locked
DHCP 2012 mp and monitoring the number available addresses in a scope RRS feed

  • Question

  • I have scopes that 9% of their ip addresses are available and the health of the "DHCP Server 2012 IPV4 Individual Scope Addresses Available Percentage Monitor" is indicating a healthy state.  No overrides are in place and the default threshold for warning is 30% and critical is 10%.  This does not appear to me that it is working the way it should.  When I open a performance view on this scope I see that perfromance data is being collected and it shows "Addresses in Use" at 189 and "Addresses Available" at 21.  Any ideas why the monitor is not working the way it should? 
    Wednesday, April 16, 2014 1:18 PM

All replies

  • Hi,

    Based on my understanding, the performance data indicates that the available percentage is only 10%, but it is still in a healthy state, as the monitor's default threshold for warning is 30% and critical is 10% .

    And based on my research, there is a known issue for the MP,

    Known issues:

    There are some challenges with this MP.  For instance – all scopes are discovered and monitored as a single object.  If a scope fills up – this will turn the monitor to an unhealthy state, and generate an alert.  If a subsequent scope fills up – you will NOT get another alert as the monitor is already unhealthy, and since it is a manual reset monitor… it is highly likely it will stay in this state unless someone resets it after resolving the issue.

    One option for this – is to rebuild these scope-full monitors as rules – and simply allow them to create a new alert for each event, so you wont miss any scopes that are full.  The downside is you will see multiple events/alerts as the DHCP service logs these events on a regular basis (once per hour)

    To understand more about the MP, I would like to suggest you go through the guide for the MP:

    http://www.microsoft.com/en-hk/download/details.aspx?id=39062

    Regards,

    Yan Li


    Regards, Yan Li

    Friday, April 18, 2014 6:06 AM
  • I am aware of the issue you mention, however the issue I have is that all monitors remain in a healthy state and no alerts have been generated for this monitor. 
    Monday, April 21, 2014 11:11 AM
  • Still not sure why this monitor is not working, but it is for this very reason we implented this mp.  Any ideas would be appreciated.
    Wednesday, April 30, 2014 6:03 PM
  • Hi,

    please to to DHCP server and veify if there are relevant warning logs.

    Regards,

    Mike


    Please remember to click “Mark as Answer” on the post that helps you, and to click “Unmark as Answer” if a marked post does not actually answer your question. This can be beneficial to other community members reading the thread.

    Thursday, May 1, 2014 10:16 AM
  • you asked and I try to answer :-). Let's try to set a common play field. First of all, I have the MP in production (version 6.0.7063.0) and it seems to work properly in the sense that I have alerting on close to exhaustion scopes.

    The monitor I assume you’re referring to is “DHCP Server 2012 IPV4 Individual Scope Addresses Available Percentage Monitor”. It is a dual states monitor: warning under 30% and error under 10% of free IP addresses. It fires if the threshold is breached for three consecutive samples. It samples every 15 minutes. It is a monitor ultimately based on a powershell probe. So you won’t find the performance counter you can read in the monitor configuration (IPV4Scope-PercentageAvailable) in perfmon, since this is a script generated counter internal to the management pack.

    The same probe collects the performance data for the scope: IPV4Scope-AddressesInUse and IPV4Scope-AddressesAvailable. The probe uses the standard Get-DhcpServerv4ScopeStatistics from the DHCP powershell module.

    If all this is true, to tourbleshoot the issue you should:

    -          Run the Get-DhcpServerv4ScopeStatistics from an administrative powershell console on one of the faulty DHCP servers. This way you can check if it’s a DHCP instrumentation issue.

    -          Check if the agent is restarting or reloading (event id 21015) very frequently. A restarting agent breaks the consecutive time threshold used by this monitor. So if your agent restarts every 30’ you will never get any alert since the monitor fires after 3 consecutive samples each every 15’

    Let me know how this evolves.

    Ciao

    Daniele


    - Daniele
    Microsoft MVP System Center Cloud and Datacenter Management

    Unisciti alla community italiana per System Center http://www.ugisystemcenter.org

    http://nocentdocent.wordpress.com
    This posting is provided “AS IS” with no warranties, and confers no rights.
    Please remember to click “Mark as Answer” on the post that helps you, and to click “Unmark as Answer” if a marked post does not actually answer your question. This can be beneficial to other community members reading the thread.

    Friday, May 2, 2014 2:36 PM
  • Thanks Daniele.  I am able to run Get-DhcpServerv4ScopeStatistics on the DC.  I can see several scopes listed above 90 for percentage in use.  No issues with the agent restarting.
    Friday, May 2, 2014 8:21 PM
  • When I look at a performance view of a scope I see data for both IPV4Scope-AddressesInUse and IPV4Scope-AddressesAvailable too.
    Friday, May 2, 2014 8:43 PM
  • Hi Keith,

    this confirms that the OS side is OK and that the probe is ok too. I suggest: set an override so that the alert fires the first time the scope gets under threshold (consecutive samples=1). If the alert fires then you're in the case where the monitor doesn't get a chance to run at least 3 times in a row. If you don't get the alert is highly probable the monitoring rule crashes, you should find an error event in the operations manager log, and in any case try to delete the agent cache.

    -Daniele


    - Daniele
    Microsoft MVP System Center Cloud and Datacenter Management

    Unisciti alla community italiana per System Center http://www.ugisystemcenter.org

    http://nocentdocent.wordpress.com
    This posting is provided “AS IS” with no warranties, and confers no rights.
    Please remember to click “Mark as Answer” on the post that helps you, and to click “Unmark as Answer” if a marked post does not actually answer your question. This can be beneficial to other community members reading the thread.

    Saturday, May 3, 2014 8:37 AM
  • Ok, made the following overrides:

    Reduced consecutive samples to 1.

    Increased critical threshold to 30 and warning to 60.

    Recreated the Agent Health Service Store on the DHCP server.

    No errors in the OpsMgr Agent log.

    After an hour no state changes on the "DHCP Server 2012 IPV4 Individual Scope Addresses Available Percentage Monitor".  

    I am aware of several scopes that are beyond this threshold and should have changed.  

    Keith

    Saturday, May 3, 2014 6:05 PM
  • Interesting.  There are a couple dhcp servers with warnings.  One of the DHCP servers has about 50 scopes and all of them are in a warning state.  I checked one of the scopes and they have not crossed the threshold.  Does it take only one scope to cross the threshold to cause all of them to change state?  Seems like it.  
    Monday, May 5, 2014 2:58 PM
  • Hi Keith,

    it depends on the rollup algorithm, as a starting point you should check then "Microsoft Server DHCP 2012\Scope Health" view. Honestly if the scopes don't turn to unhealthy there and you don't have any error in OpsMgr event log, then we must suppose an hang in the monitoringhost process in charge of the powershell engine. If this is the case you have other rules that aren't running. It's a very uncommon scenario, but it happened to me in a few cases. To tshoot such a situation you should turn on OPpsMgr Tracing at debug elvel for the healthservice and powershell modules. If the problem persist it is probably a good idea to open a CSS case.


    - Daniele
    Microsoft MVP System Center Cloud and Datacenter Management

    Unisciti alla community italiana per System Center http://www.ugisystemcenter.org

    http://nocentdocent.wordpress.com
    This posting is provided “AS IS” with no warranties, and confers no rights.
    Please remember to click “Mark as Answer” on the post that helps you, and to click “Unmark as Answer” if a marked post does not actually answer your question. This can be beneficial to other community members reading the thread.

    Monday, May 5, 2014 3:49 PM
  • Daniele,

    Many thanks for the guidance here.  I will continue down this path and will let you know how things work out.

    Keith


    Thursday, May 8, 2014 6:07 PM