Just been playing around with the HA command line and found hundred of supervisor jobs in the jobs list. For example:
ha jobs info | grep supervisor_check_connectivity | wc -l
This returns a count of 430, each job looks like this
- child_jobs: []
done: null
name: supervisor_check_connectivity
progress: 0
reference: null
stage: null
uuid: <some hex string for the id>
Anyone else have this…, why so many…?
Ah damn. So those are all actually the jobs that never ran. The way it works is all jobs have conditions and limits on when they can run and how often. For instance, supervisors connectivity check is throttled like so:
It is only allowed to run if the last check failed or it’s been at least 5 minutes since the last success.
However there’s a bug it seems. Jobs which are rejected due to limits or conditions are not being properly cleaned up. That’s why done is null on all of them, it becomes false when they actually get started.
Can you submit an issue here for this? I’m traveling currently but will fix it when I get a moment. Thanks for reporting this
Thanks Mike, reported here