Server side packaging is broken in the lab |
|||||
Issue descriptionHi Dan, is there any recent change on server side package ? As shown in https://wmatrix.googleplex.com/matrix/unfiltered?suites=chameleon_audio_perbuild&days_back=20&hide_missing=True some of 8607.0 and all of 8608.0 audio tests failed from "sox command not installed" Log: 07/19 16:45:07.815 WARNI| lxc:0363| This function is obsoleted, please use install_packages instead. 07/19 16:45:07.847 INFO | audio_test:0080| Can not install sox outside of container. 07/19 16:45:07.847 DEBUG| base_utils:0185| Running 'sox --help' 07/19 16:45:07.874 WARNI| test:0606| Autotest caught exception when running test: Traceback (most recent call last): File "/usr/local/autotest/client/common_lib/test.py", line 587, in _exec *args, **dargs) File "/usr/local/autotest/client/common_lib/test.py", line 804, in _call_test_function return func(*args, **dargs) File "/usr/local/autotest/client/common_lib/test.py", line 719, in _cherry_pick_call return func(*p_args, **p_dargs) File "/usr/local/autotest/server/cros/audio/audio_test.py", line 31, in warmup audio_test_requirement() File "/usr/local/autotest/server/cros/audio/audio_test.py", line 72, in audio_test_requirement check_sox_installed() File "/usr/local/autotest/server/cros/audio/audio_test.py", line 96, in check_sox_installed raise error.TestError(error_message) TestError: sox command is not installed. 07/19 16:45:07.876 DEBUG| logging_manager:0627| Logging subprocess finished 07/19 16:45:07.888 DEBUG| logging_manager:0627| Logging subprocess finished It seems that lxc.install_package failed. def install_sox(): """Install sox command on autotest drone.""" try: lxc.install_package('sox') except error.ContainerError: logging.info('Can not install sox outside of container.') Could you please take a look what is wrong ? Thanks a lot!
,
Jul 20 2016
This CL cause all tests to skip using ssp: https://chromium-review.googlesource.com/#/c/360512/ The problem seems to be that create_host in factory.py gets an afe_host with value of EmptyAFEHost. The caller is stage_server_side_package: https://cs.corp.google.com/chromeos_public/src/third_party/autotest/files/server/control_segments/stage_server_side_package If there is an easy fix, please implement it asap. Or we have to chump this revert: https://chromium-review.googlesource.com/#/c/361843/1
,
Jul 20 2016
I filed bug 629746 so we will verify ssp is not broken before push to prod.
,
Jul 20 2016
Hi Dan, think you so much for figuring this out quickly! Hi Kevin, we really rely on you to stop the spam of perbuild test failure mails :) Thanks a lot!
,
Jul 20 2016
taking a look
,
Jul 20 2016
ok, I think I got a fix, although I'll need some help verifying the fix in the test lab. Will get a cl up in about 30 mins.
,
Jul 20 2016
,
Jul 20 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/autotest/+/adc99f928c3edca69af33d8f3b3309af10769be5 commit adc99f928c3edca69af33d8f3b3309af10769be5 Author: Kevin Cheng <kevcheng@chromium.org> Date: Wed Jul 20 15:21:58 2016 [autotest] Pass in afe_host when staging ssp. When running stage_server_side_package, only a list of machine names is passed through and not a list of machine dicts. This causes the afe_utils methods to fail when grabbing the attributes/labels since they by default depend on the afe_host in the host object. Update the staging to grab the afe_host object when running in the lab. BUG= chromium:629702 TEST=not yet Change-Id: I59c7732c78215537893567376cee982b68bc1647 Reviewed-on: https://chromium-review.googlesource.com/361539 Reviewed-by: Dan Shi <dshi@google.com> Commit-Queue: Dan Shi <dshi@google.com> Tested-by: Dan Shi <dshi@google.com> [modify] https://crrev.com/adc99f928c3edca69af33d8f3b3309af10769be5/server/autoserv
,
Jul 20 2016
Issue 629886 has been merged into this issue.
,
Jul 25 2016
Hi Kevin, thanks a lot for the fix! The test looks good now. https://wmatrix.googleplex.com/matrix/unfiltered?suites=chameleon_audio_perbuild&days_back=20&hide_missing=True |
|||||
►
Sign in to add a comment |
|||||
Comment 1 by cychiang@chromium.org
, Jul 20 2016