New Dell Hardware MP released, 3.1.1 A00


Dell has updated their hardware management pack.  The old version was 3.1.0.118.  The new version is 3.1.1.3.  I blogged about the old MP HERE

 

So – first thing – I did the unthinkableI read the release notes.  Here is what Dell says is new:

 

#####################################################################
RELEASE HIGHLIGHTS
#####################################################################

* A known defect with storage component discovery, where in some
scenarios (especially with Server Administrator version 5.5 or above
installed),storage details could not be viewed under servers in the
diagram view of the SCOM console, has been fixed.(DF283684/DF257743)

* A defect with error handling in the DellBMCLog.exe while handling
DellBMCLog error events in a Distributed Management Server
environment has been fixed.(DF260495)

* A known defect with health of Virtual Disks, where in some
scenarios health of Virtual Disks were not shown correctly, has been
fixed.(DF279394)

 

At first glance – I can see the PDF guide has not been updated.  So no major changes there. 

Bummer – Server 2008 is still not supported, according to the release notes and guide.  🙁

If you are going to leave the SNMP stuff enabled – you absolutely still need to make sure you have KB 951526 installed – or a newer KB that updates the same file - Momnetworkmodules.dll…. and the most current hotfix KB for that is 957511

I can see that DellBMCLogSetup.exe has been updated… this should be the fix referenced above… where this tool would crash on all your management servers previously.

I can see that Dell_MDStorageArray_ABBSetup.exe has been updated… so I assume this has something to do with the fixes in storage component discovery, or virtual disks.

I can see all of the MP files have been updated with a new time/date stamp as well.

 

Next, I ran a MPDiff to see what's new in this release.

 

1.  First off – there are 62 differences.  The majority of these are a change in the discoveries – to change the “MinLength” value from “1” to “0”.  The reason for this… is because if a discovery is set with a “1” or greater, the discovery will not work correctly if the discovered value is NULL.  Essentially, some of these discovery properties probably are empty on some systems, and the fact they were set to “1” previously caused discovery data failures on agents.  You can read more about this here:  http://blogs.technet.com/smsandmom/archive/2008/09/09/opsmgr-2007-discovery-data-processing-fails-with-invalid-monitoring-class-property-value-message.aspx

 

2.  Next - DellStorageDiscovery.vbs script was updated with a few lines of code.

3.  Next - DellStorageControllerUnitMonitor.vbs arguments and script were updated a bit.

4.  Next - ControllerVirtualDiskEventTrigger.vbs arguments and script were updated a bit.

 

#2, 3, and 4 above all appear to relate to fixing the storage discovery, and Virtual Disk issue documented in the release notes.

 

 

So….. not much as far as changes go.  The rest of this post – I will talk about the most common changes I make in this MP.

 

So…. first – to really understand any management pack - I run MPViewer against the MP.  This will help me to understand the classes, tasks, dependencies, discoveries, rules, monitors, etc…

The first thing I look at – is the Object discoveries.  I want to know:

1.  What are we discovering?

2.  How often are we discovering it?

3.  What is the discovery “hierarchy”…. meaning… what is the base discovery, and what other discoveries run against the target instances of the bas discovery.  This allows us to chain together, how the MP works… and will help us troubleshoot if something isn't showing up.

(See my previous post – where I also talk about these discoveries, what they do, how they work)

 

Here is the MPviewer discovery output:

 

image

 

The core discovery is running against all “Windows Computers” every 6 hours.

Using MPviewer – I can see the core XML for this discovery:

image

 

What this tells me – is that this discovers an object class, called “DellSystem”.  That’s my first discovered class type.  I can look at this in the monitoring console -  discovered inventory:

image

 

Then – I can look above at the MPViewer output – and see that “Dell Server Discovery” targets “Dell Systems Instances”.  Using the same methods – you can see the subsequent discoveries that target “Dell Server Instance” created by this discovery, and so on, and so on.  Why is this important?  It helps us understand the MP, and helps us find out when a MP isn’t “finding” something that we think it should.

 

Ok…. so, in general, here is what I turn off in this MP.

 

First – this MP includes Dell Printer discovery and monitoring.  Let’s say that I am not interested in monitoring printers, and don't want the additional load created by this?   So – I disable the “Dell printer Discovery”. 

I also decide – I don't want to deploy the SNMP based DRAC/CMC based moniotring…. or any SNMP based stuff for that matter…. so I disable all the discoveries that are targeted to “Microsoft.SystemCenter.Network.Device”

 

image

 

Continuing on…. I look at Monitors, and disable all the Monitors for the same:

 

image

 

Next up – rules….  I disable all the SNMP based rules that are targeting “Dell.Connections.CMC, Dell.Connections.DRAC_5, and Dell.Connections.DRACMC”…. consistent with my “No SNMP” approach.

Lastly on rules – since I decided not to implement the BMCLog executables on my management servers… I disable the two rules that are running on my Management Servers.

 image

 

Lastly – and probably MOST important…. If I have a LARGE agent count…. these discoveries put a lot of data in the database.  If they run too frequently, AND collect discovery properties that change often…. I will fill my OpsDB and Warehouse with ManagedEntityProperty data.  This may not sound significant – but it is.  A management pack should NOT collect information as a discovery property of a class, if the information will change often.  For instance, The “Temperature Unit Instance” class, contains a discovered property, called “Sensor Current Reading”

image

 

I am just picking on this one as an example…. this is the current reading of the temperature sensor… and probably changes a bit between each discovery.  This means every 6 hours (and much more frequent in previous versions of the older Dell MP’s) this data is added to the database…. and this builds up over time.  It is much better to create a monitor/rule for this – and enter this data into the database as performance data via a propertybag, instead of using a discovery property on a class.

For this reason – I like to decrease the frequency of discoveries if this is a known risk.  So – using my MPViewer data – I am off to the console – to find these discoveries, and change their frequency from every 6 hours, to every 12 or 24 hours…. depending on how fast you *need* your Dell systems to be discovered once they get an agent installed.

 

The problem is, the first discovery I want to look at, the “Dell System Discovery” is nowhere to be found in the console!  This is because they did not define a discovery type.  This is required for the Object Discovery to be viewable in the console.  It’s weird – because some of their discoveries DO show up… it looks like they just left this off some of them.

Much like the IBM MP I blogged about here:  http://blogs.technet.com/kevinholman/archive/2009/02/27/issues-with-the-current-ibm-mp-2-0-0-501-and-a-workaround.aspx  We will use the same workaround to get to all the Dell discoveries.

So – using the workaround, I search for Object Discoveries using “Dell” as the keyword:

 

image

 

From here – I can click “View Knowledge” to gain access to the discovery rule in the UI – and create an override to run this every 12 or 24 hours… depending on what you want.  I can also use this to disable the discoveries I mentioned above if you elect not to use them.

 

Now…. we are ready to tune the MP… there are a LOT of alerts generated in the field, in my experience, from items that not everyone considers valuable, like voltage sensors, etc… so get ready to turn off some alert rules as well, based on your monitoring requirements.

Comments (4)

  1. Michael Humphrey says:

    "Bummer – Server 2008 is still not supported, according to the release notes and guide.  :-("

    This was my first reaction as well. But after reading on…

    "* Supported Servers: All Dell servers which are supported by

    Dell(TM) OpenManage(TM) Server Administrator version 5.3.x and 5.4

    can be managed using the Dell Server and Printer Management Pack

    v3.1.1

    The list of supported servers is available at:

      • http://support.dell.com/support/edocs/software/smsom/5.3/en/

    peosom/installa.htm#wp999358

      • http://support.dell.com/support/edocs/software/smsom/5.4/en/

    peosom/installa.htm#wp999358"

    Based on the matrices at these links, it looks like there is server 2008 support although not for every device and every edition of 2k8.  Am I reading this wrong?

  2. Enrique Moreno says:

    Have you checked it for SCOM 2007 R2?

    Regards,

  3. Jan Matejka says:

    Yes, I have checked it for R2 version. Basic HW monitoring works correctly, but classes BIOS Config Instances and BMC for Dell Servers are Not monitored. I configured MP according to pdf guide. I am running Windows Servers 2008 x64 and SQL 2008 x64. I had problems with DellStorageDiscovery.vbs but I upgraded to 3.1.1.3 version and I have not received alert.

    Jan

  4. Enrique Moreno says:

    Hi again. Like I suppose you know, there is a new version for Dell´s MP, ver. 4.0. I´m going to ckeck it in lab enviroment. Have you tested it? Have you write any article? Any suggestion? Thanks in advance.

Skip to main content