This posting is a continuation of my basics series, this one deals with Understanding Grooming and Aggregation, two very important concepts which can have great impact on your environment if they are not functioning properly.
First you need to understand the different between these two concepts, at their most basic, Grooming is the method where data older then a specific period, called the retention period, is deleted. Aggregation is the method where raw data, which is collected on short intervals i.e. 5 minutes, are averaged into hourly and daily time intervals used for reporting purposes.
The Operations Database
A stored procedure, called p_PartitioningandGrooming is run by a SCOM rule on a once a day schedule. The retention settings for each data type can be set in the console under Administration > Settings > Database Grooming.
You can see the history of the OpsDB grooming jobs with the following SQL query:
select * from InternalJobHistory order by InternalJobHistoryId desc
A Status of 1 indicates a successful completion.
The Operations Data-warehouse
The Data Warehouse is more complex, you cannot change data retention settings from the console, they exist in a table called StandardDatasetAggregation, which contains the grooming and aggregation retention periods, intervals and other related data.
Data in the DW database moves through several tables as part of the aggregations process as the data is collected it is written to a Raw Data table, from there is will move to a Staging table where it will be aggregated and then moved to a relevant aggregated data table (Hourly / Daily) where it will be available to the reporting services.
Then once the data in these tables is older then the data retention period it will be groomed out of the database.
So why is it important for aggregation to work?
Reports use the hourly and daily data in order to generate, this means that if your aggregations are not working or slow there will be gaps in your reporting. Example below:
And there you have it. the basics behind aggregation and grooming, I hope this has been informative.
This is a summary of the wave of Management Packs that were released in November 2015. Information and download location in the links provided:
OpsLogix Swift Management Pack V2.5.1.57 – here Visual Studio Team Foundation Server 2015 Management Pack v14.0.24622.0 – here SQL Server v 6.6.3.0 – here SQL Server v 6.6.4.0 – here Windows Server 2012 DHCP v6.0.7299.0–here Opslogix VMWare ManagementPack V1.3.1.1– here Management Pack for Microsoft Azure (Technical Preview) v 1.3.10.0 – here
If you know of any other Management Packs that have been released recently that I may have missed leave me a note in the comments and I’ll add them.
Tim McFadden’s maintenance mode tool has gone a long way towards filling a glaring gap in the SCOM tool set, That being the ability to easily schedule maintenance without having to hack together scheduled tasks or runbooks.
New Features in this version:
CSV Server Import – Now you can schedule a list of Windows or Unix computers for Maintenance Mode using a CSV (Comma Separated Values) file.
Intelligent Maintenance Mode – Before a computer or object is put into maintenance mode it is checked to see if it’s already in maintenance mode. If the existing maintenance window is longer then the new maintenance window, the existing window is kept.
Unix/Linux Support for Immediate Maintenance Mode – Unix Admins can now use the MMNow page to immediately put servers into Maintenance Mode using a web browser or wget from a shell script.
Management Servers in a Group are prevented from going into Maintenance Mode – If a group is scheduled for maintenance mode that contains Management Servers. The Management Servers are now skipped from going into Maintenance Mode.
New Jobs Report Fixed – Some users were experiencing an issue where the MMScheduler box was grayed out. This has now been resolved.
Improved permissions tool – It is now easier to grant new permissions as existing users and groups show up when the tool is opened.
Improved Logging – Events are created when a Schedule Maintenance Job is run or when a user puts a server into maintenance mode immediately using MM Now
With another year of innovation in the ESM space behind us, I for one am certainly looking forward to see what 2014 brings us, both in the System Center world and in the world at large.
I’d also like to say thank you again to all the supporters of this Blog, it’s been rewarding to finally get it up and running and seeing the readership climbing every month has been fantastic.