@arlakshm for awareness

Looks like the pipeline failed with an Azure authentication error:

##[error]Error Code: [1]
##[error]Error: Azure login failed
##[error]Script failed with error: ERROR: AADSTS500031: Cannot find signing certificate configured. Trace ID: b11efad7-3c3b-4607-9c3f-d81b4ce04500 Correlation ID: e76892ba-b02e-4a05-acb9-1a03c19be3b7 Timestamp: 2024-12-05 22:00:55Z
Interactive authentication is needed. Please run:
az login

/Azp run Azure.sonic-mgmt

/azpw run Azure.sonic-mgmt

/AzurePipelines run Azure.sonic-mgmt

@arista-hpandya @wenyiz2021 I suggest add the offending error logs to the issue and assess if we should white-list them instead? disabling the log analyzer will generally ignore all issues could be caught by log analyzer therefore leaving windows for regression.

@arista-hpandya @wenyiz2021 I suggest add the offending error logs to the issue and assess if we should white-list them instead? disabling the log analyzer will generally ignore all issues could be caught by log analyzer therefore leaving windows for regression.

@yxieca , here the error logs happened due to reboot, since the other reboot cases are disabling loganalyzer, and this test are all reboot tests except 1 config reload test.

I think it would be hard to whitelist all errors due to reboot.
If we really want to do whitelist, then it should be applicable for other reboot tests?

@arista-hpandya @wenyiz2021 I suggest add the offending error logs to the issue and assess if we should white-list them instead? disabling the log analyzer will generally ignore all issues could be caught by log analyzer therefore leaving windows for regression.

Hi @yxieca and @wenyiz2021 , thank you for reviewing the change. To my knowledge the current process of ignoring logs is via adding them to ansible/roles/test/files/tools/loganalyzer/loganalyzer_common_ignore.txt. The issue with this approach is that it is not test specific. As you rightly mentioned many of these issues raised by loganalyzer are valid issues/regressions to look at. In fact many of them are reproducible in other tests and are being actively worked upon. Adding these logs in the ignore file will hide these errors from all tests and not just the voq_disrupts tests.

If you wish I can open an issue for these untracked loganalyzer errors which would ensure that they don't slide away while giving the voq reboot tests the ability to run and verify reboot related checks

@arista-hpandya @wenyiz2021 I suggest add the offending error logs to the issue and assess if we should white-list them instead? disabling the log analyzer will generally ignore all issues could be caught by log analyzer therefore leaving windows for regression.

Hi @yxieca and @wenyiz2021 , thank you for reviewing the change. To my knowledge the current process of ignoring logs is via adding them to ansible/roles/test/files/tools/loganalyzer/loganalyzer_common_ignore.txt. The issue with this approach is that it is not test specific. As you rightly mentioned many of these issues raised by loganalyzer are valid issues/regressions to look at. In fact many of them are reproducible in other tests and are being actively worked upon. Adding these logs in the ignore file will hide these errors from all tests and not just the voq_disrupts tests.

The log analyzer infrastructure support adding whitelist to a single test case.

With that said, reboot related error logs is long and Wenyi is right, other tests currently also choose to disable log analyzer.

there is a way to ignore certain syslog on the testcase
https://github.com/sonic-net/sonic-mgmt/blob/4aa835e441f6c35a7c35f8f7ba62ae09971e3f76/tests/autorestart/test_container_autorestart.py#L82C5-L82C42

Thank you @wenyiz2021 and @yxieca for the prompt response. Based on the discussion can we go ahead and merge this change for consistency across all reboot test cases.

Cherry-pick PR to 202405: #16098