You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Timeouts in System.Text.Json.Tests on arm32 #67878
Console log: 'System.Text.Json.Tests' from job 0e2f0129-97a5-4449-a6f1-4f6d48f286d1 (ubuntu.1804.armarch.open) using docker image mcr.microsoft.com/dotnet-buildtools/prereqs:ubuntu-18.04-helix-arm32v7-bfcd90a-20200121150440 on ddvsotx2l053
running $HELIX_CORRELATION_PAYLOAD/scripts/437cee12e49841d097023962d57d444e/execute.sh in /ssd/work/A22408F2/w/A0840945/e max 2700 seconds
=================
Console log: 'System.Text.Json.Tests' from job cc08232a-0565-4446-a309-8e67f6b5ff0a (ubuntu.1804.armarch.open) using docker image mcr.microsoft.com/dotnet-buildtools/prereqs:alpine-3.14-helix-arm32v7-20210910135806-8a6f4f3 on ddvsotx2l143
running $HELIX_CORRELATION_PAYLOAD/scripts/00f3104607ef4bc4ab7713de646b08be/execute.sh in /ssd/work/A07A0914/w/B302093F/e max 2700 seconds
This was likely regressed by @krwq in #67838. That slowed things down, which would explain why it's only on Arm; it's also debug only, which is what we see; and it began happening fairly consistently at the moment that PR went in.
I think it's easier for me to disable those asserts on Arm.
BTW I wrote a script to read the console output for all the failures, and these are the tests that fail -- perhaps the slowest tests that run through the asserts?
Activity
jkoritzinsky commentedon Apr 12, 2022
Happening here as well: #67676 (comment)
elinor-fung commentedon Apr 12, 2022
Hit again in #67635.
A bunch of different tests being detected as long running - from one of the logs:
elinor-fung commentedon Apr 12, 2022
There over 50 hits in the past day - marking as blocking-clean-ci. Nothing in the week before that though - some recent change?
danmoseley commentedon Apr 13, 2022
cc @ViktorHofer as test monitor -- is this always 32 bit ARM? If so can we disable just for that?
eiriktsarpalis commentedon Apr 14, 2022
@krwq would it make sense to move some of the tests listed in #67878 (comment) to OuterLoop?
danmoseley commentedon Apr 14, 2022
I'll disable affected tests for Arm32.
ViktorHofer commentedon Apr 14, 2022
I didn't notice these timeouts as I only looked at test failures yesterday... Thanks Dan for disabling them.
danmoseley commentedon Apr 14, 2022
This was likely regressed by @krwq in #67838. That slowed things down, which would explain why it's only on Arm; it's also debug only, which is what we see; and it began happening fairly consistently at the moment that PR went in.
I think it's easier for me to disable those asserts on Arm.
BTW I wrote a script to read the console output for all the failures, and these are the tests that fail -- perhaps the slowest tests that run through the asserts?
danmoseley commentedon Apr 14, 2022
If someone can sign off on #68027, I'm inclined to merge it immediately since this is failing once every hour.
krwq commentedon Apr 14, 2022
Signed off, sorry for the slowdown