Found new dashboard test flakes for:

#!watchflakes
default <- pkg == "cmd/compile/internal/arm64" && test == ""

FAIL	cmd/compile/internal/arm64 [build failed]

— watchflakes

This is an OOM or timeout on netbsd/arm.

CC @golang/netbsd

There is some general flakiness on the two NetBSD/arm build machines. It appeared around the time I switched from gobuilder to LUCI, and also from NetBSD-9 to NetBSD-10. I am still working on getting to the bottom of this, sorry for the noise.

One of the symptoms I am seeing (unrelated?) is that the machine sometimes gets some processes stuck in the "tstile" state. This mostly happens with gopls tests. I suspect a kernel bug, but debugging is difficult because the issue is intermittent.

There is https://gnats.netbsd.org/58871 on the NetBSD side to track this.

Found new dashboard test flakes for:

#!watchflakes
default <- pkg == "cmd/compile/internal/arm64" && test == ""

FAIL	cmd/compile/internal/arm64 [build failed]
cmd/compile/internal/ssa: /home/swarming/.swarming/w/ir/x/w/goroot/pkg/tool/netbsd_arm/compile: signal: killed

— watchflakes

Found new dashboard test flakes for:

#!watchflakes
default <- pkg == "cmd/compile/internal/arm64" && test == ""

FAIL	cmd/compile/internal/arm64 [build failed]
cmd/compile/internal/ssa: /home/swarming/.swarming/w/ir/x/w/goroot/pkg/tool/netbsd_arm/compile: signal: killed

— watchflakes

Found new dashboard test flakes for:

#!watchflakes
default <- pkg == "cmd/compile/internal/arm64" && test == ""

FAIL	cmd/compile/internal/arm64 [build failed]
cmd/compile/internal/ssa: /home/swarming/.swarming/w/ir/x/w/goroot/pkg/tool/netbsd_arm/compile: signal: killed

— watchflakes