1275607 - add OS X 10.6-10.8 deprecation rule when 48.0 ships to release

Status: RESOLVED FIXED

(Release Engineering :: Release Requests, defect)

Description

[[e-mail] (:bhearsum)]

When we ship 48.0, create two rules:
** 1) Matches version < 48.0 and OS X 10.6-10.8, and points them at the Firefox 48.0 blob.
** 2) Matches version = 48.0 and OS X 10.6-10.8, and points them at the OSX-10.6-10.8-Desupport blob.

The first rule needs to be updated if we ship a point release.

Comment 1

[Liz Henry (:lizzard Please n-i to RyanVM, jcristau, or pascal)]

Marking this as a 48 blocker (should be done before the end of july when 48 will ship)

Comment 2

[Liz Henry (:lizzard Please n-i to RyanVM, jcristau, or pascal)]

Ben,  or catlee, I just noticed this is still a blocker for 48 shipping, and is not assigned to anyone. Can you find an owner for it? Thanks!

Comment 3

[Rail Aliiev [:rail]]

Let's assign it to Callek, who is on the hook for 48.

Comment 4

[Justin Wood (:Callek)]

:lizzard, :bhearsum,

My read of this is that the update rules should be created *after* 48.0 ships...

If I am wrong please tell me and I'll adjust. (Looks like we want OSX 10.6-10.8 to be on 48, and once there 10.6-10.8 get a deprecation notice)

Comment 5

[Justin Wood (:Callek)]

self-ni so I can be sure to do finish this once the dep bug is done.

Comment 6

[Justin Wood (:Callek)]

Done

Comment 7

[Sylvestre Ledru [:Sylvestre]]

Implemented

Comment 8

[[e-mail] (:bhearsum)]

(In reply to Ben Hearsum (:bhearsum) from comment #0)
> When we ship 48.0, create two rules:
> ** 1) Matches version < 48.0 and OS X 10.6-10.8, and points them at the
> Firefox 48.0 blob.
> ** 2) Matches version = 48.0 and OS X 10.6-10.8, and points them at the
> OSX-10.6-10.8-Desupport blob.

I had to revert part 2 due to bug 1292165. We'll need to do it prior to shipping 49.0 instead.

Comment 9

[[SV Manager] Florin Mezei, QA (:FlorinMezei)]

Was able to confirm the 48.0 watershed when running ondemand update tests for 48.0.1 on release-localtest [1], and release-cdntest [2]. The watershed was bumped to 48.0.1 when we released officially so tests on the release channel [3] have passed as expected for versions <48.0.1.

[1] - https://treeherder.mozilla.org/#/jobs?repo=mozilla-release&revision=f36f7ace6f48&group_state=expanded&filter-tier=3&filter-searchStr=Fxup-release-localtest(
[2] - https://treeherder.mozilla.org/#/jobs?repo=mozilla-release&revision=f36f7ace6f48&group_state=expanded&filter-tier=3&filter-searchStr=Fxup-release-cdntest(
[3] - https://treeherder.mozilla.org/#/jobs?repo=mozilla-release&revision=f36f7ace6f48&group_state=expanded&filter-tier=3&filter-searchStr=Fxup-release(

Comment 10

[Nick Thomas [:nthomas] (UTC+12)]

tl;dr - the background rate was set to 0 on the rule for the desupport notice, so this hadn't been working as expected (ie only for manually initiated requrests).

12:01 <rstrong> nthomas: I am looking into the increase of no updates found on 48.0.2 and it looks like there is a large number of Mac users that are getting this.
12:02 <@nthomas> ok, I’ll go have a look at the rules
12:06 <@nthomas> rstrong: I only see the 10.6 - 10.8 deprecation watershed at 48.0.2, then a desupport notice. Do you have any more info on OS versions or so on ?
12:08 <rstrong> nthomas: I'm seeing 10.8.0 and other versions getting no updates found instead of the unsupported update xml for last week's dataset
12:09 <rstrong> and other weeks
12:09 <@nthomas> ok, I’ll mock up some urls
12:12 <rstrong> nthomas: I'm looking at a few of the records and for 48.0.2 the majority of no updates found is on Mac. I also see other versions such as 11.4.2 in the mix. Also, these are clients that have only seen no updates found when running Firefox 48.0.2
12:13 <rstrong> and had a subsession last week.
12:13 <@nthomas> 11.4.2 for OSX version ?
12:16 <rstrong> nthomas: that is what I see in telemetry. Row(name=u'Darwin', version=u'11.4.2'
12:16 <@nthomas> ok
12:17 <@nthomas> so https://aus5.mozilla.org/update/6/Firefox/48.0.2/20160823121617/Darwin_x86_64-gcc3-u-i386-x86_64/en-US/release/Darwin%2012.1.0/NA/default/default/update.xml should be a 48.0.2 request for 10.8.1, and indeed no update offered
12:20 <rstrong> nthomas: those should be unsupported messages... right?
12:20 <@nthomas> yeah, as the rules are configured. I’ll dig into it and look for exceptions/errors
12:20 <@nthomas> oh, the rate is set to 0
12:21 <@nthomas> with ?force=1 the desupport is offered
12:22 <@nthomas> that change was made 5 months ago though
12:23 <@nthomas> the client only shows the desupport once, right ? So there’s no need to limit the notice to manual check for updates
12:23 <rstrong> nthomas: correct
12:31 <@nthomas> rstrong: can’t find anything in bugs as to why, so I’ve set the rate to 100
12:31 <@nthomas> https://aus5.mozilla.org/update/6/Firefox/48.0.2/20160823121617/Darwin_x86_64-gcc3-u-i386-x86_64/en-US/release/Darwin%2012.1.0/NA/default/default/update.xml works now
12:35 <@nthomas> is this something that became obvious after we squashed other problems ?
12:37 <rstrong> nthomas: I wasn't planning on commenting. It became obvious in the app update out of date dashboard after 48 was considered out of date a couple of weeks ago.