Removing active/running jobs

[ferronrsmith]

Tried using the following to remove an active job. Is there a better way to do this ? I am still seeing the job in redis

    var q = queue.get(qName);
    return q.getJob(id).then(function (job) {
      job.releaseLock(job.lockKey()).then(function () {
        return job.remove();
      });
    });

[manast]

The code seems to be correct. If you can write a test case that fails I can look into it.

[ferronrsmith]

Ok will do

[ferronrsmith]

It's strange that i tried to release the lock, then delete, but shows errors in my stacktrace

[queue-manager-1]2017-03-30T05:56:25.229464043Z Unhandled rejection Error: Could not get lock for job: 3ebadf00-14c0-11e7-b2b5-91515b0b1873. Cannot remove job.

[queue-manager-1]2017-03-30T05:56:25.229507753Z     at /usr/src/app/node_modules/bull/lib/job.js:316:13

[queue-manager-1]2017-03-30T05:56:25.229518898Z     at tryCatcher (/usr/src/app/node_modules/bluebird/js/release/util.js:16:23)

[queue-manager-1]2017-03-30T05:56:25.229527773Z     at Promise._settlePromiseFromHandler (/usr/src/app/node_modules/bluebird/js/release/promise.js:510:31)

[queue-manager-1]2017-03-30T05:56:25.229537154Z     at Promise._settlePromise (/usr/src/app/node_modules/bluebird/js/release/promise.js:567:18)

[queue-manager-1]2017-03-30T05:56:25.229548847Z     at Promise._settlePromise0 (/usr/src/app/node_modules/bluebird/js/release/promise.js:612:10)

[queue-manager-1]2017-03-30T05:56:25.229555024Z     at Promise._settlePromises (/usr/src/app/node_modules/bluebird/js/release/promise.js:691:18)

[queue-manager-1]2017-03-30T05:56:25.229559872Z     at Promise._fulfill (/usr/src/app/node_modules/bluebird/js/release/promise.js:636:18)

[manast]

Does this happens all the time? since releasing the lock and then removing is not performed atomically, there may be another worker trying to lock the job during that small window. In which context are you tying to remove the jobs?

[ferronrsmith]

yea seems to happen all the time for me. I am trying to remove a running job. What happens is there are long running jobs that might get stuck without returning a status, I wrote the above code to allow developers to kill the active job to free the queue.

[manast]

why don't you try job.moveToFailed instead? You can use it giving it a reason as well:

job.moveToFailed(Error('job has been stuck for too long'))

It seems that the feature that you would like to have is some kind of TTL, where if a job takes more than the given time it is failed automatically. I will consider it for an upcoming release.

[manast]

Feature request: #479

[ferronrsmith]

I do realize that then I moveToFailed, the queue doesn't automatically start pending jobs, is this expected behavior ?

[pariola]

Please any update on removing all jobs?

[dhamaniasad]

One way to do this is that you could spawn child processes from your main process that implement Queue.process(), and hold a reference to the child process in memory. When you want to kill active jobs, just kill the child process. You could use a worker for each queue, or a single worker for all queues, or any combination.

const execa = require("execa");

let filePath = path.join(__dirname, "./worker.js");
worker = execa(`node`, filePath);

// kill job
worker.kill();

[Torsten-ha1]

A simple way of cleaning all queues, freeing them from all jobs is to use following class:

export class CleanQueues {
	constructor() {}

	private getKeys = async (q) => {
		const multi = q.multi()
		multi.keys('*')
		const keys = await multi.exec()
		return keys[0][1]
	}

	private filterQueueKeys = (q, keys) => {
		const prefix = `${q.keyPrefix}:${q.name}`
		return keys.filter((k) => k.includes(prefix))
	}

	private deleteKeys = async (q, keys) => {
		const multi = q.multi()
		keys.forEach((k) => multi.del(k))
		await multi.exec()
	}

	/**
	 * @param queue
	 */
	emptyQueue = async (q) => {
		const keys = await this.getKeys(q)
		const queueKeys = this.filterQueueKeys(q, keys)
		await this.deleteKeys(q, queueKeys)
	}
}