Crash of mono interpreter tests with SIGSEV when running ODBC tests

[ericstj]

Build Information

Build: https://dev.azure.com/dnceng-public/cbb18261-c48f-4abb-8651-8cdcb5474649/_build/results?buildId=1063262 Build error leg or test failing: Work Item System.Data.Odbc.Tests Pull request: https://github.com/dotnet/runtime/pull/116459

Error Message

Fill the error message using step by step known issues guidance.

{
  "ErrorMessage": ["MONO_ENV_OPTIONS=--interpreter", "Got a SIGSEGV while executing native code."],
  "ErrorPattern": "",
  "BuildRetry": false,
  "ExcludeConsoleLog": false
}

Could be related to https://github.com/dotnet/runtime/issues/114123

Two test logs, both System.Data.Odbc from the same commit:

• Test log dump
• Test log dump

Seems too consistent to be a coincidence.

Known issue validation

Build: :mag_right: https://dev.azure.com/dnceng-public/public/_build/results?buildId=1063262 Error message validated: [MONO_ENV_OPTIONS=--interpreter Got a SIGSEGV while executing native code.] Result validation: :white_check_mark: Known issue matched with the provided build. Validation performed at: 6/10/2025 3:27:56 PM UTC

Report

Build	Definition	Test	Pull Request
1064508	dotnet/runtime	System.Data.Odbc.Tests.WorkItemExecution
1064196	dotnet/runtime	System.Data.Odbc.Tests.WorkItemExecution
1063870	dotnet/runtime	System.Data.Odbc.Tests.WorkItemExecution
1063262	dotnet/runtime	System.Data.Odbc.Tests.WorkItemExecution

Summary

24-Hour Hit Count	7-Day Hit Count	1-Month Count
0	4	4

[lewing]

@BrzVlad I don't recall seeing this crash before, can you take a look?

[dotnet-policy-service[bot]]

Tagging subscribers to this area: @brzvlad, @kotlarmilos See info in area-owners.md if you want to be subscribed.

[BrzVlad]

This is same issue as https://github.com/dotnet/runtime/issues/101370. This is fixed on main, but didn't seem particularly important for a backport. 9.0.5 also manifests this crash.

[ericstj]

Not sure how common a customer application is to hit this, but typically crashing bugs like this meet the bar for servicing. I raised this because we didn't have a known issue tracking it and we need to be able to sort through the noise in servicing - these test runs are a big portion of what gives us a quality read in servicing where we have less folks using the builds before release.

[BrzVlad]

This issue would be reproducible if you have a PInvoke that fails to resolve. So, in practice, this is probably not a likely scenario since I guess it means you already have an issue with your application. I wouldn't consider it high priority, but we could backport it as well, the fix is low risk.

[vitek-karas]

@BrzVlad could you please decide if we will backport this? If not, please close this.

[BrzVlad]

In my opinion it is not worth to backport the fix. I would still keep this issue open while net9 is active so it is referenced for failing builds. Alternatively I guess we could disable this test suite on interpreter, on net9 branch if you think this is a better approach.

[vitek-karas]

I'd be fine disabling the test - either a single test case if it can be focused, or even the whole suite if necessary. There's no value in running it as it will fail all the time.

[vitek-karas]

@BrzVlad Should we resolve this - it's fixed in 10 and we're not planning to fix it in 9.