Our retention time for Incidents for the CMDB is 90 days after which they get moved to the data warehouse for retrieval. What happens to the attachments that were part of that Incident? Looks to me like they get stripped but wasn't sure.

When the work item that the attachment is related to is groomed out of the CMDB, the file attachment is deleted from the database.

• Proposed as answer by Andreas BaumgartenMVP, Moderator 9 hours 16 minutes ago

When the work item that the attachment is related to is groomed out of the CMDB, the file attachment is deleted from the database.

• Proposed as answer by Andreas BaumgartenMVP, Moderator Wednesday, August 19, 2015 10:00 PM

When the work item that the attachment is related to is groomed out of the CMDB, the file attachment is deleted from the database.

• Proposed as answer by Andreas BaumgartenMVP, Moderator Wednesday, August 19, 2015 10:00 PM

When the work item that the attachment is related to is groomed out of the CMDB, the file attachment is deleted from the database.

• Proposed as answer by Andreas BaumgartenMVP, Moderator Wednesday, August 19, 2015 10:00 PM
• Marked as answer by Xin GuoMicrosoft contingent staff, Moderator 23 hours 26 minutes ago

Thanks. Speaking of grooming, I just looked at our close incidents and I noticed that we have a number in the CMDB (hundreds actually) that are well past the 90 incident retention setting. They are definitely closed. Why are they not being removed from the CMDB and into the data warehouse?

I looked at Data Retention Settings and Incident retention time is 90 days. History is configured for 365 days though. I think this is the default. Where can I look to see why incidents are not being groomed properly?

Grooming should only occur on work items where Status = Closed and the Closed date is outside of your data retention settings.

You can also check the status of the grooming workflows with this sql statement:

SELECT * FROM dbo.InternalJobHistory WHERE TimeFinished IS NULL

If you get a lot of returns or see frequent failures, it could indicate that there is trouble in your data warehouse.  2 to 3 failures in a week is acceptable while 2 to 3 failures a day is not.

I did run that query against the CMDB and it came back with all zeros for a status code which I am hoping means it ran OK. I took a closer look at all our incidents marked Closed and we have 700+ that are older than 90 days. I guess next question is which job is responsible for cleaning up these incidents and moving them to the DW?

A zero indicates failure.  For instance, since the 6th, there have only been 24 Grooming jobs that failed in our SCSM instance.  How many failures are you seeing?

Here is the result of the query. Its important to know also that the regular data warehouse jobs that you configure via the console do not run. For some reason, the person that maintained Service Manager before me runs the ETL job and the MPSync just via a script that runs using task manager.

Well, that would indicate that your data warehouse is definitely in need of repair.

If you have time to spare on this, the following link is helpful to get you started.

http://blogs.technet.com/b/servicemanager/archive/2010/06/07/troubleshooting-the-data-warehouse-an-overview.aspx

Actually, I found out why the Closed Incidents are still in the CMDB. I found this blog below which explains the criteria for Incidents is 90 days AND last modified. I then displayed a Last Modified column in the view and sure enough all the Incidents I was referring to have been modified within the last 90 days which is why they are still in the CMDB. I have a couple that are close to their limit (Modified 5/23/2015) so in a few days, they should get purged.

I am a bit concerned though about the query I ran showing all zeros. I configured with the writer of the script that it does nothing concerning grooming and the article clearly states its just a couple of Management Packs that take care of grooming so its all a built in process based off of values you specify in the Data Retention setting in the console. Maybe I am running the query wrong? Our CMDB is stored on the Service Manager server so I am just selecting the Service Manager database and executing the query you gave me. Should I be doing it that way?

This part of the query, WHERE TimeFinished IS NULL, limits the returns to failed jobs.  Remove it and you'll see all the jobs.

OK, thanks for that. Big difference. Almost 1800 jobs with a status of 1. Only the ones above had zero but they were randomly scattered about. Not going to worry about them.