Jobs for no reason stuck in pending status

[klinux]

I have workspaces that stuck in pending status.

The api server has this message in the logs:
Job 348 is waiting for previous jobs to be completed

Theres no others jobs running or pending before 348:

Theres some caches or others things that can lock the execution jobs?

Thank you

[alfespa17]

Hello @klinux, that one is a bug in version 2.19.0, you can check this comment

We foget to add the status cancelled in this validation

terrakube/api/src/main/java/org/terrakube/api/plugin/scheduler/ScheduleJob.java

Line 86 in 191689a

Optional<List<Job>> previousJobs = jobRepository.findByWorkspaceAndStatusNotInAndIdLessThan(job.getWorkspace(), Arrays.asList(JobStatus.failed, JobStatus.completed), job.getId());

As a temporary fix you could go to the database and check if you have some status as cancelled for a particular workspace and change that to failed, I will release a hotfix later today

[klinux]

Thank you @alfespa17

[klinux]

@alfespa17
I think that I have other problem here, I updated all jobs em steps to failed instead cancelled, but the stuck job still here, can I look to other thing?

[alfespa17]

Check all the previous jobs from a particualr workspace and check the status to see if all are in faild/completed or any other state

[klinux]

You 're right, theres some columns as rejected status. The workaround works.

[alfespa17]

The job status issue was fixed in 2.19.1, the release is running the container images will be ready in a few minutes