666798 - sporadic reftest-2 failure on Android with "Not granting permission android.permission.[foo] to package com.mozilla.browser" (also "com.mozilla.SUTAgentAndroid")

Status: RESOLVED FIXED

(Release Engineering :: General, defect)

Description

[Daniel Holbert [:dholbert]]

http://tinderbox.mozilla.org/showlog.cgi?log=Mozilla-Inbound/1308869793.1308869948.4176.gz
Android Tegra 250 mozilla-inbound opt test reftest-2 on 2011/06/23 15:56:33
s: tegra-032

{
06-23 15:57:08.226 W/PackageManager( 1017): Not granting permission android.permission.SEND_DOWNLOAD_COMPLETED_INTENTS to package com.android.browser (protectionLevel=2 flags=0x1be45)
06-23 15:57:08.226 W/PackageManager( 1017): Not granting permission android.permission.REBOOT to package com.mozilla.SUTAgentAndroid (protectionLevel=3 flags=0xbe46)
06-23 15:57:08.226 W/PackageManager( 1017): Not granting permission android.permission.DEVICE_POWER to package com.mozilla.SUTAgentAndroid (protectionLevel=2 flags=0xbe46)
06-23 15:57:08.226 W/PackageManager( 1017): Not granting permission android.permission.INSTALL_PACKAGES to package com.mozilla.SUTAgentAndroid (protectionLevel=3 flags=0xbe46)
06-23 15:57:08.226 W/PackageManager( 1017): Not granting permission android.permission.STATUS_BAR to package com.mozilla.SUTAgentAndroid (protectionLevel=3 flags=0xbe46)
06-23 15:57:08.226 W/PackageManager( 1017): Not granting permission android.permission.SET_TIME to package com.mozilla.SUTAgentAndroid (protectionLevel=3 flags=0xbe46)
06-23 15:57:08.226 W/PackageManager( 1017): Not granting permission android.permission.WRITE_SECURE_SETTINGS to package com.mozilla.SUTAgentAndroid (protectionLevel=3 flags=0xbe46)
}

program finished with exit code -1
elapsedTime=83.979327

Comment 1

[Daniel Holbert [:dholbert]]

(CC'ing some mobile people.  Not sure if RelEng is the right home for this; feel free to move if appropriate)

Comment 2

[Daniel Holbert [:dholbert]]

(I don't actually know if the "not granting permission" is related at all to the "exit code -1", but it's the last mozilla-related thing in the logcat output before the exit.)

Comment 3

[Phil Ringnalda (:philor)]

I'd be more suspicious of the

Cleanup Device exception
=== Output ===
python /builds/sut_tools/cleanup.py 10.250.49.19
...
reconnecting socket
unable to connect socket
devroot None
/builds/tegra-032/test/../error.flg
Remote Device Error: devRoot from devicemanager [None] is not correct
program finished with exit code 1

Comment 4

[John Ford [:jhford] CET/CEST Berlin Time]

I think this could be related to the certificate issues we were experiencing earlier, or else machines running out of space, as at least one device is mentioned in 666821.

Comment 5

[John O'Duinn [:joduinn] (please use "needinfo?" flag)]

(In reply to comment #4)
> I think this could be related to the certificate issues we were experiencing
> earlier, 
This work was tracked in bug#666760, and was rolled out across all builders approx 3.30-4pm today. Test tegras should have been picking up builds generated with those new keys from then onwards, so we believe the key problem is now resolved. As you are seeing these "not granting permissions" error in the same timeframe, it *might* have been fixed by bug#666760. However, if you see this again, please let us know.

Keeping this bug open to track.



> or else machines running out of space, as at least one device is
> mentioned in 666821.

Comment 6

[Mike Taylor [:bear]]

(In reply to comment #5)
> (In reply to comment #4)
> > I think this could be related to the certificate issues we were experiencing
> > earlier, 
> This work was tracked in bug#666760, and was rolled out across all builders
> approx 3.30-4pm today. Test tegras should have been picking up builds
> generated with those new keys from then onwards, so we believe the key
> problem is now resolved. As you are seeing these "not granting permissions"
> error in the same timeframe, it *might* have been fixed by bug#666760.
> However, if you see this again, please let us know.
> 
> Keeping this bug open to track.

we didn't really roll out any new keys that deal with the fennec apk or the tegras - the change was to the sutagent.apk that is built which is not used at all by the android testing process.

I would like the ateam to look at this first as the sutagent and it's actions during test runs is controlled by them and they are also the better ones to tell us what the real issue is (I hope :)

> 
> > or else machines running out of space, as at least one device is
> > mentioned in 666821.

Comment 7

[Mike Taylor [:bear]]

this hasn't popped up since the fix was implemented - marking as resolved

(where fix is the removal of the sdk auto-created signing key)