823532 - android 4.0 panda boards reboot in the middle of the test

Status: RESOLVED WORKSFORME

(Testing :: General, defect)

Description

[Joel Maher ( :jmaher ) (UTC -8)]

for the 8th time I am filing an identical bug, just a different time and place.

On two different occasions today I found jsreftests timing out without output in 2400 seconds, then in the proceeding logcat for the reboot step I see a clear path of a reboot taking place instead of gecko related output.

https://tbpl.mozilla.org/php/getParsedLog.php?id=18125653&tree=Firefox&full=1
https://tbpl.mozilla.org/php/getParsedLog.php?id=18125772&tree=Firefox

Comment 1

[Joel Maher ( :jmaher ) (UTC -8)]

To determine this if you look at the log and the testcase times out, you need to look at the full log and see what the next step has to say.  

The next step would be the reboot step and it will dump the logcat information from the device out, then it will attempt to reboot the device.

You should see in the logcat messages from Gecko* about pages, drawing, cleanup.  For a successful run, you will see:
12-19 15:10:26.210 I/WindowManager( 1401): WIN DEATH: Window{41525878 org.mozilla.fennec/org.mozilla.fennec.App paused=false}


If you see messages more like this:
01-01 00:00:15.648 E/EthernetService( 1401): EthernetManager.ETHERNET_STATE_ENABLED
01-01 00:00:15.664 E/EthernetStateMachine( 1401): startMonitoring resetInterface()

or:
01-01 00:00:19.218 W/ActivityManager( 1401): Permission Denial: receiving Intent { act=android.intent.action.BOOT_COMPLETED flg=0x10 } to com.android.musicfx requires android.permission.RECEIVE_BOOT_COMPLETED due to sender null (uid 1000)


then it shows that the device has rebooted.  Right now we care about finding patterns about which devices are rebooting on what jobs and how frequently.  There might be a simple pattern which emerges.  One factor we have seen in the past was heat and there is some work underway to monitor that and populate the logfile with information to help diagnose that as a root cause if it really is.

Comment 2

[Ed Morley [:emorley]]

Resolving WFM keyword:intermittent-failure bugs last modified >3 months ago, whose whiteboard contains none of:
{random,disabled,marked,fuzzy,todo,fails,failing,annotated,time-bomb,leave open}

There will inevitably be some false positives; for that (and the bugspam) I apologise. Filter on orangewfm.