Bugzilla

Andrew McCreight [:mccr8]

Comment 2

•

4 years ago

Yes, sorry, I meant to file a follow-up for this and got distracted! I believe this is a bug with Fission that the update to the test has caught. I can't (personally) see a way that the test is doing something wrong here.

We seem to have told the content process to quit, and it is hung and not quitting. The change to the test simply adds two out-of-process iframes to the tresize test. There's nothing that I can think of in that that would cause a content process to not exit.

Andrew, is there any chance you could take a brief look at this and confirm, or kick this to someone who can?

Flags: needinfo?(dothayer) → needinfo?(continuation)

Comment 3

•

4 years ago

Yeah, that does seem like it is more likely the fault of Fission than anything else. I think the first step to investigate this would be see what happens in a debug build. The crash stacks don't look very useful. Unfortunately, a lot of the Fission team is going on PTO soon, so maybe you could just back this out, or at least the iframe part (which is less likely to bit rot), and somebody can look at it when we get back? Thanks.

Fission Milestone: --- → ?

Flags: needinfo?(continuation)

Andrew McCreight [:mccr8]

Comment 4

•

4 years ago

I'll post this in the Fission IRC channel and see if anybody has some free time today or tomorrow to look at it.

Chris Peterson [:cpeterson]

Comment 5

•

4 years ago

Tracking Talos crash for Fission Nightly (M6)

Fission Milestone: ? → M6

Andrew McCreight [:mccr8]

Comment 6

•

4 years ago

(In reply to Andrew McCreight [:mccr8] from comment #3)

Yeah, that does seem like it is more likely the fault of Fission than anything else. I think the first step to investigate this would be see what happens in a debug build. The crash stacks don't look very useful. Unfortunately, a lot of the Fission team is going on PTO soon, so maybe you could just back this out, or at least the iframe part (which is less likely to bit rot), and somebody can look at it when we get back? Thanks.

Does it make sense to back out for this? It's a tier 2 failure and it's highlighting a problem with fission, which was in a sense the point of the change to the test anyway. I don't think running the test is likely to be relevant for measuring fission performance if we don't keep this change in, either.

Comment 7

•

4 years ago

I think it is against the spirit of tier 2 to leave a failing test in for 3 or more weeks, but I could be wrong.

Comment 8

•

4 years ago

Attached file Bug 1604972 - Remove tresize iframes temporarily r?mccr8 — Details

Phabricator Automation

Updated

•

4 years ago

Assignee: nobody → dothayer

Status: NEW → ASSIGNED

Bogdan Tara[:bogdan_tara | bogdant]

Updated

•

4 years ago

Keywords: leave-open

Pulsebot

Comment 9

•

4 years ago

Pushed by dothayer@mozilla.com:
https://hg.mozilla.org/integration/autoland/rev/bb2ea10279b3
Remove tresize iframes temporarily r=mccr8

Comment 10

•

4 years ago

bugherder

https://hg.mozilla.org/mozilla-central/rev/bb2ea10279b3

Comment hidden (Intermittent Failures Robot)

BugBot [:suhaib / :marco/ :calixte]

Comment 12

•

4 years ago

The priority flag is not set for this bug.
:rwood, could you have a look please?

For more information, please visit auto_nag documentation.

Flags: needinfo?(rwood)

Robert Wood [:rwood]

Updated

•

4 years ago

Flags: needinfo?(rwood)

Priority: -- → P3

Florin Strugariu [:Bebe]

Comment 13

•

4 years ago

https://wiki.mozilla.org/Bug_Triage#Intermittent_Test_Failure_Cleanup

Status: ASSIGNED → RESOLVED

Closed: 4 years ago

Resolution: --- → INCOMPLETE

BugBot [:suhaib / :marco/ :calixte]

Updated

•

4 years ago

Keywords: leave-open

u608768

Comment 14

•

4 years ago

Reopening since the failing test wasn't actually fixed yet, just disabled.

Status: RESOLVED → REOPENED

Keywords: leave-open

Resolution: INCOMPLETE → ---

Chris Peterson [:cpeterson]

Comment 15

•

4 years ago

Also removing myself as the assignee as I believe a fission person will need to take this.

Assignee: dothayer → nobody

Comment 16

•

4 years ago

Tracking for Fission M6b Nightly because we should very whether this test is still crashing with Fission. A lot of Fission code has changed since tresize was disabled six months ago.

https://searchfox.org/mozilla-central/rev/8118a1f3390b232e70497770d372de062e44dd5f/testing/talos/talos/tests/tresize/addon/chrome/tresize-test.html#20-22

Fission Milestone: M6 → M6b

Comment 17

•

4 years ago

Randell, can you try re-enabling this and checking if it still fails?

Flags: needinfo?(rjesup)

Updated

•

4 years ago

Assignee: nobody → rjesup

Flags: needinfo?(rjesup)

Assignee

Updated

•

4 years ago

Flags: needinfo?(rjesup)

Assignee

Comment 18

•

4 years ago

Per https://treeherder.mozilla.org/#/jobs?repo=try&revision=2cdea9730b0a1227dd493fc4753afc5a20fd99bd it appears that we can revert this disabling. I'll check it isn;'t intermittent before committing

Flags: needinfo?(rjesup)

Comment 19

•

4 years ago

Randell confirmed that with his latest try run, it seems to no longer be a problem so we can re-enable it.

Flags: needinfo?(rjesup)

Updated

•

4 years ago

Whiteboard: will reenable the test

Comment 20

•

4 years ago

Enabling the test doesn't have to block M6b, so moving to M6c so we do this soon.

Fission Milestone: M6b → M6c

Assignee

Comment 21

•

4 years ago

Attached file Bug 1604972: re-enable iframes in tresize-test.html r=smaug — Details

Cristian Brindusan [:cbrindusan]

Assignee

Comment 22

•

4 years ago

fission/talos try: https://treeherder.mozilla.org/#/jobs?repo=try&author=rjesup%40wgate.com&selectedTaskRun=LqK6xvcDT1C5s_kph7tsug.0 doesn't seem to have this problem anymore. There are 'outlook' crashes; looking at that (that wouldn't be this bug).

Flags: needinfo?(rjesup)

Pulsebot

Comment 23

•

4 years ago

Pushed by rjesup@wgate.com:
https://hg.mozilla.org/integration/autoland/rev/2269b7fe6cee
re-enable iframes in tresize-test.html r=perftest-reviewers,sparky

Comment 24

•

4 years ago

bugherder

https://hg.mozilla.org/mozilla-central/rev/2269b7fe6cee