897152 - tracking bug for automatic reimaging -> production

Status: RESOLVED FIXED

(Infrastructure & Operations Graveyard :: CIDuty, task)

Description

[[e-mail] (:bhearsum)]

That is to say: this bug tracks all the work involved to get us to a point where a slave will automatically go back into production after being reimaged. The big points are:
* Double check the hostname (machines should come back from re-image with the correct hostname, but we should bark if it's not set. we should explicitly _not_ try to set it ourselves.)
* Secrets management for Puppet-based machines (ssh keys, google api key, passwords, etc.)
* Ability to detect when a slave has been re-imaged, and thus requires re-enabling. (probably depends on the new multistate flag we've talked about; probably done in kittenherder)
* Watching to make sure the slave attaches to a master, and closing the problem tracking bug (probably done in kittenherder)

Comment 1

[[e-mail] (:bhearsum)]

I probably won't be getting to this anytime soon. Bug 792836 makes things a lot less painful though.

Comment 2

[Chris Cooper [:coop] (he/him)]

Re-imaged machines now go directly into production, provided they are enabled in slavealloc.