runtime: TestGdbBacktrace failures with "No such process." since 2021-10-15

[bcmills]

greplogs --dashboard -md -l -e 'FAIL: TestGdbBacktrace .*(?:\n .*)*: No such process\.'

2022-01-24T17:07:30-f88c3b9/linux-386
2021-11-05T20:59:43-dbd3cf8/linux-amd64-nocgo
2021-10-15T03:49:17-8dab959/linux-amd64-nocgo

This is a new failure mode for a known-flaky test (see #37405; CC @aclements, @jeremyfaller). Unfortunately, this failure mode appears to predate the CLs added for that issue.

[mknyszek]

Since this failure mode is old, I'm putting this in backlog. But if it looks like this is plaguing the dashboard enough, we might want to consider moving it into the 1.18 milestone.

[bcmills]

Since this failure mode is old,

Note that this particular failure mode is not old — it's new as of October. (The older failures for this test have different failure modes.)

[bcmills]

One more:

greplogs --dashboard -md -l -e 'FAIL: TestGdbBacktrace .*(?:\n .*)*: No such process\.' --since=2022-01-26

2022-02-10T01:08:52-656d3f4/linux-386

Note that these failures are so far all on first-class ports.

[bcmills]

This is a release blocker at least until a skip is added for this failure mode, then it can be moved back to the backlog.

(Given the other failure modes of this test, I don't think it's worth trying to get in a fix for the root cause during the freeze.)

[bcmills]

This looks like https://sourceware.org/bugzilla/show_bug.cgi?id=9086, open against gdb since 2005-08-03. 😩

[mknyszek]

Oof. I'll add the skip.

[bcmills]

I've got a CL pending already. I should also update the other skip with a comment for the GDB bug...

[mknyszek]

Ah OK. Feel free to send it to me.

[gopherbot]

Change https://go.dev/cl/385175 mentions this issue: runtime: update TestGdbBacktrace skips for known GDB bugs