PFQ failure with error "Failed to set up adb connection" |
|||||||||
Issue descriptionveyron_minnie-chrome-pfq:1741 failed Builders failed on: - veyron_minnie-chrome-pfq: https://luci-milo.appspot.com/buildbot/chromeos/veyron_minnie-chrome-pfq/1741 Error: Failed to set up adb connection It looks like Android is started and adb connect succeeded. But device shows up as "offline". 10/04 00:41:44.676 INFO | autotest:0601| END GOOD cheets_StartAndroid cheets_StartAndroid timestamp=1507102896 localtime=Oct 04 00:41:36 10/04 00:41:44.676 INFO | tradefed_test:0435| Setting up adb connection. 10/04 00:41:44.677 DEBUG| utils:0212| Running 'adb keygen '/tmp/tmpL12yL0_cheets_GTS.GtsAdminTestCases/test_key'' 10/04 00:41:45.290 DEBUG| ssh_host:0296| Running (ssh) 'android-sh -c 'echo '"'"'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 root@test-chromeos4-row9-rack9-host5'"'"' > /data/misc/adb/adb_keys'' from '_ready_arc|_connect_adb|_write_android_file|_android_shell|run|run_very_slowly' 10/04 00:41:46.102 DEBUG| ssh_host:0296| Running (ssh) 'android-sh -c 'restorecon /data/misc/adb/adb_keys'' from 'run_once|_ready_arc|_connect_adb|_android_shell|run|run_very_slowly' 10/04 00:41:46.879 ERROR| utils:0280| [stderr] SELinux: Loaded file_contexts contexts from /file_contexts.bin. 10/04 00:41:47.067 DEBUG| utils:0212| Running 'adb kill-server' 10/04 00:41:47.171 DEBUG| ssh_host:0296| Running (ssh) 'android-sh -c 'setprop sys.usb.config mtp,adb'' from 'run_once|_ready_arc|_connect_adb|_android_shell|run|run_very_slowly' 10/04 00:41:48.061 DEBUG| ssh_host:0296| Running (ssh) 'android-sh -c 'setprop persist.sys.usb.config mtp,adb'' from 'run_once|_ready_arc|_connect_adb|_android_shell|run|run_very_slowly' 10/04 00:41:48.913 DEBUG| utils:0212| Running 'adb connect 'chromeos4-row9-rack9-host5:22'' 10/04 00:41:52.348 INFO | tradefed_test:0398| adb connect chromeos4-row9-rack9-host5:22: * daemon not running. starting it now on port 5037 * * daemon started successfully * connected to chromeos4-row9-rack9-host5:22 10/04 00:41:52.363 DEBUG| utils:0212| Running 'adb devices' 10/04 00:41:52.428 INFO | tradefed_test:0403| adb devices: List of devices attached chromeos4-row9-rack9-host5:22 offline ... Similar issues happened veyron_minnie-tot-chrome-pfq-informational yesterday, e.g. Failed to set up adb connection: https://uberchromegw.corp.google.com/i/chromeos.chrome/builders/veyron_minnie-tot-chrome-pfq-informational/builds/6219 Android did not boot: https://uberchromegw.corp.google.com/i/chromeos.chrome/builders/veyron_minnie-tot-chrome-pfq-informational/builds/6218 It is a bit flaky on the informational, Build 6229 is green. But 6230 is failing again.
,
Oct 5 2017
Looks like betty-chrome-pfq and caroline-chrome-pfq failed for the same error in VMTest cheets_:
Error: Failed to set up adb connection
Another error found was:
Will raise error TestFail('adb is not ready in 60 seconds.',) due to unexpected return: False
https://uberchromegw.corp.google.com/i/chromeos/builders/betty-chrome-pfq/builds/256
https://uberchromegw.corp.google.com/i/chromeos/builders/caroline-chrome-pfq/builds/562
,
Oct 5 2017
,
Oct 5 2017
,
Oct 5 2017
,
Oct 5 2017
So it is consistently failing to connect to adb on ToT Chrome ? PFQ seems to be consistently failing.
,
Oct 5 2017
Luis and I took a quick look at logs, and suspect something related to vold is holding up the boot process some of the time. Of course this would result in the test timing out in this manner. Can someone more familiar with vold take a look? (cc'ing folks from Tokyo, as well as next week's constables)
,
Oct 5 2017
FWIW, there are at least two symptoms being talked about - the logs in comment #7 match the error in comment #2. Not sure if the same root cause as the original report or not. But it is happening on chrome PFQ.
,
Oct 6 2017
As an aside, some tests use arc.py which uses a heuristic of "have we seen the event 'ArcAppLauncer:started' in the Android event log", while others use tradefed_test.py which uses the heuristic "is org.chromium.arc.intent_helper process running" for the "has android booted yet". The former (ArcAppLauncher:started) is wired up to a BOOT_COMPLETED intent receiver, while the latter appears to happen earlier (before BOOT_COMPLETED). So it is possible that different tests could fail in slightly different ways ("adb is not ready in 60 seconds" vs "Error: Failed to set up adb connection") but have the same symptom (android didn't fully boot), and possibly even the same root cause.
,
Oct 6 2017
According to veyron_minnie-tot-chrome-pfq-informational, it was broken between https://chromium.googlesource.com/chromium/src/+log/073de6fde992cb746985430230e8a02558d1862f..73dd2ff8e610aff52ab6edd1bf8e312e39b98aca and it was fixed between https://chromium.googlesource.com/chromium/src/+log/da69d8843dc3e59e17a998231cfea40bd7d1f61c..bc438f0b2dbf9cf2ccb5f622256b3fd3ea2fe2f0 My wild guess is the following changes: Reland support for Mojom message ID scrambling https://chromium-review.googlesource.com/693037 Mojo: Fix scrambled message ID build conditions https://chromium-review.googlesource.com/699854 Description of the latter change says: > This corrects the condition used to enable scrambled message > IDs, ensuring that they are only enabled on Windows, Mac, and > Linux desktop builds, and NOT on Chrome OS.
,
Oct 6 2017
Yes, that would probably make sense.
,
Oct 6 2017
Chrome PFQ runs for veyron_minnie has cycled green on the latest PFQ runs. Given comments #10 and #11 and this fact, I will mark this fixed.
,
Jan 22 2018
,
Jan 23 2018
|
|||||||||
►
Sign in to add a comment |
|||||||||
Comment 1 by xiy...@chromium.org
, Oct 4 2017