Flaky buildbot clang-x86_64-linux-selfhost-abi-test

Project Infrastructure LLVM Dev List Archives
1

This bot seems flaky too. In the last 100 builds there were 11 failures because on of these test had been failed:

FAIL: SanitizerCommon-tsan::signal_segv_handler.cc
FAIL: SanitizerCommon-msan::signal_segv_handler.cc
FAIL: SanitizerCommon-lsan::signal_segv_handler.cc

There were also 5 testing timeouts:

[325/325] Running all regression tests
command timed out: 1200 seconds without output, attempting to kill

  • PB
2

Thanks for bringing this up. We've seen these tests misbehave under
load on sanitizer bots, but not in the last few weeks, and we could
never reproduce it locally. I'll disable the test for now.

Btw, we've also seen this test survive whatever buildbot does in the
case of a step timeout and consuming the CPU cycles forever after
that. Please check the bot for runaway processes.

Sorry about that.