* use ec2 runner for kube acceptance tests
* add missing newline
* log outputs
* add user and home env vars
* log more
* use root user
* fix ec2 runner
* remaining debugging
* try overwrite forcing
* fail fast
* get kubectl location
* even more
* just look for it
* see if it's a symlink
* symlink
* make sure it's started
* try with overrides
* Revert "try with overrides"
This reverts commit 123e3c033e.
* clean up
* describe pods
* display exception when getting address in use error
* retry installing socat
* try inet4 address specifically
* switch order of install
* use unique ports for each
* try to detect locations with home and user set
* STOPTTTTTTTTTTT
* fix typo
* move socat back up one more
* add update