovery as well?

http://jama00.wordpress.com/2014/05/13/microsoft-exchange-server-2013-discovery/

Marnix Wolf said...

Hi Rob.

Will pass your question on to someone I know who's involved with the development of this MP.

Cheers,
Marnix

Marnix Wolf said...

Hi Rob1974.

This is the answer I got (copied and pasted below):

'...Hi Marnix,

The answer for this question is simple: there are just 2 changes in the discovery:

1. PS script discovers more objects (everything that was introduced with the update)

2. A condition detection which blocks discovery for virtual nodes (cluster names) has been added.


Frankly speaking, I do not understand Rob’s point: in the previous version, discovery was a composition of Microsoft.Windows.Discovery.RegistryProvider and PS script. It first checks for the existence of value for SOFTWARE\Microsoft\ExchangeServer\v15\Setup\MsiInstallPath\InstallPath registry attribute, which we need to run the PS part of the discovery. Exchange 2013 requires Windows Server 2008 R2 or higher which has PS2 preinstalled, so we assume that every Exchange 2013 Server has PS (by the way, Exchange 2013 needs PS3 as a prerequisite). Thus, I don’t understand how the discovery can get to PS on Windows Server 2003.

Re targeting to the base class (Microsoft.Windows.Server.Computer) – this is a common design pattern. Another option is to target to the each version-specific class, but in this case we’ll end with multiple discovery workflows and references to a bunch of version-specific MPs. This is obviously a bad design J

Finally, if someone wants to override the discovery for not supported OSs or disable it completely and enable only for the set of selected servers – I’m totally fine with that – it will [slightly] reduce the monitoring footprint for non-exchange systems and absolutely doesn’t hurt.

Hope this helps...'

bsuresh said...

Hi Marnix,

I could see the latest version (15.0.652.19) of Exchange MP is being supported in SCOM 2012 SP1 as per MP guide.

As Richard explained, I am also getting Alert Name as {2} in Email notifications though the Alert description is correct in console itself.

Do we have any fix for Alert name in Email notifications?

Note: I am running with SCOM 2012 SP1.

Regards,
Balaji

Warren Kahn said...

Has anyone managed to resolve the issue with the alert name displaying {2}?