Upstreaming question

[jhohertz]

@inve1 I have a few little fixes as I synced up my fork of buri, wanted to ask you something.

C* was disabled/drained on the "not image_build" case, which makes the Vagrantfile / all_in_one_* dev VMs not bring up C*, i've just reverted that locally for now, but in the interest of upstreaming some fixes, what can we do here? Thinking add a boolean for whatever you want to call the behaviour you set in the local dir? (cassandra_drain_on_deploy?).

Thoughts?

[inve1]

yeah that stuff is a little messy now, the purpose of drain there is for upgrades [which is basically the only use case of not image_build for us]. I forget why I commented out the stuff that starts cassandra back up.

A switch sounds good to me if starting cassandra back up is not viable for whatever reason [I'll look into that tomorrow]. otherwise having it start at the end of that role should solve your problem?

[jhohertz]

Yup, I might still make a flag in the inverted sense then, cassandra_skip_upgrade_drain, and apply it just to the VM/Image case. I see why you'd want it for in-place updates.