Closed Bug 855400 Opened 12 years ago Closed 11 years ago

Intermittent Windows mochitest-2,3 "command timed out: 7200 seconds elapsed, attempting to kill" after "[|WebrtcVideoSessionConduit] VideoConduit.cpp:633: mozilla::WebrtcVideoConduit::SendRTCPPacket Failed"

Categories

(Core :: WebRTC, defect)

Product:
Core
Component:
WebRTC
Platform:
x86
Windows 7
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

()

Status:
RESOLVED WORKSFORME

People

(Reporter: RyanVM, Assigned: abr)

Details

(Keywords: intermittent-failure, Whiteboard: [webrtc][blocking-webrtc-][webrtc-uplift]) 

Whole lotta WebRTC spew in this log...

https://tbpl.mozilla.org/php/getParsedLog.php?id=21165284&full=1&branch=mozilla-inbound#error3

Rev3 WINNT 6.1 mozilla-inbound opt test mochitest-2 on 2013-03-27 08:23:08 PDT for push 771c895e8b55
slave: talos-r3-w7-030

Note the extreme repetition of the following line until the process is finally killed.

08:47:35     INFO -  0[150f140]: [main|def] ccapi_call.c:56: SIPCC-SIP_CC_PROV: CCAPI_Call_releaseCallInfo: ref=0x1b58e480: count=1
By our rules for intermittent failure, this wouldn't block, but what's causing me to throw this into question is that the log here looks quite concerning. Maybe discuss first?
Whiteboard: [WebRTC][blocking-webrtc?]
actually this is the repetition:
10:23:20     INFO -  0[d20bb80]: [|WebrtcVideoSessionConduit] VideoConduit.cpp:624: mozilla::WebrtcVideoConduit::SendRTCPPacket : channel 0 , len 40
10:23:20     INFO -  0[d20bb80]: [|WebrtcVideoSessionConduit] VideoConduit.cpp:633: mozilla::WebrtcVideoConduit::SendRTCPPacket Failed

This appears to be another variant of a low-frequency intermittent "doesn't finish shutting down the call" bug (some sort of missed message?)

abr, can you link this to any related bugs?  It probably needs to stay open for starring reasons.
Assignee: nobody → adam
Whiteboard: [WebRTC][blocking-webrtc?] → [WebRTC][blocking-webrtc-]
https://tbpl.mozilla.org/php/getParsedLog.php?id=21881393&tree=Birch
The last one here isn't webrtc (not sure who); the others are a failure-to-shut-down instance.

Adam, is this another instance of a known bug?  (if so, we may need to leave it open for now for starring purposes, but we should link them.)
Summary: Intermittent Windows mochitest-2 shutdown "command timed out: 7200 seconds elapsed, attempting to kill" → Intermittent Windows mochitest-2,3 "command timed out: 7200 seconds elapsed, attempting to kill" after "[|WebrtcVideoSessionConduit] VideoConduit.cpp:633: mozilla::WebrtcVideoConduit::SendRTCPPacket Failed"
The frequency of this intermittent failure has really jumped up. Did something regress here?

Adam - Thoughts?
Flags: needinfo?(adam)
Whiteboard: [WebRTC][blocking-webrtc-] → [WebRTC][blocking-webrtc?]
(In reply to Jason Smith [:jsmith] from comment #18)
> The frequency of this intermittent failure has really jumped up. Did
> something regress here?
> 
> Adam - Thoughts?

I suspect that what happened here is that the sheriffs finally figured out that this failure they've been seeing corresponds to this bug number. Note that the pace *really* picked up after Ed changed things around to make it match TBPL runs for easier starring...
Flags: needinfo?(adam)
Whiteboard: [WebRTC][blocking-webrtc?] → [webrtc][blocking-webrtc-][webrtc-uplift]
...and then it mysteriously stops.
Sorry, all of the recent starrings were bad muscle memory, they were in fact bug 857127.
(In reply to TinderboxPushlog Robot from comment #45)
> gwagner
> https://tbpl.mozilla.org/php/getParsedLog.php?id=24508652&tree=Mozilla-B2g18
> Rev4 MacOSX Lion 10.7 mozilla-b2g18 opt test xpcshell on 2013-06-24 00:13:07

No, that's some kind of timeout in test_content_annotation.js on MacOS. The only thing it shares in common with this bug is that it caused a test timeout.

It's been on the order of six weeks since we've seen a repro on this issue. It think it's safe to say it's gone. I'm going to close it to prevent it from gathering errant stars, such as this one.
Status: NEW → RESOLVED
Closed: 11 years ago
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.