When jobs.process was interrupted by errors, the job still in active state

[jason-son]

So how to clear those jobs state,which I wanna put them back to the Queued state.
If there could be some action in web UI is great, and API is also ok.
And also for the failed jobs.

[tj]

hmm there is a API for max attempts, however that defaults to 1 so if you get an error it should stay in the failure state, is that not happening?

[jason-son]

Actually,I use ctrl+c to interrupt the jobs.process scripts. So the states always to be active.
and, I wanna to run another script to attempts the failure job,but don't know how to do.

[jason-son]

Got the state change

PUT '/job/:id/state/:state'
job.states()

Job.get(id, function(err, job){
if (err) return res.send({ error: err.message });
job.status(state);
job.save(function(err){
if (err) return res.send({ error: err.message });
res.send({ message: 'updated state' });
});
});

BTW, if the web UI has some action to change the state, that would be nice ~

update! found that ! just click the every job states in the web UI,which you can get selections;

[tj]

yup, it's a little obscure since it's just the text, I should probably make that more obvious lol

[Qard]

Is it possible to somehow change the state of a job automatically if it's been "active" for a long period of time? I have kue running a thousand jobs at any given time. If I interrupt it, those jobs stay in the queue forever. My host kills and restarts my instances regularly, so I end up overflowing redis after about a day and have to manually clear it.

[tj]

not with kue's api really, you'd have to do some direct redis stuff, or do some cron job style lua scripting etc

[Qard]

What about having another job run every hour or so that searches queued jobs with the active state and checks the updated_at field? Would that be reliable, or could updated_at be changed by other things? I assume the progress field would change it, but if no progress has been made, the same issue probably exists.

[tj]

I dont think we have progress updating it ATM, I cant remember, but either way yeah that should be a good indication of a stuck job