1035040 - [MTBF][Marionette] Connection to Marionette server is lost. Check gecko.log (desktop firefox) or logcat (b2g) for errors.

Status: RESOLVED WORKSFORME

(Remote Protocol :: Marionette, defect)

Description

[Walter Chen[:ypwalter][:wachen]]

Hi,  I will open a bug for mdas to investigate.

http://mtbf-1:8080/job/flame.v200.mtbf/label=mtbf-2/7/consoleFull
Suspected Reason: Marionette server/client issue (ADB is still alive for us to get minidump)

23:29:30 07-05 23:29:26.747 E/ACDB-LOADER( 7943): Error: ACDB ANC returned = -1
23:29:30 07-05 23:29:26.747 E/ACDB-LOADER( 7943): Error: ACDB ANC returned = -1
23:29:30 07-05 23:29:26.747 E/ACDB-LOADER( 7943): Error: ACDB ANC returned = -1
23:29:30 07-05 23:29:26.747 E/ACDB-LOADER( 7943): Error: ACDB ANC returned = -1
23:29:30 07-05 23:29:27.327 E/ACDB-LOADER( 7943): Error: ACDB AudProc vol returned = -19

IOError: Connection to Marionette server is lost. Check gecko.log (desktop firefox) or logcat (b2g) for errors.

Comment 1

[Walter Chen[:ypwalter][:wachen]]

Malini,

Can you help with this bug? What are information that can help you with investigation?

Comment 2

[Malini Das [:mdas] - Away, not checking bugmail]

(In reply to Walter Chen[:ypwalter][:wachen] from comment #1)
> Malini,
> 
> Can you help with this bug? What are information that can help you with
> investigation?

Do you have the full logcat? Also, I'm not aware of what ACDB-LOADER is or who's responsible for it. If you can find who they are, they may be able to diagnose this issue.

Comment 3

[Walter Chen[:ypwalter][:wachen]]

Unfortunately, I didn't turned on the adb logcat. I will try to turn on logcat for you and see if I can reproduce it...

Comment 4

[Walter Chen[:ypwalter][:wachen]]

http://mtbf-1:8080/job/flame.v200.mtbf/label=mtbf-2/33/consoleFull
Reproduced, but no logcat, trying to reproduce again

Comment 5

[Paul Yang [:pyang] (away)]

Attachment: logcat_5

logcat attached

Comment 6

[Malini Das [:mdas] - Away, not checking bugmail]

I'll show you how to diagnose problems like this: The way to find out the probable cause of a crash is to look for errors and exceptions in the log.

By doing `cat logcat_5 | grep -i "error\|exception"` you'll see some 'normal' errors like 'I/Gecko   (  296): ###!!! [Parent][MessageChannel] Error: Channel error: cannot send/recv' which is harmless, you may see salient messages like:

E/ACDB-LOADER( 3451): Error: ACDB get asm topologies returned = -18
E/ACDB-LOADER( 3451): Error: ACDB ANC returned = -1
E/ACDB-LOADER( 3451): Error: ACDB ANC returned = -1
E/ACDB-LOADER( 3451): Error: ACDB ANC returned = -1

Which is unusual. You can then look at the log and see where this happens to get more context:


D/ACDB-LOADER( 3451): ACDB -> ACDB_CMD_GET_ANC_SETTING
D/        ( 3451): Invalid AV file. It doesnt contain LUT for tblid 0000000A
D/        ( 3451): Failed to fetch the lookup information of the device 0000001C
E/ACDB-LOADER( 3451): Error: ACDB ANC returned = -1
D/ACDB-LOADER( 3451): ACDB -> ACDB_CMD_GET_ANC_SETTING
D/        ( 3451): Invalid AV file. It doesnt contain LUT for tblid 0000000A
D/        ( 3451): Failed to fetch the lookup information of the device 0000001D
E/ACDB-LOADER( 3451): Error: ACDB ANC returned = -1
D/ACDB-LOADER( 3451): ACDB -> ACDB_CMD_GET_ANC_SETTING
D/        ( 3451): Invalid AV file. It doesnt contain LUT for tblid 0000000A
D/        ( 3451): Failed to fetch the lookup information of the device 0000001E
E/ACDB-LOADER( 3451): Error: ACDB ANC returned = -1
D/ACDB-LOADER( 3451): ACDB -> ACDB_CMD_GET_ANC_SETTING
D/        ( 3451): Invalid AV file. It doesnt contain LUT for tblid 0000000A
D/        ( 3451): Failed to fetch the lookup information of the device 0000001F
E/ACDB-LOADER( 3451): Error: ACDB ANC returned = -1
D/ACDB-LOADER( 3451): ACDB -> ACDB_CMD_GET_ANC_SETTING
D/        ( 3451): Invalid AV file. It doesnt contain LUT for tblid 0000000A
D/        ( 3451): Failed to fetch the lookup information of the device 00000067
E/ACDB-LOADER( 3451): Error: ACDB ANC returned = -1
D/ACDB-LOADER( 3451): ACDB -> init done!


Looks like ACDB isn't being configured correctly, and a reboot followed. There's another error after that reboot:
E/GeckoConsole( 3640): Content JS ERROR at app://verticalhome.gaiamobile.org/gaia_build_defer_index.js:384 in loadSVConfFileError: Failed parsing singl     eVariant configuration file [js/singlevariantconf.json]:  [Exception... "File error: Not found"  nsresult: "0x80520012 (NS_ERROR_FILE_NOT_FOUND)"  loca     tion: "JS frame :: app://verticalhome.gaiamobile.org/gaia_build_defer_index.js :: loadFile :: line 374"  data: no]

and it *seems* the fallout from that triggered the crash:
E/GeckoConsole( 3640): Content JS ERROR at app://verticalhome.gaiamobile.org/gaia_build_defer_index.js:468 in GridItem.prototype.renderIcon/<: Error fe     tching icon Error: Error while HTTP GET:
...
...
I/GeckoDump( 3512): Crash reporter : Can't fetch app.reportCrashes. Exception: [Exception... "Component returned failure code: 0x8000ffff (NS_ERROR_UNE XPECTED) [nsIPrefBranch.getBoolPref]"  nsresult: "0x8000ffff (NS_ERROR_UNEXPECTED)"  location: "JS frame :: chrome://b2g/content/shell.js :: shell_repo     rtCrash :: line 149"  data: no]
W/AudioPolicyManagerALSA( 3500): setPhoneState() setting same state 0
1908 I/GeckoDump( 3512): Crash reporter : Can't fetch app.reportCrashes. Exception: [Exception... "Component returned failure code: 0x8000ffff (NS_ERROR_UNEXPECTED) [nsIPrefBranch.getBoolPref]"  nsresult: "0x8000ffff (NS_ERROR_UNEXPECTED)"  location: "JS frame :: chrome://b2g/content/shell.js :: shell_repo     rtCrash :: line 149"  data: no]


So it looks like it's either this ACDB bit got into a weird state, and after the reboot, gaia was unable to get a file resource and it crashed. Contacting whoever deals with ACDB and having them look at the ACDB failure, and asking the gaia team about this particular file error should be the way forward.

Comment 7

[Paul Yang [:pyang] (away)]

Thanks Malini, after investigating by gecko engineer, it looks like an ipc error in graphic module.

Comment 8

[Walter Chen[:ypwalter][:wachen]]

Attachment: logcat_47

http://mtbf-1:8080/job/flame.v200.mtbf/47/label=mtbf-2/

It appearred as the same bug.

Comment 9

[David Burns :automatedtester]

Is this still an issue? Seeing comment 7 it suggests that it's not a marionette bug

Comment 10

[Walter Chen[:ypwalter][:wachen]]

nope, we can close as wfm then.