Repository navigation
[ci-scan] Hang: System.Diagnostics.Process.Tests times out on windows-x86 Checked CoreCLR #135001
Description
Build Information
Build: https://dev.azure.com/dnceng-public/public/_build/results?buildId=1618965
Build error leg or test failing: windows-x86 Debug Libraries_CheckedCoreCLR-System.Diagnostics.Process.Tests
Pull request: #134960
This is a recurrence of #131389 (the same System.Diagnostics.Process.Tests whole-work-item hang on Windows), which was closed on 2026-09-21 after 30 days without failures. That one was on windows-x64 Release NativeAOT_Libs; this one is on CoreCLR windows-x86 Checked. Build Analysis reports a 2.61% failure rate for this work item.
KBE authoring guidance (ci-failure-scan)
Error Detailsis for readers. Paste the full exception, stack trace, or build error excerpt so the failure is understandable without opening the raw log.- Build Analysis parses only the single JSON block under
Error Message. ErrorMessageis a case-sensitive ordinalString.Containssubstring copied verbatim from the failing log.- Set
BuildRetrytotrueonly for a clear infrastructure retry case.ExcludeConsoleLogdisables Helix console-log scanning.
Error Details
The work item starts and reports a couple of [SKIP] lines. After that no further output appears, until the Helix executor kills it (exit code -3, about 15m35s). No [FAIL] line is emitted, so the hanging test is not identified.
"...\dotnet.exe" exec --runtimeconfig System.Diagnostics.Process.Tests.runtimeconfig.json --depsfile System.Diagnostics.Process.Tests.deps.json xunit.console.dll System.Diagnostics.Process.Tests.dll -xml testResults.xml -nologo -nocolor -notrait category=IgnoreForCI -notrait category=OuterLoop -notrait category=failing
Discovering: System.Diagnostics.Process.Tests (method display = ClassAndMethod, method display options = None)
Discovered: System.Diagnostics.Process.Tests (found 423 of 451 test cases)
Starting: System.Diagnostics.Process.Tests (parallel test collections = on [4 threads], stop on fail = off)
System.Diagnostics.Tests.ProcessStartInfoTests.ShellExecute_Nano_Fails_Start [SKIP]
Condition(s) not met: "IsWindowsNanoServer"
System.Diagnostics.Tests.ProcessTests.NonElevatedUser_QueryProcessNameOfSystemProcess [SKIP]
Condition(s) not met: "IsNotPrivilegedProcess"
Invalid number of parameters
0 File(s) copied
['System.Diagnostics.Process.Tests' END OF WORK ITEM LOG: Command timed out, and was killed]
Configuration: net11.0-windows-Debug-x86-coreclr_checked-Windows.10.Amd64.Open, Helix job e45775cb-103a-475f-9f07-250ab1a48c98.
Error Message
{
"ErrorMessage": ["Starting: System.Diagnostics.Process.Tests", "['System.Diagnostics.Process.Tests' END OF WORK ITEM LOG: Command timed out, and was killed]"],
"ErrorPattern": "",
"BuildRetry": false,
"ExcludeConsoleLog": false
}Agentic workflow metadata (ci-failure-scan)
Workflow artifact: ci-failure-scan
Artifact kind: kbe-verification
Verified match count: 1 hits in failure.log
Note
This issue was drafted with GitHub Copilot assistance.
Known issue validation
Build: 🔎 https://dev.azure.com/dnceng-public/public/_build/results?buildId=1618965
Error message validated: [Starting: System.Diagnostics.Process.Tests ['System.Diagnostics.Process.Tests' END OF WORK ITEM LOG: Command timed out, and was killed]]
Result validation: ✅ Known issue matched with the provided build.
Validation performed at: 10/1/2026 4:30:43 AM UTC
Report
Summary
| 24-Hour Hit Count | 7-Day Hit Count | 1-Month Count |
|---|---|---|
| 0 | 7 | 215 |
Activity
dotnet-policy-service commentedon Oct 1, 2026
Tagging subscribers to this area: @dotnet/area-system-diagnostics-process
See info in area-owners.md if you want to be subscribed.
steveisok commentedon Oct 3, 2026
Investigation update
The observed failure is consistent with the suite exhausting its 900-second budget through accumulated execution time, not a demonstrated single-test hang. In build 1621730, 632 cases finished with reported durations totaling about 883 seconds before Helix killed the work item (exit -3). The eight-minute diagnostic snapshot did not terminate the host; execution continued afterward. The last test started at the cutoff is not established as the culprit.
A passing sample on the same x86 Checked configuration and Windows Server 2016 version took 640 seconds. Across 546 matching passed cases, the timeout run took 707 seconds versus 470 seconds in the passing run. The cause of that broad slowdown remains unknown, and these observations do not establish the cause of every historical match. The relationship to #131389 is unproven: its previously identified
StartInfo_BadExe(true)shell/StateRepository hang did not reproduce in our instrumented runs.The mitigation is #135160: six balanced Windows x86 CoreCLR Helix work items, preserving coverage and sequential execution within each shard, with no new skips or timeout increase. In build 1622015, all six passed in 2m52s-3m51s including setup; combined results preserved 691 cases (688 passed, 3 skipped) with no missing or duplicated shard coverage. This validated the identical sharding logic in investigation PR #131907; the new focused PR removes its temporary diagnostic CI routing. #131907 is now closed as superseded.
The mitigation is proposed, not yet merged; it addresses timeout exposure without claiming to fix or explain the underlying slowdown.
Note
This investigation update was prepared with GitHub Copilot assistance.
Enable class-level parallelism in Process tests (#135160)
[release/11.0] Enable class-level parallelism in Process tests (#135315)