Powerwash not triggered by TPM failure
Reported by
jrbarnette@chromium.org,
Mar 24 2016
|
|||
Issue description
Observe this repair failure:
https://pantheon.corp.google.com/storage/browser/chromeos-autotest-results/hosts/chromeos2-row4-rack6-host3/52705964-repair/
It fails because of a failure in the 'tpm' verifier.
(The failure is the wrong exception because of bug 597444 ,
but it's still a failure).
Failures in 'tpm' are supposed to trigger the 'powerwash' repair
action. But, the source code doesn't say that:
repair_actions = [
# ...
(AutoUpdateRepair, 'au', ['ssh', 'good_au'], ['cros']),
(PowerWashRepair, 'powerwash', ['ssh'], ['good_au', 'cros']),
(ServoInstallRepair, 'usb', [], ['ssh', 'good_au', 'cros']),
]
Somewhere along the way, a merge got lost.
,
Mar 26 2016
Devices are coming back online.
Sample successful repair:
START ---- repair timestamp=1458947802 localtime=Mar 25 16:16:42
GOOD ---- verify.ssh timestamp=1458947819 localtime=Mar 25 16:16:59
GOOD ---- verify.power timestamp=1458947819 localtime=Mar 25 16:16:59
GOOD ---- verify.cros timestamp=1458947821 localtime=Mar 25 16:17:01
GOOD ---- verify.good_au timestamp=1458947821 localtime=Mar 25 16:17:01
FAIL ---- verify.tpm timestamp=1458947822 localtime=Mar 25 16:17:02 Cannot load the TPM SRC public key
START ---- repair.powerwash timestamp=1458947822 localtime=Mar 25 16:17:02
START ---- reboot timestamp=1458947822 localtime=Mar 25 16:17:02
GOOD ---- reboot.start timestamp=1458947822 localtime=Mar 25 16:17:02
GOOD ---- reboot.verify timestamp=1458947857 localtime=Mar 25 16:17:37
END GOOD ---- reboot kernel=3.14.0 localtime=Mar 25 16:17:37 timestamp=1458947857
START ---- reboot timestamp=1458947867 localtime=Mar 25 16:17:47
GOOD ---- reboot.start timestamp=1458947867 localtime=Mar 25 16:17:47
GOOD ---- reboot.verify timestamp=1458947885 localtime=Mar 25 16:18:05
END GOOD ---- reboot kernel=3.14.0 localtime=Mar 25 16:18:05 timestamp=1458947885
START ---- reboot timestamp=1458948065 localtime=Mar 25 16:21:05
GOOD ---- reboot.start timestamp=1458948065 localtime=Mar 25 16:21:05
GOOD ---- reboot.verify timestamp=1458948099 localtime=Mar 25 16:21:39
END GOOD ---- reboot kernel=3.14.0 localtime=Mar 25 16:21:40 timestamp=1458948100
GOOD ---- verify.ssh timestamp=1458948131 localtime=Mar 25 16:22:11
GOOD ---- verify.tpm timestamp=1458948131 localtime=Mar 25 16:22:11
GOOD ---- verify.good_au timestamp=1458948131 localtime=Mar 25 16:22:11
GOOD ---- verify.cros timestamp=1458948134 localtime=Mar 25 16:22:14
END GOOD ---- repair.powerwash timestamp=1458948134 localtime=Mar 25 16:22:14
GOOD ---- verify.PASS timestamp=1458948134 localtime=Mar 25 16:22:14
START ---- reboot timestamp=1458948134 localtime=Mar 25 16:22:14
GOOD ---- reboot.start timestamp=1458948134 localtime=Mar 25 16:22:14
GOOD ---- reboot.verify timestamp=1458948153 localtime=Mar 25 16:22:33
END GOOD ---- reboot kernel=3.14.0 localtime=Mar 25 16:22:34 timestamp=1458948154
END GOOD ---- repair timestamp=1458948154 localtime=Mar 25 16:22:34 {'host_attributes': {'job_repo_url': 'http://100.107.160.1:8082/static/samus-release/R50-7978.18.0/autotest/packages', 'powerunit_hostname': 'chromeos2-row1-rack5-rpm1', 'hydra_hostname': 'chromeos-197-hydra1.cros', 'powerunit_outlet': '.A9'}, 'hostname': 'chromeos2-row1-rack5-host4'} repaired successfully
,
Apr 27 2016
|
|||
►
Sign in to add a comment |
|||
Comment 1 by jrbarnette@chromium.org
, Mar 24 2016It's a three line fix: (AutoUpdateRepair, 'au', ['ssh', 'good_au', 'tpm'], ['cros']), (PowerWashRepair, 'powerwash', ['ssh'], ['good_au', 'tpm', 'cros']), (ServoInstallRepair, 'usb', [], ['ssh', 'good_au', 'tpm', 'cros']), That is 'tpm' blocks AutoUpdateRepair, and triggers both PowerWashRepair and ServoInstallRepair