Catch exception on system.health update check

Description

The system.health state information appears to be getting out of sync with the system after a period of time. In particular the "update" field/information never changes after that initial subscription - even though the system itself does actually report that updates are available via other mechanisms.

Chain of events:
1. Subscribe to the system.health events. NAS currently does not have any updates available ("update" : "UNAVAILABLE")
2. Wait XX hours depending on the update frequency of the train
3. The system.health events continue to report "update" : "UNAVAILABLE", even though other API calls or the NAS webUI both report that updates are available.

Problem/Justification

None

Impact

None

SmartDraw Connector

Katalon Manual Tests (BETA)

Activity

Show:

Bug Clerk November 8, 2019 at 4:25 PM

11.3 PR: https://github.com/freenas/freenas/pull/3776

William Gryzbowski November 1, 2019 at 4:24 PM

can you attach the Debug of that freenas, please?

Ken Moore November 1, 2019 at 4:00 PM

Confirmed: After leaving a system on the 11.3 nightly train for 3 days, the system.health subscription is still reporting updates as unavailable when updates have been available for almost 48 hours now. Appears that the subscription cache is not getting updated at all.

Ken Moore October 30, 2019 at 2:01 PM

Yes, my hunch is that the cached value for the system.health subscription is never getting updated at all - at least that is what is appears after inspecting all the events on Bonnie's system. I am running a couple tests with the 11.3 nightly train on my VM's now and should have verification by tomorrow for you as well.

William Gryzbowski October 30, 2019 at 1:52 PM

each subscription has its own cache.

To be clear, 24h cache is fine but somehow its not working? Do you have a debug of one of the FreeNAS that is not working? I am hoping for some clue in the log.