| dholbach | good morning | 07:34 |
|---|---|---|
| dpm | dholbach, davidcalle, I followed the conversation on the community team channel, and this is something I actually wanted to discuss a while ago - could we document the way we use branches (and the deployment procedure) somewhere in a doc? Perhaps in a README.md file on the spec branch or on the d.u.c branches | 09:05 |
| davidcalle | dpm: the current - actually impossible to do - way is documented in deploy.md | 09:09 |
| davidcalle | dpm: so yeah, this will be replaced | 09:10 |
| dholbach | cool | 09:11 |
| dpm | awesome, thanks davidcalle | 09:14 |
| mhall119 | davidcalle: what info do you need from me? | 13:18 |
| davidcalle | Hey mhall119 :) I've found the RT from january and it was revno 162 (hence the staging branch), so I guess we should keep it this way and request a deployment of the same branch, what do you think? | 13:38 |
| davidcalle | (which is actually what is deployed on staging: (lp:staging/164) | 13:39 |
| mhall119 | davidcalle: we don't deploy to production from branches, only from the tarball in swift | 13:41 |
| davidcalle | mhall119: we don't pull anything at all? | 13:44 |
| mhall119 | nope | 13:45 |
| mhall119 | they just do a `juju set build_label=$version` and it pulls the tarball out of swift | 13:45 |
| davidcalle | mhall119: I was expecting something to be pulled when this config is changed :) Great | 13:46 |
| mhall119 | so the production branch should be identical to what's in production, but it's not the source of production deployments itself | 13:47 |
| davidcalle | Ok then. /me starts to file the RT | 13:50 |
| davidcalle | mhall119: does the following steps seem clear to you for an RT? (end of doc) http://pad.ubuntu.com/zdvmD4Ekb6 | 14:18 |
| mhall119 | starting at line 68? | 14:19 |
| davidcalle | mhall119: yes | 14:20 |
| mhall119 | hmmm, does piping yes to a manage.py command work? Does it terminate when the manage.py command exits? | 14:21 |
| davidcalle | mhall119: it returns "yes: write error" but actually works | 14:23 |
| mhall119 | ok | 14:23 |
| mhall119 | otherwise it looks fine, just make sure they have a db backup of the last stuff before starting on this | 14:24 |
| davidcalle | mhall119: that's the first step ;) Maybe I should make step 2: MAKE SURE YOU HAVE BACKED UP THE DATABASE | 14:25 |
| davidcalle | Alright, let's explode prod ;) | 14:26 |
| mhall119 | \o/ | 14:30 |
| mhall119 | burn it down! | 14:31 |
| * dholbach crosses fingers | 14:44 | |
| davidcalle | dpm: mhall119: dholbach: you are CCed on the RT ;) | 14:47 |
| * davidcalle throws RT over the wall and runs | 14:47 | |
| dholbach | yep, saw it :) | 14:47 |
| mhall119 | just saw it | 14:47 |
| mhall119 | now run! | 14:47 |
| mhall119 | just kidding, it'll probably take a few days before anybody tries to run it | 14:47 |
| davidcalle | mhall119: too late I've wished a happy deployment day to Daniel this morning, they don't have a choice :) | 14:48 |
| dpm | great | 14:48 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!