Join GitHub today
Fix inconsistency with "jid" on minion scheduled jobs and the returners output #47471
What does this PR do?
This PR proposes a fix to an inconsistent behavior happening at the moment with scheduled jobs and the returners output for these jobs.
Currently the "jid" which is included in the returner output is generated by the standalone job created on the minion, but it's not generated and stored on the Salt master. Salt does the following for scheduled jobs on minions:
This behavior produces inconsistent references to "jid" between the Salt event bus (which includes the "jid" generated on the master) and the returner output (with a different one generated by the minion) for the same Salt job.
Given this scheduled job:
I got this returner output:
While I get this on the Salt event bus:
Notice that the two
The proposed behavior is to included a new
and this returner output:
I'm targeting this PR to 2017.7 branch since this could be considered as a bug. Downside is that this would produce a change on the output schema of the job return event. Let me know if I should point this directly to
Once this PR is merged, a change on https://github.com/SUSE/salt-netapi-client/blob/master/src/main/java/com/suse/salt/netapi/event/JobReturnEvent.java to reflect the new attribute there.
What do you think about this proposal? Is this feasible or do you think we should use another approach?
Commits signed with GPG?
I am not a big fan of this solution, as it adds a field that will lead to confusion in the future. But we do need to figure out a solution here.
I spent some time looking into the root cause. The