749359 - [tracker] deploy Django webapp on crash-stats.m.o

Status: RESOLVED FIXED

(Socorro :: General, task)

Description

[Lonnen :lonnen]

It's time to remake the webapp as a Django site.

The plan is to reimplement pages in Django, serving them from Django instead of PHP as they become ready, until the whole webapp is being served from django. Then we can delete the PHP webapp.

Use this bug to track individual pieces of the rewrite.

Comment 1

[Robert Kaiser]

If bug 788048 is about correlations not working at all in socorro-crashstats, I'd see that as a blocker of deployment as primary UI, not as a dependency on it.

Bug 778863, as eager as I am to have it, is the other way round, from what I can tell: It's already a bug on PHP, so not a blocker, but we'll only solve it on django, so a dependeny.

I'd see bug 850985 as blocker, but that WFM right now.

The sorting bugs all seem to be blockers to me: bug 850988 (only soft blocker), bug 850991, bug 845439 (I see that one is marked fixed, yay).

Bug 848497 is probably bad enough to warrant blocking as well.

Bug 849897 is a blocker for a numbers of devs.


And bug 821179 is a blocker of course (same as bug 844140, btw?), but I guess that blocks bug 788053 from working anyhow.

Comment 2

[Robert Helmer [:rhelmer]]

(In reply to Robert Kaiser (:[e-mail]) from comment #1)
> If bug 788048 is about correlations not working at all in
> socorro-crashstats, I'd see that as a blocker of deployment as primary UI,
> not as a dependency on it.

Correlations has been intermittently busted in general for a while due to bug 836671, have you seen specific cases where it works on old crash-stats and not new? Please add to that bug if so and we can block on it (I have a hard time reproducing correlations report behavior in general...)

Right now correlations are pulled via the old PHP app, so it should be working the same.

Comment 3

[Robert Kaiser]

(In reply to Robert Helmer [:rhelmer] from comment #2)
> Correlations has been intermittently busted in general for a while due to
> bug 836671, have you seen specific cases where it works on old crash-stats
> and not new?

I've seen that (i.e. working on PHP and not on django) at least when I tried in one case, but I'll need to take another look and find specific cases to link - of course, those usually are only trustworthy to be reproducible as long as it's the same day of data we're looking at...

Comment 4

[Robert Helmer [:rhelmer]]

Filed blocker bug 875992 for things that don't block turning django on, but do block turning PHP off (just correlations right now)

Comment 5

[Robert Helmer [:rhelmer]]

Changing the summary to avoid confusion - we can deploy the django app without turning off PHP, items that blocks turning off PHP are now in bug 875992

Comment 6

[Laura Thomson :laura]

We are ready to ship. Just need the php domain set up - bug 886604.