Open tanganellilore opened 3 weeks ago
hi @fosterseth , reformat as per discussion above.
@fosterseth with external and unstable execution node (i rebooted it multiple times) this happes one mor time, and with my last commint I should cover all things
SUMMARY
In case we have some issue beetween execution node and AWX, and AWX will not catch that execution node is not working well or nor reachave or simply delete workunit (I don't identify exactly the use case but appen to me in 24.2.0 with execution node and ansible runne 1.4.3), workflow still wait the running state. if we try to cancel the job/workflow via UI, we receive error below on awx-task pod and job never cancelled/stopped.
In thi PR i simply try/except the for cycle and demand the release to workunit reaper, where I put the
force-release
command instead of simplerelease
.I think that we need to force the release inside the for-cycle, because administrative_workunit_reaper check a lot of things on work unit side, that to me is not much sense because we already filter by ACTIVE_STATES on UnifiedJob filter.
If this is true, i can change it adding a force-relase command on exception in that way we are shure that works will be relased when cancel will be clicked on UI.
ISSUE TYPE
COMPONENT NAME
AWX VERSION
ADDITIONAL INFORMATION