SCDPM Online

The Microsoft DPM 2010 management pack for System Center Operations Manager 2007 was released a couple of months ago now and itt is available on this download page, http://www.microsoft.com/downloads/en/details.aspx?FamilyID=32077d99-618f-43d0-843d-4ba4f8019f84.  It is also now available on the web service catalogue.

So what's new and improved in this MP?  Well, there have been a number of significant improvements that have been detailed by Jonathan Almquist at http://blogs.technet.com/b/jonathanalmquist/archive/2010/10/21/data-protection-manager-2010-management-pack-just-released-things-you-need-to-know.aspx.

There is also a summary of all the new features on the download page which I have pasted below, just for you. 

Other improvements I have noticed are:

1. The number of object properties discovered has been significantly decreased. In larger environments this will prevent significant performance issues on the RMS and database.
2. The number of object properties discovering dynamic data has also been significantly reduced. Although the number of discovered object properties has been reduced, the biggest cause of data churn with this MP was that is was discovering dynamic data.  This has also been significantly reduced with data on volume sizes and total recovery points no longer being collected as a data source object property.
3. Three new state views. There are now new state views for Datasource State (Hyper-V), Datasource State (Secondary) and Protected Computer State (Secondary).
   
4. The default discovery interval has been increased. This again was a needed change.  One of the first things you noticed with the previous MP was that the data discovery ran every two hours and this in turn a high level of configuration churn. The data discovery is now every 12 hours.
5. Updated images.  The DPM server images in the MP have been updated and look better.  This is only a cosmetic improvement but I think cosmetic improvements are always welcome. 
   
6. SLA-based, Standard or Ticketing System. This MP is the first to offer monitoring options for Standard monitoring, SLA monitoring and monitoring for use with a ticketing system.  This is really nice and shows that the authors of this MP have thought about the different ways in which monitoring is performed in different companies.  I will not go over the details here as Jonathan Almquist has done a very good write-up in the blog post mentioned above.
7. More monitors. There are now many more monitors in this new MP which means you can now truly use Operations Manager to accurately monitor you DPM servers.
8. Consolidation, correlation and suppression. This MP has consolidation, correlation and suppression implemented in many of its workflows.  This is really useful as it does not mean you will necessarily get an alert for every Error or Warning event posted to the DPM Alerts log.  Again though, this has been detailed by Jonathan in his blog so I wont go through it again here.
9. Documentation. Not much to say here other than the documentation is very good and seems to be constantly improved.

Now for the not so positive side, here are the issues I have with the MP:

1. A number of the views do not have the State column enabled by default. Although this is relatively minor as you only need to right click in the State View and select Personalize View to enable the viewing of the state column, I think the State column is one of the most important columns in a state view!!!
2. Views do noy have friendly names. This isn't a viewable issue unless you look at the MP at an authoring level but it is a big frustration of mine.  The majority of views in this MP have nice friendly names which you see in the console, but break open the MP and the real view names have the horrible "View_hahie883r4...." names so you can't tell which view is which.  This occurs when views are created in the Operations Manager console and saved to the MP rather than created in the XML or authoring console; it would have been nice if these had been created properly.
3. Still using Events in DPM Alerts Log . The DPM management pack provides monitoring by monitoring the DPM Alerts event log.  When Alert Publishing has been enabled on a DPM server, event log messages will start getting posted to the DPM Alerts log whenever there is a failure/alert and when a failure/alert has been resolved.  The monitors in the DPM management pack are based on these log entries and the state of the monitored data sources change based on these monitors.  I personally am not a fan of this type of monitoring, I would much rather perform some kind of data query to accurately return the state of what is being monitored.  The problem with monitors looking at event logs is that if an event isn't posted to log or a event is missed by the monitor, the state of the monitor will not accurately reflect the state of what is being monitored.
4. Dynamic data is still discovered for disk objects. Although the number of object properties discovering dynamic data has been significantly reduced in this MP, there is one exception.  The disk objects still have properties that collect dynamic data such as Used Space, Unallocated Space, % Allocated and % Unallocated.  This data will change all the time so will have to be updated on every discovery which increases data churn; this is bad!  To put this in some context though, if you have a small number of storage pool disks in a small number of DPM servers, you probably wont notice any performance issues as the amount of object property data that will need updating will still be relatively small.  If however, you have a larger deployment with a lot of storage pool disks, the amount of object property data that will need updating is significantly larger and will cause more noticeable performance impacts.  I do understand that this data is needed in the MP, but using a performance collection rule would be a better way of collecting this data.
5. No Reports. This is a real shame.  As Operations Manager can monitor all your DPM servers centrally, it would be great to have some cool DPM availability, performance and SLA reports!

Overall this MP is a significant step forward and should definitely be implemented.  The improvements in this MP by far out way the negative points that I mentioned above, and I hope that future DPM MP upgrades are as good as this one!

I'm sure I have missed a few things but hopefully I have covered all the major points and this has proved helpful to you.  And, if you have any feedback on this MP please feel free to post a comment below.

Happy monitoring.

David Allen MVP System Center Operations Manager http://www.scdpmonline.org http://wmug.co.uk/blogs/aquilaweb