v1.1 known issues / FAQ accumulator

[dchen1107]

This issue is intended to accumulate known errata / issues / workarounds so we can produce a useful doc upon release. If you have a known issue that is worth putting in our 1.1 release notes, please add a comment here of the form "NEW: ..." and write a paragraph explaining the issue and the workaround or repercussions or whatever.

cc/ @thockin @bgrant0607

[lavalamp]

I thought we were going to scrape the cherry pick issues for this? Or is this meant for things that are known broken?

[lavalamp]

Known broken in 1.1: #15726

NEW:

Stdin redirection (pod exec, attach etc) does not work as expected between kubectl 1.0.x clients and Kubernetes 1.1 apiservers, as well as between kubectl 1.1 and 1.0.x apiservers. Workaround: use a kubectl that corresponds to your apiserver's version.

EDIT: It looks like we may be able to get a fix for this, a PR is in progress.

[dchen1107]

@lavalamp This is for known breakage in 1.1 release, not for cherry-picks.

[vishh]

#10959 will still be unresolved in v1.1.

[vishh]

Grafana automated dashboards for containers are at times not working.
A reload of the browser has resolved these issues in the past.
We suspect that the issues are related to access via proxy.
We have documented how to access Grafana directly using an external Service.

[dchen1107]

#15960: Mirror pods in a delete create loop in version skewed cluster.

[dchen1107]

#15175: docker pull can stall and leading cluster bringup issue. Restarting docker daemon should fix the issue.

[brendandburns]

Closing this since 1.1 is released. @dchen1107 re-open if you need this for some reason.

[bgrant0607]

This was never done

[bgrant0607]

Didn't see anything here worth documenting at this point. Might have been useful 3 months ago. Closing.