Fix already existing random job name.

[Carreau]

Matthias Bussonnier @Carreau 09:52
Morning/Evening all, when running helm upgrade ... I get a
 Error: jobs.batch "pull-all-nodes-<mynamespace>-1" already exists
Any clue why, and what I can do ?

Matthias Bussonnier @Carreau 09:58
Response : kubectl delete jobs pull-all-nodes-<mynamespace>-1 --namespace=<mynamespace>
and job were list-able with kubectl get jobs --namespace=<mynamespace>

Yuvi Panda @yuvipanda 10:00
thanks @Carreau
that seems to happen when a helm install or upgrade is interrupted, or when you re-install into same namespace. should fix them both
can you open an issue in the helm-chart repo?

Matthias Bussonnier @Carreau 10:01
Yep. Seem to be the case when the docker image is huge. I got a timeout but the kubectl --namespace=<YOUR_NAMESPACE> get pod listed them as successful after a few minutes.

[yuvipanda]

This has been fixed, we now put the timestamp in there too.