You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
[scheduling]
[[graph]]
R1 ="foo => a & b & c"
[runtime]
[[foo]]
script="sleep 360"
[[a, b, c]]
script=falseexecution retry delays=PT5S
Run this sucker, and manually trigger a while foo is still running:
a will run, fail, and return to the waiting state to wait on its retry timer
Then on master:
a will not run job 02 immediately when the retry timer is done
(it is waiting on foo:succeeded ... manual trigger of job 01 means "run now regardless of prerequisites")
a will run job 02 immediately when the retry timer is done
(manual trigger satisfies the task's "off-flow" prerequisites)
I think the upcoming new behaviour is correct (although you can possibly argue it both ways). But it's certainly something we should document for users, because it could be surprising. I was surprised by the current behaviour on master, myself, until thinking a bit!
Question: do we agree the upcoming new behaviour correct?
If yes, close this issue after merging #6395 and documenting this aspect of it.
The text was updated successfully, but these errors were encountered:
Run this sucker, and manually trigger
a
whilefoo
is still running:a
will run, fail, and return to the waiting state to wait on its retry timerThen on master:
a
will not run job 02 immediately when the retry timer is done(it is waiting on
foo:succeeded
... manual trigger of job 01 means "run now regardless of prerequisites")Or, after #6395
a
will run job 02 immediately when the retry timer is done(manual trigger satisfies the task's "off-flow" prerequisites)
I think the upcoming new behaviour is correct (although you can possibly argue it both ways). But it's certainly something we should document for users, because it could be surprising. I was surprised by the current behaviour on master, myself, until thinking a bit!
Question: do we agree the upcoming new behaviour correct?
If yes, close this issue after merging #6395 and documenting this aspect of it.
The text was updated successfully, but these errors were encountered: