Error: Missing lock for job 84 failed

[ujwaldhakal]

Hello i am using 2 node process one main and one for queue in supervisor with docker where i have set autorestart and start to true. When some job takes longer time I get . I was little confused on where to start and if anyone has similar issue that would be great.

A 2021-04-21T23:36:04.598667Z Error: Missing lock for job 84 failed
at Function.finishedErrors (/var/www/node_modules/bullmq/dist/classes/scripts.js:135:24)
at Job.moveToFailed (/var/www/node_modules/bullmq/dist/classes/job.js:207:41)
at runMicrotasks ()
at processTicksAndRejections (internal/process/task_queues.js:97:5)
at async handleFailed (/var/www/node_modules/bullmq/dist/classes/worker.js:220:17)
at async Worker.run (/var/www/node_modules/bullmq/dist/classes/worker.js:99:34)
A 2021-04-21T23:37:54.192317Z Error: Missing lock for job 85 failed
at Function.finishedErrors (/var/www/node_modules/bullmq/dist/classes/scripts.js:135:24)
at Job.moveToFailed (/var/www/node_modules/bullmq/dist/classes/job.js:207:41)
at runMicrotasks ()
at processTicksAndRejections (internal/process/task_queues.js:97:5)
at async handleFailed (/var/www/node_modules/bullmq/dist/classes/worker.js:220:17)
at async Worker.run (/var/www/node_modules/bullmq/dist/classes/worker.js:99:34)

[manast]

Probably due to this: https://docs.bullmq.io/guide/jobs/stalled

[ujwaldhakal]

@manast but there is no stalled flag in my current queue job .. yep my some of the background task take long time and they are processed even though i get this issue. And once i get this error the job will be still there and no one will proceed i am running sandbox processor that is running function in different file

[manast]

You should provide a minimal test case that reproduces the issue so we can help better.

[ujwaldhakal]

Ok sure i will provide with some codes that will replicate.Thank you !

[ujwaldhakal]

OptimalBits/bull#1591

I got it fixed by increasing the lockDuration from 30 sec to the 5 mins max and it is working great

[edisoncpp]

I'm getting this error after 30 minutes even if I set the lockDuration to 90 minutes, AND the worker has async requests and database calls in it freeing up the event loop.

[hgeldenhuys]

I suspect that when bull does the stalled check, Redis might not be available for a moment due to swapping out something, the timestamp is then null, which results in a 1970 date, so Bull thinks the job has been stalling for 53+ years.