SMA connector occasionally fails to start runbooks.

Adam_Dzyacky

Curious if anyone else is running into this issue wherein very, very, very occasionally the SMA connector fails at kicking off the runbook. The result returned in the runbooks is always the following:

Execution of the job failed.
The 'Start-SMARunbook' command was found in the module 'Microsoft.SystemCenter.ServiceManagementAutomation', but the module could not be loaded. For more information, run 'Import-Module Microsoft.SystemCenter.ServiceManagementAutomation'.

The error alone seems to point to the PS module being at fault, but the fact it works an overwhelmingly large majority of the time only further clouds the troubleshooting here.

Adam_Dzyacky

Is there any verbose logging that could be enabled for this connector beyond the above output?

Leigh_Kilday

I've had no experience with SMA, but I've had SCSM/SCORCH issues recently. With any luck this is also relevant to your instance.

1. "Monitor" type activities in SCORCH can have filters, but are limited to the object properties and no relationships. This causes it to identify a large quantity of objects and proceeds to process them through the activities. Sometimes this payload is so large that the runbook hasn't finished processing before the next payload has begun. The CPU on the Primary Workflow server cannot keep up and this results in scripts partially completing.
2. Connections to the Primary Workflow server are limited by the MaxShellsPerUser setting. We've had to increase this value to stop the connections from being dropped. Without error handling in the remote session script, the connections are dropped and the runbook itself doesn't care. With error handling, these remote session scripts queue up and cause delays.
3. Runbooks can go "Missing" for no logical reason. The default sync time is midnight and there's something preventing certain runbooks from being found at this time. I've implemented a scheduled task to check for missing runbooks and sync the connector if any are found. There's a period of about 10 minutes where synchronising the connector doesn't resolve the issue, but any other time of the day it works perfectly. We are still investigating this...

Adam_Dzyacky

Actually @Leigh_Kilday, I believe point number 3 of yours I think was fixed in a UR release. But if you're on the latest version across the board, I'm clearly in error.

I wonder if the best way to go about this (although I'm not even sure what i'm looking for at this point) is procdumping this, and even still if that would reveal anything than the above error message.

Leigh_Kilday

Thanks @Adam_Dzyacky, we haven't applied UR9 yet. Between the reissue and us going live with SCSM/Cireson we forgot about it.

Adam_Dzyacky

To any onlookers, this has been acknowledged as a bug by Cireson with a fix in the works.