CI

Andrew and I were talking about ways of confirming a chacra node is safe to perform maintenance on other than

1. touch /tmp/fail_check
2. Wait an hour or confirm no running builds are slated to push packages to the node by checking build console logs in Jenkins

He suggested:

<andrewschoen> what we could do is have shaman ping chacra back when chacra registers to verify it is actually healthy
<andrewschoen> then all you'd need to do is touch /tmp/fail_check && sudo service chacra restart
<andrewschoen> and it'd be out of the shaman pool