Fork me on GitHub

anybody have a tenancy id strategy for a job that runs every 1 hour? For example, should the tenancy id be different each time the job is run? I've had some issues in the past with reusing tenancy id for the same job on different runs so I was wondering if anyone had some experience dealing with this use case


@dbernal when you say "same job", do you mean resuming a job ? or submitting a new job ?


there should be no reason why you have to change the tenancy id


@lmergen In this case I was submitting a new job. I'm trying to see if there would be any issues with resubmitting a new job with the same tenancy id. In this case it's a batch job so it needs to run on a schedule. And also, in the instance that the job fails, could the new job reuse the tenancy id or will that cause problems


no, tenancy id is useful in case of multiple tenants, or when you want to update your code while using the same zookeeper cluster


(the last part is how i use it)


what kind of errors are you seeing if onyx does not pick up a second job submission ?


I'm not sure if it was a specific error but it had difficulty starting. It was kind of hard to debug but the problem was fixed after I deleted the znode for that tenancy id and without knowing specifically what Onyx does when reusing a tenancy id I assumed it got into a bad state. But I guess my assumption is that the znode stores state for the running job


Is it possible that the previous job never completed / wasn’t killed?


@lucasbradstreet yea, I think that was the case in this instance