runtime: Linq tests timing out in WASM leg
Linq tests appear to be timing out (or possibly hanging) in WASM
https://dev.azure.com/dnceng/public/_build/results?buildId=780635&view=ms.vss-test-web.build-test-results-tab&runId=24492990&resultId=164319&paneView=dotnet-dnceng.dnceng-build-release-tasks.helix-test-information-tab https://helixre107v0xdeko0k025g8.blob.core.windows.net/dotnet-runtime-refs-pull-41038-merge-558fe2197ef6431186/System.Linq.Tests/console.01a47de4.log?sv=2019-02-02&se=2020-09-09T20%3A59%3A51Z&sr=c&sp=rl&sig=Nv2ILnn693jzHBwsZrPiM8NJrLkYK7Qe2su3kYUaY78%3D
Executed on a003RYO
+ export PATH=/home/helixbot/work/A66808D0/p/dotnet:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/snap/bin
+ export DOTNET_ROOT=/home/helixbot/work/A66808D0/p/dotnet
+ export DOTNET_CLI_TELEMETRY_OPTOUT=1
+ export PATH=/home/helixbot/work/A66808D0/p/xharness-cli:/home/helixbot/work/A66808D0/p/dotnet:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/snap/bin
+ export XHARNESS_DISABLE_COLORED_OUTPUT=true
+ export XHARNESS_LOG_WITH_TIMESTAMPS=true
+ export XHARNESS_CLI_PATH=/home/helixbot/work/A66808D0/p/microsoft.dotnet.xharness.cli/1.0.0-prerelease.20411.1/tools/netcoreapp3.1/any/Microsoft.DotNet.XHarness.CLI.dll
+ ./RunTests.sh --runtime-path /home/helixbot/work/A66808D0/p
XHarness command issued: wasm test --engine=V8 --engine-arg=--stack-trace-limit=1000 --js-file=runtime.js -v --output-directory=./xharness-output -- --run WasmTestRunner.dll System.Linq.Tests.dll -notrait category=IgnoreForCI -notrait category=OuterLoop -notrait category=failing
[21:00:14] info: test[0]
21:00:14.6178730 v8 --expose_wasm --stack-trace-limit=1000 runtime.js -- --run WasmTestRunner.dll System.Linq.Tests.dll -notrait category=IgnoreForCI -notrait category=OuterLoop -notrait category=failing
[21:00:14] dbug: test[0]
Arguments: --run,WasmTestRunner.dll,System.Linq.Tests.dll,-notrait,category=IgnoreForCI,-notrait,category=OuterLoop,-notrait,category=failing
[21:00:15] dbug: test[0]
console.debug: MONO_WASM: Initializing mono runtime
[21:00:15] dbug: test[0]
console.debug: MONO_WASM: ICU data archive(s) loaded, disabling invariant mode
[21:00:15] dbug: test[0]
console.debug: mono_wasm_runtime_ready fe00e07a-5519-4dfe-b35a-f867dbaf2e28
[21:00:15] dbug: test[0]
Initializing.....
[21:00:16] dbug: test[0]
Discovering tests for System.Linq.Tests.dll
,...................
[21:15:10] dbug: test[0]
[PASS] System.Linq.Tests.SkipLastTests.SkipLast(source: [0, 1, 2, 3, 4, ...], count: 8)
[21:15:12] dbug: test[0]
[PASS] System.Linq.Tests.SkipLastTests.SkipLast(source: [0, 1, 2, 3, 4, ...], count: 13)
[21:15:13] dbug: test[0]
[PASS] System.Linq.Tests.SkipLastTests.SkipLast(source: [0, 1, 2, 3, 4, ...], count: 21)
Killed
About this issue
- Original URL
- State: closed
- Created 4 years ago
- Comments: 23 (23 by maintainers)
Commits related to this issue
- [Wasm] Send ToArrayShouldWorkWithSpecialLengthLazyEnumerables Linq test to the outer loop Addresses the occasional timeouts in System.Linq tests. The last 3-4 iterations of ToArrayShouldWorkWithSpec... — committed to steveisok/runtime by deleted user 4 years ago
- [Wasm] Send ToArrayShouldWorkWithSpecialLengthLazyEnumerables Linq test to the outer loop Addresses the occasional timeouts in System.Linq tests. The last 3-4 iterations of ToArrayShouldWorkWithSpec... — committed to dotnet/runtime by deleted user 4 years ago
- [Wasm] Reduce max length on browser for ToArrayShouldWorkWithSpecialLengthLazyEnumerables Linq test (#41241) Addresses the occasional timeouts in System.Linq tests. The last 3-4 iterations of ToArra... — committed to dotnet/runtime by steveisok 4 years ago
- [release/5.0] [Wasm] Reduce max length on browser for ToArrayShouldWorkWithSpecialLengthLazyEnumerables Linq test (#41268) Addresses the occasional timeouts in System.Linq tests. The last 3-4 iterat... — committed to dotnet/runtime by github-actions[bot] 4 years ago
- Runtime ingestion (#73) * Applied BinaryPrimitives throughout CoreLib (#37582) * Flaky multiple HTTP/2 connection tests read full request body (#40844) There are 2 flaky HTTP/2 tests verifying ... — committed to dotnet/runtimelab by joperezr 4 years ago
My reservation about increasing the timeout is that it potentially increases the time of our PR validation when it’s not clear whether we’re getting value in return. If I added some new Linq tests today, and it caused the full set to take 16 minutes on WASM, I wouldn’t push another change to increase the timeout. I’d scale back the test. Is there any reason to believe that 15 mins isn’t a reasonable quota to test all of Linq innerloop (on a machine that’s not occupied with any other task) – when no other library we have needs so long?
I took a closer look at the Kusto data between a run that times out and a good run. It seems both have the same amount of records and their durations are ~13 min. I suspect a longer GC plays a part in the runs that time out. Wasm does not have the benefit of a separate GC thread.
I’m asking around if there are some GC knobs to tweak, but I feel like increasing the timeout might actually be the prudent choice.
We should disable and link to this issue while we investigate.