Aleksejs Spiridonovs

Official Docker container description mentions: https://hub.docker.com/r/sonatype/nexus/ : > _Mount a host directory as the volume._ This is not portable, as it relies on the directory existing with correct permissions on...

P.S. After taking a second look... This solution is still not solid enough. [Inotifyd](https://wiki.alpinelinux.org/wiki/Inotifyd), however, would serve the purpose perfectly. ~~Should I change this PR or create new (if reload...

Thank you, it is indeed poor-man process, an attempt to solves a use case that spawned a lot of wild workarounds with some standard approach (if PR would be accepted)....

That's what I thought. I'll come up with something we can document. Don't blame me, if it becomes popular! ;)

Exactly, scaling deployment will effectively stop pod from rescheduling, and from crashlooping. This should be applicable for deployment with multiple pods too (CrashLoopBackOff appears for majority cases due to internal...

P.S. Unfortunately, we cannot set `restartPolicy` to `never` for deployment, only supported value is "Always": https://github.com/kubernetes/kubernetes/issues/24725

Taking that this option have relatively **big downsides for any permanent setting (tail: 1000 - may overwhelm, as derailed said, tail: 100 - very often unusable due to large stacktrace)**,...

Feature request for update script also made here: #2695

Duplicate of #2664/#2694/#2695

Generalizing request: this is not only about namespaces, but all object types: ConfigMap, ServiceAccount, ClusterRole, ClusterRoleBinding, etc, etc, even if those are not under exemption.