Closed
Bug 1238802
Opened 8 years ago
Closed 8 years ago
Re-run e10s A/B experiment on Beta 44
Categories
(Firefox :: General, defect)
Tracking
()
RESOLVED
WONTFIX
People
(Reporter: Felipe, Assigned: Felipe)
References
Details
Attachments
(3 files)
We should re-run the e10s A/B experiment on Beta 44 for the remainder of Beta. More precisely until Jan 26th according to Blassey in #e10s. Also should make sure that the minbuildid includes bug 1236754.
Assignee | ||
Comment 1•8 years ago
|
||
I've checked that bug 1236754 made it to beta7, which has build id = 20160107144911 I'll do some sanity checks with the experiment on this build now
Summary: [meta] Run an e10s A/B experiment on Beta 44 → Re-run e10s A/B experiment on Beta 44
Comment 2•8 years ago
|
||
Should we get RyanVM's team, who took over QAing telemetry experiments, to take a quick look to make sure the experiment is still working as expected with these changes? From the looks of bug # 1236754, the following prefs should be enabled when someone already has e10s enabled or is placed into the e10s experiment group correct? >> extensions.interposition.enabled >> extensions.interposition.prefetching >> dom.compartment_per_addon >> dom.ipc.plugins.asyncInit.enabled
Assignee | ||
Comment 3•8 years ago
|
||
New experiment patch. Only the add-on ID and manifest values (endDate, minBuildID) were changed.
Attachment #8707001 -
Flags: review+
Assignee | ||
Comment 4•8 years ago
|
||
Hey Jason, could you sign this new experiment xpi?
Attachment #8707003 -
Flags: feedback?(jthomas)
Assignee | ||
Comment 5•8 years ago
|
||
(In reply to Kamil Jozwiak [:kjozwiak] from comment #2) > Should we get RyanVM's team, who took over QAing telemetry experiments, to > take a quick look to make sure the experiment is still working as expected > with these changes? > > From the looks of bug # 1236754, the following prefs should be enabled when > someone already has e10s enabled or is placed into the e10s experiment group > correct? > > >> extensions.interposition.enabled > >> extensions.interposition.prefetching > >> dom.compartment_per_addon > >> dom.ipc.plugins.asyncInit.enabled Yeah that would be nice. When we get the signed xpi I'll push the new experiment to the staging server for some testing
Comment 6•8 years ago
|
||
Signed experiment.xpi
Updated•8 years ago
|
Attachment #8707003 -
Flags: feedback?(jthomas) → feedback+
Comment 7•8 years ago
|
||
> Yeah that would be nice. When we get the signed xpi I'll push the new
> experiment to the staging server for some testing
RyanVM, would your team be able to look at this once everything is ready? I can provide you with the test cases I originally used when we launched this experiment.
Flags: needinfo?(ryanvm)
Assignee | ||
Comment 8•8 years ago
|
||
Pushed https://hg.mozilla.org/webtools/telemetry-experiment-server/rev/ab54b1891657 In some moments the new experiment should appear alongside the URLBar experiment at the staging server https://telemetry-experiment-dev.allizom.org/
Comment 9•8 years ago
|
||
I can confirm extensions.interposition.enabled(workaround) kicks in now on beta. CanvasBlocker add-on giving 30s hangs (most prolific I have experienced.) Bug 1196975 breaks the add-on in 45.
Updated•8 years ago
|
Flags: needinfo?(ryanvm)
Felipe, as per the decision on email, we have decided to postpone this experiment to the start of Beta45 cycle. Beta8 - RC week does not seem like the place to start any experiments especially if it requires code changes to the binaries we ship (pref changes in this case). Could you revert the changes that were made? Thanks!
Flags: needinfo?(felipc)
Assignee | ||
Comment 11•8 years ago
|
||
Backed out the experiment in https://hg.mozilla.org/webtools/telemetry-experiment-server/rev/7b8a1908c85e I believe the prefs do not need to be backed out because they won't affect non-e10s, but I'll let that be figured out in bug 1238802.
Status: ASSIGNED → RESOLVED
Closed: 8 years ago
Flags: needinfo?(felipc)
Resolution: --- → WONTFIX
Assignee | ||
Comment 12•8 years ago
|
||
(In reply to :Felipe Gomes (needinfo me!) from comment #11) > I believe the prefs do not need to be backed out because they won't affect > non-e10s, but I'll let that be figured out in bug 1238802. I mean bug 1236754
(Is this resolved wontfix (wontdo?) If you want heartbeat or other survey stuff, come talk to us :) )
You need to log in
before you can comment on or make changes to this bug.
Description
•