Kevin Holman's System Center Blog

Posts in this blog are provided "AS IS" with no warranties, and confers no rights. Use of included script samples are subject to the terms specified in the Terms of UseAre you interested in having a dedicated engineer that will be your Mic

OpsMgr: MP Update: New Base OS MP 6.0.7026.0

OpsMgr: MP Update: New Base OS MP 6.0.7026.0

  • Comments 21
  • Likes

 

A new Base OS MP Version 6.0.7026.0 has shipped.  This management pack includes updated MP’s for Windows 2003 through Windows 2012 operating systems.  This updated MP will import into OpsMgr 2007 or 2012 management groups.

 

http://www.microsoft.com/en-us/download/details.aspx?id=9296

 

image

 

 

Ok – so what's new in this MP?

 

The April 2013 update (version 6.0.7026.0) of the Windows Server Operating System Management Pack contains the following changes:

  • Fixed a bug in Microsoft.Windows.Server.2008.Monitoring.mp where the performance information for Processor was not getting collected.
  • Made monitoring of Cluster Shared Volume consistent with monitoring of Logical Disks by adding performance collection rules. (“Cluster Shared Volume - Free space / MB”,”Cluster Shared Volume - Total size / MB”,”Cluster Shared Volume - Free space / %”,”Cluster Disk - Total size / MB”,”Cluster Disk - Free space / MB”,”Cluster Disk - Free space / %”)
  • Fixed bug in Microsoft.Windows.Server.ClusterSharedVolumeMonitoring.mp where the Cluster disks running on Windows Server 2008 (non R2) were not discovered.
  • Fixed bug 'Cluster Disk Free Space Percent' and Cluster Disk Free Space MB' monitors generate alerts with bad descriptions when the volume label of a cluster disk is empty.
  • Added feature to raise event when NTLM requests time out and customers are unable to use mailboxes, outlook stops responding, due to the low default value for Max Concurrent API registry Key (HLM\SYSTEM\CurrentControlSet\Services\Netlogon\Parameters) , which is a ceiling for the maximum NTLM or Kerberos PAC password validations a server can take care of at a time. It uses the “Netlogon” performance counter to check for the issue.

 

These fixes address the majority of known issues discussed in my last article on the Base OS MP:

http://blogs.technet.com/b/kevinholman/archive/2012/09/27/opsmgr-mp-update-new-base-os-mp-6-0-6989-0-adds-support-for-monitoring-windows-server-2012-os-and-fixes-some-previous-issues.aspx

 

A note on Processor utilization monitoring and collection:

Distinct rules and monitors were created for Windows Server 2008, and 2008 R2.  Server 2008 will monitor and collect “Processor\% Processor Time” while Server 2008 R2 will monitor and collect “Processor Information\% Processor Time”.  Overrides were included in the MP to disable the “2008” rules and monitors for the 2008 R2 instances.  If for some reason you prefer to collect and monitor "from “Processor” instead of “Processor Information”, for instance if this breaks some of your existing reports, it is very simple to just override those rules and monitors back to enabled.  An unsealed override will always trump a sealed override.

 

 

 

Known Issues in this MP:

1.  The knowledge for the 2008 and 2008 R2 Total CPU Utilization Percentage is incorrect – the monitor was updated to a default value of 3 samples but the knowledge still reflects 5 samples.  This is still an issue (no biggie)  The 2012 monitors use 5 samples by default with correct knowledge.

2.  There are now collection rules for Cluster disks and CSV for free space (MB), free space (%), and total size (MB),  If you want performance reports on other perfmon objects that are available in perfmon but not included in our MP, such as disk latency, idle time, etc., you will need to create these.  Since this can be complicated to get it right – I wrote an article on how to do this correctly, and offer a sample MP for download:  http://blogs.technet.com/b/kevinholman/archive/2012/09/27/opsmgr-authoring-performance-collection-rules-for-cluster-disks-the-right-way.aspx

3.  The new monitor for Max Concurrent API has some issues and will generate a false alert in some cases.  If you have servers where this is happening – disable this monitor and it will be addressed in the next release of the MP.

Comments
  • Fixed bug 'Cluster Disk Free Space Percent' and Cluster Disk Free Space MB' monitors generate alerts with bad descriptions when the volume label of a cluster disk is empty.

    -->Could it be that the label is not right in the Alert Text in the new Managementpack!

    there are only letters like 3,4,5 and not the labelText!

    :-(

  • The Max Concurrent API monitor seems to alert on all non-domain joined computers, due to a lack of NetLogon performance counters, which the data source script does not handle well.  Our workaround at present is to create a group of Workgroup computers (manually unfortunately, as we cant identify workgroup computers dynamically), and disable the monitor for this group.

    Could you feed this back to the MP developers?

  • Hi Steve - I will.  There are other problems with this monitor where it goes red when the script encounters an error... which isn't terribly uncommon.

  • I'm seeing issues with the Windows Server 2012 Max Concurrent API Monitor also.  Using perfmon I don't see any related counter under netlogon to see what this current value is.

  • We are also experieincing issues with API monitor. In each case that we receive alerts, and then go back to verify that there is an issue, everything is fine. We have been following below as verification:

    support.microsoft.com/.../2688798

    In order to determine the best MaxConcurrentApi value for your servers, several data points must be brought together and calculated by using a formula. The data to be used to estimate MaxConcurrentApi is as follows:

    • Net Logon semaphore acquires

    • Net Logon semaphore time-outs

    • Net Logon average semaphore hold time

    • Duration of the performance logging that is completed, measured in seconds

    After the data is obtained, the following formula can be used to estimate the correct MaxConcurrentApi value:

    (semaphore_acquires + semaphore_time-outs) * average_semaphore_hold_time / time_collection_length = < New_MaxConcurrentApi_setting

    After you collect the Net Logon performance data from when the server was under authentication load, you should determine the duration of the data-collecting process by looking at the Line View beginning and end times.

  • Hello,

    Same here we are having a lot of alerts with the "Windows Server 2008 Max Concurrent API Monitor".

    Should the update done server per server or could we have a parameter in the Monitor to change?

    For now what I see on the monitor is:

    Alert On State, Alert Priority, Alert Severity, Auto-Resolve Alert, Enabled, Generates Alert, Interval seconds, Sync Time, Time Seconds.

    Anything missing on this monitor: Windows Server 2008 Max Concurrent API Monitor?

    Which parameter should be fixed?

    Thanks,

    Dom

  • I recommend disabling the API monitor until the next release of the MP at this point.

  • Kevin;  

    I was using your 2003 and 2008 addemdums to 6.0.6958.0.  Are they still needed with this new version?

    Thanks,

    Hector

  • Kevin,

    does MS plan to release a Performance By Utilization report that includes Processor Information as well as Processor, or is the long-term "fix" for broken 2008 R2 performance reports to disable the Processor Information performance collection rules and re-enable the Processor performance collection rules?

  • ii want to discover my cluster disks also as Windows server 2008 logical disk. Is it mandatory to include the cluster shared volume MP while importing or i can ignore it?

  • @Hector Hernandez  - yes my addendums are still necessary if you want modified alert descriptions.

    @John_Curtiss - I don't know if they plan to update any reports or not. I'd hope these are addressed in a future update.  otherwise - most customers I work with build their own custom reports.

    @ Vijayh - I don't recommend that.  The only way to accomplish that is to use an older MP... we disabled discovery of cluster disks in the logical disk discovery script.  Cluster disks as logical disks have some significant drawbacks, which is why they were changed to a new class type.

  • Hi Kevin,

    We found out yesterday that update in 6.0.7026 overrides the the Total CPU Utilization Percentage for Windows Server 2008 R2 Operating System, and the parameter Enabled changed to False! Is this a bug? Why it is disabling it (by overriding the Enabled to False)?

    Thanks,

    Ogun

  • @ogun -

    I covered this topic above in the section titled:  A note on Processor utilization monitoring and collection:

  • Thanks for your quick response Kevin. Yes, I read that part already, but apparently Total CPU Utilization Percentage monitor is also affected.  We overrode it in our management pack with proper threshold value, interval, etc, but noticed the monitor didn't generate alarms when the CPU utilization was beyond our threshold value (it was working before with the 6.0.6989) after upgrading to the version 6.0.7026. The solution that we found override the Enabled parameter to True, now the Effective Value shows as True. We decided to add the override for the "Total CPU Utilization Percentage" on Windows Server 2008 Operating System as it looked like a good generic target for our system. Should we do that override in a different target? As a side note, Windows Operating Sytem doesn't have this mentioned performance monitor but Windows Server 2008 Core Operating System and Windows Server 2008 Operating System do.

  • www.microsoft.com/.../details.aspx - doesn't work, do you have an updated link?

Your comment has been posted.   Close
Thank you, your comment requires moderation so it may take a while to appear.   Close
Leave a Comment
Search Blogs