Fix Dashboard crash after exec into pod when using proxy

[floreks]

Related to #2029

Dashboard was crashing when trying to exec into pod. This was caused by missing tmp directory in the dashboard container. I have added /tmp mount that will be stored in memory on disk.

Tested on k8s 1.7.7. Exec works fine when accessing Dashboard using NodePort and kubectl proxy.

[codecov]

Codecov Report

Merging #2430 into master will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master    #2430   +/-   ##
=======================================
  Coverage   58.82%   58.82%           
=======================================
  Files         596      596           
  Lines       12896    12896           
=======================================
  Hits         7586     7586           
  Misses       5116     5116           
  Partials      194      194

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 64b8874...ff69951. Read the comment docs.

[maciaszczykm]

+1 to keeping it on disk rather than in memory.

[adrianchifor]

@floreks How can we view exec logs in the /tmp dir if we can't exec into the dashboard pod (no sh or bash) ? Can we maybe get an alpine based docker image?

[floreks]

It is only required to make exec feature work in dashboard. These files are created when you exec into pod (any pod) and deleted when exec session is closed.