Closed Bug 1289091 Opened 8 years ago Closed 8 years ago

Intermittent test_switch_frame.py TestSwitchFrame.test_after_switching_to_child_frame_navigates_changes_top | IOError: Process killed because the connection to Marionette server is lost. Check gecko.log for errors (Reason: Connection timed out after 360s

Categories

(Remote Protocol :: Marionette, defect)

Product:
Remote Protocol
Component:
Marionette
Version:
Version 3
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

(firefox50 fixed, firefox51 fixed)

Status:
RESOLVED FIXED
Milestone:
mozilla51
Tracking Flags:
Tracking Status
firefox50 --- fixed
firefox51 --- fixed

People

(Reporter: intermittent-bug-filer, Unassigned)

References

Details

(Whiteboard: [blocked by bug 1294456]) 

This failure message has actually be changed to the one from bug 1291127 which has better details now.
Status: NEW → RESOLVED
Closed: 8 years ago
Resolution: --- → DUPLICATE
Looks like I accidentally duped it against a wrong bug. The failures as seen do not match. The real failure message I will set as the summary now.
Status: RESOLVED → REOPENED
Resolution: DUPLICATE → ---
Summary: Intermittent test_switch_frame.py TestSwitchFrame.test_after_switching_to_child_frame_navigates_changes_top | IOError: process died with returncode None → Intermittent test_switch_frame.py TestSwitchFrame.test_after_switching_to_child_frame_navigates_changes_top | IOError: Process killed because the connection to Marionette server is lost. Check gecko.log for errors (Reason: Connection timed out after 360s
From the postings of Orange factor it looks like an asan, debug only failure.
Whiteboard: [asan, debug only]
Here an excerpt from the gecko.log:

1471362096192	Marionette	TRACE	conn801 -> [0,7,"switchToFrame",{"focus":true,"element":"48ac0344-497e-cc4c-9d44-3bf8cf4fa260"}]
1471362096300	Marionette	TRACE	conn801 <- [1,7,null,{}]
1471362096310	Marionette	TRACE	conn801 -> [0,8,"getTitle",null]
1471362096336	Marionette	TRACE	conn801 <- [1,8,null,{"value":"Unique title"}]
1471362096348	Marionette	TRACE	conn801 -> [0,9,"get",{"url":"http://127.0.0.1:49290/test.html"}]
++DOMWINDOW == 269 (0x1297a6c00) [pid = 1642] [serial = 835] [outer = 0x134960800]
JavaScript error: chrome://marionette/content/listener.js, line 1026: NS_ERROR_FAILURE: Component returned failure code: 0x80004005 (NS_ERROR_FAILURE) [nsIWebProgress.removeProgressListener]
[..]
1471362456368	Marionette	DEBUG	Closed connection conn801

It looks like that the call to 'marionette.get()' never returns the result after we switched to another frame. Looks to be the similar underlying issue as bug 1294456.
Status: REOPENED → NEW
Depends on: 1294456
Whiteboard: [asan, debug only]
Whiteboard: [blocked by bug 1294456]
This should no longer be an issue since the patch on bug 1294456 has been landed.
Status: NEW → RESOLVED
Closed: 8 years ago → 8 years ago
status-firefox51: --- → fixed
Resolution: --- → FIXED
Target Milestone: --- → mozilla51
This just hit on Aurora tip, which seems worrying.
https://treeherder.mozilla.org/logviewer.html#?job_id=3448582&repo=mozilla-aurora
Flags: needinfo?(hskupin)
Ryan, there can be various issues why we hang in this test. It's not clear to see if this is the same or another problem. Feel free to file a new bug or reopen this one. Personally I would prefer a new one due to the unknown nature. Once we have a failure for a debug build I can say more.
Flags: needinfo?(hskupin) → needinfo?(ryanvm)
Spun off to bug 1300076. Removing the intermittent-failure keyword from this bug so that it doesn't continue getting suggested.
Flags: needinfo?(ryanvm)
Keywords: intermittent-failure
(And in general, please remove intermittent-failure from bugs you're duping so that they don't keep getting suggested on Treeherder)
Product: Testing → Remote Protocol
You need to log in before you can comment on or make changes to this bug.