Samsung devices on chromium.perf all stuck at login screen |
|||||
Issue descriptionLink to buildbot status page: https://build.chromium.org/p/chromium.perf/builders/Android%20Galaxy%20S5%20Perf%20%281%29 https://build.chromium.org/p/chromium.perf/builders/Android%20Galaxy%20S5%20Perf%20%282%29 https://build.chromium.org/p/chromium.perf/builders/Android%20Galaxy%20S5%20Perf%20%283%29 All the non-reference tests are failing. When I click through to the logs, I see screenshot links that all look like this: https://console.developers.google.com/m/cloudstorage/b/chrome-telemetry-output/o/profiler-file-id_0-2016-10-24_03-18-4072413.png Not sure if infra labs or clank infra should take a look?
,
Oct 24 2016
What you are seeing is the default language panel after the devices have been provisioned. I touched a couple and they were responsive. I checked uptime on the phones: Note the pattern within each host. Let me know if you want me to do something hands on. In the meantime I'm going to set it back to untriaged and added a couple of clank infra folks. build21-b1: Checking 3208851faca351f3... up time: 04:54:34, idle time: 1 days, 14:38:38, sleep time: 00:00:00 Checking 3208e0a226fa51b7... up time: 04:06:22, idle time: 1 days, 08:24:07, sleep time: 00:00:00 Checking 32085a73842c615b... up time: 03:43:38, idle time: 1 days, 05:16:14, sleep time: 00:00:00 Checking 3208df23b0c251e1... up time: 02:53:50, idle time: 22:45:42, sleep time: 00:00:00 Checking 3208e0600bb251f3... up time: 02:48:28, idle time: 22:04:32, sleep time: 00:00:00 Checking 3208e623c90051f7... up time: 01:55:30, idle time: 15:05:03, sleep time: 00:00:00 Checking 3208584f952c61ef... up time: 01:16:23, idle time: 08:56:17, sleep time: 00:00:00 On build22-b1: Checking 32081d5f765c510d... up time: 00:21:04, idle time: 02:34:25, sleep time: 00:00:00 Checking 3208dd33a9c25169... up time: 00:21:03, idle time: 02:35:34, sleep time: 00:00:00 Checking 3208531995be5145... up time: 00:16:44, idle time: 01:50:50, sleep time: 00:00:00 Checking 3208521990be511d... up time: 00:16:37, idle time: 01:55:01, sleep time: 00:00:00 Checking 32085d1787be514b... up time: 00:08:04, idle time: 00:50:55, sleep time: 00:00:00 Checking 3208611192be51c3... up time: 00:06:08, idle time: 00:35:30, sleep time: 00:00:00 Checking 320851777626611b... up time: 00:04:17, idle time: 00:28:41, sleep time: 00:00:00 On build23-b1: Checking 3208086b755c5135... up time: 01:07:36, idle time: 08:44:45, sleep time: 00:00:00 Checking 3208d22daac2518f... up time: 00:50:27, idle time: 06:30:20, sleep time: 00:00:00 Checking 32089d0db6a351b7... up time: 00:45:11, idle time: 05:44:46, sleep time: 00:00:00 Checking 32082171745c5147... up time: 00:36:29, idle time: 04:42:07, sleep time: 00:00:00 Checking 3208e46a04b25131... up time: 00:24:35, idle time: 03:08:26, sleep time: 00:00:00 Checking 3208cd5005b25183... up time: 00:23:15, idle time: 02:45:31, sleep time: 00:00:00 Checking 32089f2db2a351c5... up time: 00:00:56, idle time: 00:04:13, sleep time: 00:00:00
,
Oct 24 2016
Is part of provisioning making sure they're set up passed OOBE?
,
Oct 24 2016
#3: Yes, though it's possible we're not doing it correctly (or not doing so all the time) on S* devices.
,
Oct 24 2016
Just wondering, as this is in the CIT camp if so, right? Vince - can you find someone to help drive the rest of this? Also, who's on the hook for figuring out if we missed something or if we're doing it correctly?
,
Oct 25 2016
jbudorick@: is this part of provision_devices, or some other step?
,
Oct 25 2016
FRE/OOBE disabling? it's part of provision_devices.
,
Oct 26 2016
Issue 659786 has been merged into this issue.
,
Oct 26 2016
Issue 659784 has been merged into this issue.
,
Oct 26 2016
sheriff ping, is anyone actively looking at this? It's causing a pretty large amount of test failures.
,
Oct 26 2016
,
Oct 26 2016
I don't think this is a chrome-labs issue?
,
Oct 26 2016
Nope. Definitely infra-android's. Someone here needs to investigate the device provisioning and ensure it's correctly getting past the devices past the OOBE walkthrough. I can do that, but I'm trooper for the rest of the week, so progress will be slow.
,
Oct 26 2016
You're trooper today and tomorrow. That's hardly the rest of the week.
,
Oct 26 2016
And also Friday. Correct me if I'm wrong, but that's literally the definition of "rest of the week" But more importantly, perhaps the fact that some of these devices have their system language set to Kazakh could be contributing to the failures: chrome-bot@build21-b1:(Linux 14.04):~$ /b/rr/tmpS_bw1c/w/src/third_party/catapult/devil/bin/deps/linux2/x86_64/bin/adb -s 3208584f952c61ef shell getprop | gr ep language [persist.sys.language]: [kk] [ro.product.locale.language]: [en]
,
Oct 26 2016
I think this is the default language presented with a default provision. And it's been like this since day one. So maybe this is a non-factor?
,
Oct 26 2016
Could be. But it looks like some of these devices haven't made it past the language selection screen since as far back as our logs go. Ctrl-f for "language" on https://luci-logdog.appspot.com/v/?s=chrome%2Fbb%2Fchromium.perf%2FAndroid_Galaxy_S5_Perf__1_%2F3300%2F%2B%2Frecipes%2Fsteps%2Fprovision_devices%2F0%2Fstdout# You'll see that after rebooting for the last time, the devices' setupwizard are in the LANGUAGE mode. It should be in the FINISH state, I believe. I'm waiting for https://build.chromium.org/p/chromium.perf/builders/Android%20Galaxy%20S5%20Perf%20%281%29/builds/4315 finish so I can hop on the bot and poke at the devices for a minute. Sure is taking a while though.
,
Oct 28 2016
Well it seems whatever firmware we've got on these things is a build for the KZ locale, so I guess that would explain the whole language thing: [ril.product_code]: [SM-G900HZWASKZ] I'm still trying to get it past the OOBE screen, but nothing seems to stick. It's definitely something with the eng build since we can get past it on user builds. I can keep poking at it, but I've been warned that these bots might be dropped from the perf waterfalls soon. sullivan@, can you confirm? How high of a priority are these devices for your team right now? Is this bug still relevant?
,
Nov 15 2016
Has this been worked on?
,
Nov 15 2016
It's on hold while we discuss whether these devices are worth the maintenance costs.
,
Nov 18 2016
|
|||||
►
Sign in to add a comment |
|||||
Comment 1 by vhang@chromium.org
, Oct 24 2016Status: Assigned (was: Untriaged)