AU tests failing with: "Omaha Response does not have version in manifest!" |
||||
Issue descriptionA change in 67 has broken some of the AU tests that rely on nano_omaha_devserver. You can see the tests started failing in R67-10503.0.0: https://stainless.corp.google.com/search?view=matrix&row=model&col=build&first_date=2018-03-10&last_date=2018-04-06&test=%5Eautoupdate%5C_CatchBadSignatures%24&exclude_cts=true&exclude_not_run=false&exclude_non_release=true&exclude_au=true&exclude_acts=true&exclude_retried=true&exclude_non_production=false Here is the changelist: https://crosland.corp.google.com/log/10502.0.0..10503.0.0 Last working update_engine log: https://pantheon.corp.google.com/storage/browser/chromeos-autotest-results/184835544-chromeos-test/chromeos6-row1-rack7-host5/autoupdate_CatchBadSignatures/autoupdate_CannedOmahaUpdate/results/ First broken update engine log: https://pantheon.corp.google.com/storage/browser/chromeos-autotest-results/184985490-chromeos-test/chromeos6-row1-rack7-host1/autoupdate_CatchBadSignatures/autoupdate_CannedOmahaUpdate/results/
,
Apr 6 2018
I have only caught this as my test device was on 66. The most recent test checked in about switching urls is also failing with the same error: https://stainless.corp.google.com/search?test=%5Eautoupdate%5C_OmahaResponse%5C.url%5C_switch%5C.full%24&exclude_non_release=true&exclude_cts=true&col=build&row=model&view=matrix&first_date=2018-03-31&last_date=2018-04-06
,
Apr 6 2018
Amin, can you help me figure out what broke here?
,
Apr 6 2018
autoupdate_CannedOmahaUpdate uses nano_omaha_devserver, which hardcodes the wrong appid: http://cs/chromeos_public/src/third_party/autotest/files/client/cros/update_engine/nano_omaha_devserver.py?l=37&rcl=f63161c201b6c5eb818184697f1153d680f16748 so this may be similar to https://bugs.chromium.org/p/chromium/issues/detail?id=829691, b/76117708
,
Apr 6 2018
If you are using an older version of devserver, these two changes need to be cherry picked: https://chromium-review.googlesource.com/c/chromiumos/platform/dev-util/+/961585 https://chromium-review.googlesource.com/c/chromiumos/platform/dev-util/+/961586 The nano_omaha_devserver should have these changes. Is nano_omaha_devserver has a different code base than devserver?
,
Apr 6 2018
Yeah, I didn't know that nano_omaha_devserver hardcode the app id. It needs to be changed.
,
Apr 6 2018
Thanks for the pointers. When I changed the file to return the appid and hash_sha256 it fixed the problem. nano_omaha_devserver has no code shared with the devservers. It is a dumb class that is used to return a simple response. It had no logic for figuring stuff out.
,
Apr 6 2018
,
Apr 6 2018
I'm sorry, I wasn't aware of the nano_omaha_devserver code when I added that change. I think you are the author of nano_omaha_devserver and you probably know better how to fix and test it. Sorry for the trouble.
,
Apr 7 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/autotest/+/a8512d0b30036159f71fc25260b9f0fb3050c382 commit a8512d0b30036159f71fc25260b9f0fb3050c382 Author: David Haddock <dhaddock@chromium.org> Date: Sat Apr 07 05:00:28 2018 Return correct app id and hash_sha256 in nano_omaha_devserver response. Previously we just returned a default app id for every device. In 67 update_engine has gotten stricter about what it will allow. So we need to parse the appid from the request and return it in the response. TEST=Hacking locally with autoupdate_CatchBadSignatures. BUG= chromium:830025 Change-Id: Idfec1edaa842f74db44105b7d24f86b01aa32da0 Reviewed-on: https://chromium-review.googlesource.com/1000537 Commit-Ready: David Haddock <dhaddock@chromium.org> Tested-by: David Haddock <dhaddock@chromium.org> Reviewed-by: Amin Hassani <ahassani@chromium.org> [modify] https://crrev.com/a8512d0b30036159f71fc25260b9f0fb3050c382/client/cros/update_engine/nano_omaha_devserver.py
,
Apr 9 2018
,
Dec 7
|
||||
►
Sign in to add a comment |
||||
Comment 1 by dhadd...@chromium.org
, Apr 6 2018