bymodude
commented
8 years ago I believe I am encountering the same issue, I simply want to process jobs one at a time but wait for 5 seconds after each job of that type before processing the next one, so I have this example code:
Jobs.process('testJob', function(job, ctx, done) {
setTimeout(function() {
done();
ctx.pause(500, function(err) {
setTimeout(function() { ctx.resume(); }, 5000);
});
}, 1000);
});what happens is that when ctx.pause() is called it never completes the job in redis, that means I never receive the Jobs.on('complete') event even though the job obviously completed
if I dont call ctx.pause() directly after the done(), e.g I delay it by 1000ms, then I do get the event, so it seems as soon as ctx.pause() is called it will no longer update active jobs as complete when they are marked as done
for my use case I would already be happy with a hint on how to properly implement the pattern "process jobs one at a time but wait x seconds in between", but I do agree with OP that this behaviour of .pause() is not expected
behrad
Hi,
So pause and resume works in that when I have a queue of jobs and pause them, no other jobs in the queue are processed until I resume, perfect.
The problem is that the jobs that are "active" and successfully complete never get updated to "complete" and stay "active" even after resume.
My job queue processes 10 jobs asynchronously, while they're processing (I'm just using a timeOut for testing) I pause them and resume them and this is then my output.
After the "done!" output a "Job #x completed!" output is expected in the console but these jobs are not updated to have completed even though they do complete (as seen by "done!").
So I end up having lots of "active" jobs in my queue even though they have completed! This is surely not intended behaviour? And if it is, then it shouldn't be...
Any help on this issue? Thanks!