726575 - Setup dev/stage/prod masters for Releng APIs/Tools

Status: RESOLVED INVALID

(Infrastructure & Operations Graveyard :: WebOps: Other, task)

Description

[Mike Taylor [:bear]]

Need to spin up a dev vm for the new Releng Dashboard/API

requirements are being stored here

https://mana.mozilla.org/wiki/display/websites/RelEng+Dashboard

Comment 1

[Amy Rich [:arr] [:arich]]

Adding a couple ccs (laura and corey) to keep other folks in the loop.

Releng is creating a new mobile dashboard that will integrate with the work Laura's folks are doing to make an overall releng/metrics dashboard.  This will be critical infrastructure (tree closer) for releng and should be supported by infra (like other web sites) and conform to the dev/staging/prod methodology that corey's group is requiring.

This vm request is to create the dev environment for that service.

Corey, how deeply do you want to be involved at this point?  I can spin up a RHEL 6 vm on the ganeti cluster in mountain view (where all of the mobile stuff currently is).  Or would you prefer to spin something up elsewhere?

Comment 2

[Laura Thomson :laura]

Another point to note is that in addition to the API, there will also be a webapp.  Should we use the same dev VM for everything?  I'm happy with that for dev but would probably want separate boxes as we proceed to staging, in order to more closely mirror a production environment.

Comment 3

[Mike Taylor [:bear]]

(In reply to Laura Thomson :laura from comment #2)
> Another point to note is that in addition to the API, there will also be a
> webapp.  Should we use the same dev VM for everything?  I'm happy with that
> for dev but would probably want separate boxes as we proceed to staging, in
> order to more closely mirror a production environment.

I'm ok with one vm for dev knowing that we will need to test on multiple for staging

Comment 4

[Amy Rich [:arr] [:arich]]

After discussion on irc, it seems like the best course of action is to hook this into the existing web properties dev/staging/prod puppet configs and cluster setup.  Reassigning to serverops for continued work.

Comment 5

[Lukas Blakk [:lsblakk] use ?needinfo]

We're going to run autoland on this setup as well so we can share the dev/stage/prod envs with briar-patch and future developer tools releng implements.  We already have a puppet script ready for what needs deploying on the masters, Marc can provide those to whomever is setting these up.  Adding in the autoland project here bumps the urgency of getting these machines ready since we're on a Q1 deadline and we really need to get these environments online so current try-only production can start running from a puppetized master and so that Marc can continue his work on staging branch landings from the dev/staging machines.

Comment 6

[Mike Taylor [:bear]]

closing this bug - it morphed into many paths and after todays meeting with dustin and webdev and webops we have a new roadmap

new bugs are being created for autoland staging/production and releng dashboard staging/production