1202867 - test_play_3gp_video.py is failing all the time in the Jenkins labs

Status: RESOLVED FIXED

(Firefox OS Graveyard :: Gaia::UI Tests, defect)

Description

[Martijn Wargers (dead)]

This test is failing all the time in the automation labs, examples here:
http://jenkins1.qa.scl3.mozilla.com/job/flame-kk-319.mozilla-central.nightly.ui.functional.smoke.bitbar/394/HTML_Report/
http://jenkins1.qa.scl3.mozilla.com/job/flame-kk-319.mozilla-central.nightly.ui.functional.smoke/423/HTML_Report/
It shows that the video is somehow not on the device.

Could it be that the file is somehow not in the repo there?

Comment 1

[No-Jun Park [:njpark]]

I'll see whether I can find any clues

Comment 2

[No-Jun Park [:njpark]]

When I ran this in jenkins by itself, the test was passing all 2 runs that it ran with different video files.
Logcat from successful single-script run:
http://jenkins1.qa.scl3.mozilla.com/view/Graphics/job/flame-kk.mozilla-central.nightly.ui.graphics.reference.adhoc.bitbar/70/artifact/tests/python/gaia-ui-tests/logcat.txt

But in the smoketest, it consistenly fails here. (A bunch of scripts are running prior to this one, but b2g is restarted for each test run)  I'm suspecting something that happens before this script run might be causing this issue.

Logcat from failed smoketest run: (18MB)
http://jenkins1.qa.scl3.mozilla.com/job/flame-kk-319.mozilla-central.nightly.ui.functional.smoke/ws/tests/python/gaia-ui-tests/logcat.txt

in above logcat, 09-09 05:45:23.862 is where the test_play_3gp_video.py starts.  There are some errors, but nothing really stands out to me.  The test fails because it doesn't detect any video files.  :djf, if you have time, perhaps could you take a quick look and see anything that might cause this? Thanks a lot in advance!

Comment 3

[No-Jun Park [:njpark]]

(In reply to No-Jun Park [:njpark] from comment #2)
> in above logcat, 09-09 05:45:23.862 is where the test_play_3gp_video.py
> starts.  There are some errors, but nothing really stands out to me.  The

This logcat might get replaced by more recent ones, in that case, you can search for the beginning of the test by looking for: "test_play_3gp_video.py:test_play_3gp_video".  If the line starts with TEST-START, then it's where the script starts.

script does the following before the crash:
Push VID_0001.3gp to the phone
Launch Video App
Waiting for one thumbnail to load

Comment 4

[No-Jun Park [:njpark]]

Also note that the exact same video file was loaded in previous test script, but not on this script run. (if you search for VID_0001.3gp you'll see)

Comment 5

[No-Jun Park [:njpark]]

clearing ni? for :djf since this looks more and more like jenkins issue

Comment 6

[Martijn Wargers (dead)]

It turns out this is happening, because test_settings_usb_storage.py was running before. This causes all files on the internal sdcard to not be visible anymore, until the device is restarted.
We can't do a full restart in Gaia UI tests.
We decided to set test_settings_usb_storage.py as the last test to run in a test run.
To do that, we need to reorder some entries in the manifest.ini files.

Comment 7

[GitHub Autolander Bot]

Attachment: [gaia] mwargers:1202867 > mozilla-b2g:master

Comment 8

[Martijn Wargers (dead)]

Comment on attachment 8660275 [details] [review]
[gaia] mwargers:1202867 > mozilla-b2g:master

I also moved test_wifi_multiple_connect.py to the beginning, because it's a stability test (to test for the Marionette connection lost issue that was pestering us for a while).

Comment 9

[Martijn Wargers (dead)]

Looks like test_settings_usb_storage.py needs to be updated, btw, when bug bug 1194045 lands.

Comment 10

[No-Jun Park [:njpark]]

Comment on attachment 8660275 [details] [review]
[gaia] mwargers:1202867 > mozilla-b2g:master

LGTM

Comment 11

[Martijn Wargers (dead)]

Merged: https://github.com/mozilla-b2g/gaia/commit/8d95a8f2f1b83af84e2ba56416b9568b01b1b068