Missing key error on updateData

[gee4vee]

Suddenly I am getting this error when updating the data in a job using updateData. This was working recently with no code changes.

`Error: Missing key for job repeat:609e5fbb-95f2-4977-bb73-e32cc176c666-location-us-south-repeat-every-1-MINUTE:1750344240000. updateData
    at Scripts.finishedErrors (/Users/gvalenc/git/epx/projects/packages/jobs-manager/node_modules/bullmq/dist/cjs/classes/scripts.js:1100:24)
    at Scripts.updateData (/Users/gvalenc/git/epx/projects/packages/jobs-manager/node_modules/bullmq/dist/cjs/classes/scripts.js:367:24)
    at process.processTicksAndRejections (node:internal/process/task_queues:105:5)
    at async QueueEvents.<anonymous> (/Users/gvalenc/git/epx/projects/packages/jobs-manager/scheduler/async.js:182:25)`

My code is simply replacing the value of one of the properties in the data of job.

                        const jobRunObj = jobRun.getExternalObject();
                        const updatedData = JSON.parse(JSON.stringify(job.data));
                        updatedData.projectJobRun = jobRunObj;
                        await job.updateData(updatedData);

Has something changed in BullMQ that might be causing this?

[gee4vee]

I am getting the error with Redis 6.2 but not 7.2. Is this expected?

[roggervalf]

hi @gee4vee, that validation is pretty old since v1.68.1. You should make sure to execute this method while your jobs is not removed from Redis

[gee4vee]

@roggervalf since my code is exactly the same and only breaks on the older Redis version, I don’t think I’m unintentionally removing the job. Can you shed some light on why the older version behaves differently?

[manast]

The error is telling you the key holding the job is not there anymore. Check your Redis config, or it could be that before it worked if due to asynchronicity your code was able to update before the job was removed. We do not have any more info to give based on the code you have posted. Isolate the problem and then we can look further into it.

[gee4vee]

@manast I’m not sure what I’m missing here. The only variable that is changing is the Redis version. Is there a particular config I can check in the 6.2 instance that could cause this?

[gee4vee]

@manast If it’s a synchronicity issue, why is it consistently working in 7.2 and consistently failing in 6.2 with the same code?

In this case, I’m scheduling a job to run on a one minute or longer interval, but the job is almost immediately gone when the missing key error occurs.