OpsMgr SDK Service Fails to Start with Event ID 26325

If your Operations Manager suddenly appears to have stopped functioning, your first place to check is likely going to be the services for Operations Manager.

Ops Manager is fairly picky about its environment and sometimes the smallest of changes can render the system useless.

During my most recent OpsMgr incident, my Data Access Service was basically useless to me. It would start, then about five seconds later, would stop. The Operations Manager Application Log (Event Viewer -- Applications and Services Log -- Operations Manager) was showing me a bunch of Error messages with EID 26325.

Log Name:      Operations Manager
Source:        OpsMgr SDK Service
Date:          7/8/2015 11:00:37 AM
Event ID:      26325
Task Category: None
Level:         Error
Keywords:      Classic
User:          N/A
Computer:      SCRP-SCOM12MGT1.spang.local
Description:
An authorization store exception was thrown in the System Center Data Access service. Exception message: Unable to perform the operation because of authorization store errors.

Event Xml:
  
    
    26325
    2
    0
    0x80000000000000
    
    13112630
    Operations Manager
    SCRP-SCOM12MGT1.spang.local
    
  
  
    Unable to perform the operation because of authorization store errors.
  

This definitely isn't what we were looking for!

As it turns out, the OperationsManager database on our operational SQL server is unable to have more than one db_owner. I had this issue once before, but didn't document the resolution, so I'm taking this chance to jot down the solution!


First, let's check out whether or not we have more than one db_owner on the Operations Manager database.

Connect to your SQL Server instance and right click the OperationsManager database and select to create a New Query against it. Type in the following query and press Execute (or F5 for those hotkey users out there)


select user_name(member_principal_id) from sys.database_role_members where role_principal_id = user_id('db_owner')

The output will likely have two entries in it. My broken instance looked like:



Definitely more than one entry that is a db_owner on OperationsManager. Go into the Security -- Logins page and modify (in my case) the AppaSA account. Select the OperationsManager database and de-select db_owner.



Now, when you re-run the query, the result should only show one user as being a db_owner.



Great! Now I'll remote back into my OpsMgr Management Server and attempt to start the Data Access Service. In my case, the service started like it had before, but continued to run. No more random stopping for me! We should confirm this in Event Viewer, just to be safe, though.



And as a final confirmation, I'm able to successfully connect into my Operations Manager instance! Great!

I sure hope that this is helpful for someone in the future in diagnosing their Operations Manager instance!

Comments

Popular Posts