uber / fiber

Distributed Computing for AI Made Simple

Home Page:https://uber.github.io/fiber/

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Possible issue with pi estimation example

ahutterTA opened this issue · comments

When running the pi estimation example in KIND the Job succeeds and correctly reports the value of pi but all of the workers finish with a status of Error with nothing in their logs. Looking at https://github.com/uber/fiber/blob/master/fiber/process.py#L297 it appears that it's possible to have an exit code of 1 but without err being set. Is this intentional?

@ahutterTA I have same issue here. And seems like poolworker pods can only be triggered when create Job in 'default' namespace ?

See #27 Q2