I think that test is flaky on link (at least if I'm reading wmatrix correctly).  I also don't see anything in the list of CLs in that run that looks even remotely responsible.

It also looks like the following run passed successfully: https://luci-milo.appspot.com/buildbot/chromeos/link-paladin/29297

How do I find out the list of CLs in that run?

https://luci-milo.appspot.com/buildbot/chromeos/link-paladin/29296

See the CommitQueueSync section.

Actually you can just look into the DetectRelevantChanges stage, only 3 CLs were relevant to the build.

Basically, if we can confirm https://chrome-internal-review.googlesource.com/#/c/chromeos/overlays/chromeos-overlay/+/427829/ is irrelevant to the failure, we can confirm it's a flake.

Given that the CL mentioned in #5 passed another CQ run, I'd be inclined to say the reported fail was a flake.

Moreover, looking at /var/log/messages from the device, we clearly see that something went wrong with opening user's profile (including TPM errors):

2017-08-14T17:47:10.996710+00:00 INFO cryptohomed[1468]: Asynced Mount() requested. Tracking request sequence id 16 for later PKCS#11 initialization.
2017-08-14T17:47:10.998908+00:00 WARNING cryptohomed[1468]: Could not load the device policy file.
2017-08-14T17:47:11.023934+00:00 INFO kernel: [ 1014.776233] tpm_tis tpm_tis: command 0x15 (size 14) returned code 0x0
2017-08-14T17:47:11.140981+00:00 INFO kernel: [ 1014.893222] tpm_tis tpm_tis: command 0x65 (size 18) returned code 0x0
2017-08-14T17:47:11.023352+00:00 WARNING cryptohomed[1468]: Could not load the device policy file.
2017-08-14T17:47:11.170217+00:00 ERR cryptohomed[1468]: TPM error 0x1 (Authentication failed): Error calling Tspi_Key_GetPubKey
2017-08-14T17:47:11.170969+00:00 INFO kernel: [ 1014.923122] tpm_tis tpm_tis: command 0x21 (size 14) returned code 0x1
2017-08-14T17:47:11.329943+00:00 INFO kernel: [ 1015.082222] tpm_tis tpm_tis: command 0x65 (size 18) returned code 0x0
2017-08-14T17:47:11.563326+00:00 ERR cryptohomed[1468]: TPM error 0x1 (Authentication failed): Error calling Tspi_Data_Unbind
2017-08-14T17:47:11.563973+00:00 INFO kernel: [ 1015.316033] tpm_tis tpm_tis: command 0x1e (size 274) returned code 0x1
2017-08-14T17:47:11.563490+00:00 ERR cryptohomed[1468]: The TPM failed to unwrap the intermediate key with the supplied credentials
2017-08-14T17:47:11.563631+00:00 ERR cryptohomed[1468]: Failed to decrypt any keysets for 24bd7ec77691224315ee2c0f6b83e68036a8a3f9
2017-08-14T17:47:11.564068+00:00 INFO cryptohomed[1468]: An asynchronous mount request with sequence id: 16 finished; doing PKCS11 init...
2017-08-14T17:47:11.601519+00:00 INFO session_manager[19088]: [INFO:policy_key.cc(53)] No policy key on disk at /home/root/24bd7ec77691224315ee2c0f6b83e68036a8a3f9/session_manager/policy/key
2017-08-14T17:47:11.602587+00:00 ERR session_manager[19088]: [ERROR:nss_util.cc(129)] Error opening persistent database (configDir='sql:/home/user/24bd7ec77691224315ee2c0f6b83e68036a8a3f9/.pki/nssdb' tokenDescription='/home/user/24bd7ec77691224315ee2c0f6b83e68036a8a3f9'): -8018
2017-08-14T17:47:11.602620+00:00 ERR session_manager[19088]: [ERROR:session_manager_impl.cc(1370)] Could not open the current user's NSS database.
2017-08-14T17:47:11.602652+00:00 ERR session_manager[19088]: [ERROR:dbus_util.cc(14)] CreateError(...): Domain=dbus, Code=org.chromium.SessionManagerInterface.NoUserNssDb, Message=Can't create session.

(https://storage.cloud.google.com/chromeos-autotest-results/135022432-chromeos-test/chromeos4-row5-rack13-host1/login_MultipleSessions/sysinfo/messages?_ga=2.142028021.-672223342.1502780285)

Agree this is a flake. None of the CLs could've caused this.

https://luci-milo.appspot.com/buildbot/chromeos/link-paladin/29306

http://cautotest/tko/retrieve_logs.cgi?job=/results/135203947-chromeos-test/

Failed the CQ again. A bug on TOT causes the flake? 

Based on #6 it seems like a TPM problem, which implies there's a device with bad hardware in the lab.  According to the flake dashboard (https://wmatrix.googleplex.com/testrun/?test_ids=526811726,530162520,532075189,532332965,532846096,533316388,520320236,523982219,527278666,530823779,535057942), 3 out of the 5 failures of this test in the last 30 days have happened on the same host (chromeos4-row9-rack2-host1), which further suggests a hardware problem with the tpm.

Sounds like a problem with the dut chromeos4-row5-rack13-host1, let me lock it and balance the pool.

Failed again. This time on chromeos4-row5-rack12-host21 (issue 758665).

Merge into the flakiness  issue 740839  for this test on link?

chromeos4-row5-rack13-host1 has been held locked three weeks with no action.
I'm unlocking the DUT.  If someone has cause to believe that there's something
specific wrong with the hardware, please post solid evidence before re-locking.