Closed tiborsimko closed 4 years ago
"see another ticket" above is https://github.com/reanahub/reana/issues/253
and, if for some reason reana-server cannot start it automatically -- e.g. because the message was lost; see another ticket on this -- then the end-users cannot trigger the starting action via
This doesn't happen anymore, see https://github.com/reanahub/reana/issues/253.
I think that what is left in this ticket is to return a better message, saying something like "your workflow is already queued for execution" or similar.
Sometime the workflow is in a queued state:
and, if for some reason
reana-server
cannot start it automatically -- e.g. because the message was lost; see another ticket on this -- then the end users cannot trigger the starting action viareana-client
because of:Currently, only cluster admins can unblock the situation by setting workflow status to
created
and issuing thestart
command.What we could do is to allow end users to do
start
command and on the reana-server endpoint side check the queue and if everything is OK and the job is properly waiting for execution, then do nothing (=as is the case now); however if there is something wrong and the job is still queued even though it could run, the reana-server endpoint could accept thestart
command (=new behaviour).