Some time ago, my customer set up policy actions for the ILMT “run capacity scan & upload results” fixlet by operating system, so as new clients are added and the scanners are installed, the scans would begin running automatically. These actions have been running uneventfully for quite some time. The capacity scan is scheduled to run every 30 minutes.
We recently noticed a moderate number of failures that occurred on some of the clients. The policy action is defined, when an error occurs, to try 3 times waiting a day between attempts. Evidently this failure occurred 3 times so the fixlet stopped running on these clients. The status of these clients now reports as “< error >” in the policy action results.
To remedy the situation, I suggested that an individual task be run on the failed computers to perform a one time capacity scan and upload results. This one time capacity scan ran successfully for the vast majority of clients.
My question is - for the original policy actions, should the status now become “Completed” for clients that now have a successful capacity scan? I would expect that the policy action would resume on these computers now that we have a successful capacity scan, and we should see the capacity scan kicking off every 30 minutes, correct?
After receiving data from the manually-run fixlet, you should see Complete agent on the server.
Unless there is some problem related only to running the fixlet within the schedule, or other external factors, you can expect it to start to work OK now - it is still configured to run at set intervals, so even after retry errors, at next scheduled scan everything should be OK.
It should be the case that if the action has failed the maximum number of times without a success then it will no longer run. This should be true even if a separate action that is identical is applied successfully, because that success shouldn’t apply to the original action and wouldn’t reset the number of failures.
It turns out that running the “Run Capacity Scan & Uploads Results” fixlet once and getting it to succeed was the resolution. Once this fixlet succeeded, it started running every 30 minutes again because it became relevant again after it was run once. The policy action document “saw” this and reported on the status correctly the next time it ran (through the policy action).
That makes sense. It was stuck, but all of a sudden the relevance evaluation of the policy action changed due to the other action, so then it started again. This happened because it was set to reapply when it becomes relevant again instead of reapply while waiting.
If you use reapply while waiting with policy actions like this, then it will attempt to run again after the delay as long as the relevance is still TRUE.