Closed bcmills closed 2 years ago
Compare #32534, #32919.
Same failure mode on solaris-amd64-oraclerel
: https://build.golang.org/log/05176237c5a06f48100a31a8b05997879c9f8ea1
darwin-amd64-10_12
: https://build.golang.org/log/5a6b1d2da89c39f0aa344ce5d9996aa44b8133a8
darwin-amd64-10_12
: https://build.golang.org/log/22a57819568de78654e834c54ee909a50826230b
A different (but likely related) failure mode:
--- FAIL: TestDialParallelSpuriousConnection (1.49s)
dial_test.go:456: got read tcp6 [::1]:59891->[::1]:59893: i/o timeout; want EOF
FAIL
FAIL net 4.495s
2020-02-29T17:02:40-74f8983/darwin-arm64-corellium 2020-02-24T16:39:52-3093959/darwin-arm64-corellium 2020-01-31T20:18:54-1b7fefc/darwin-arm64-corellium 2020-01-23T21:01:12-ace25f8/darwin-arm-mg912baios 2020-01-19T14:04:09-8e0be05/darwin-arm-mg912baios
More of the new failure mode: 2020-03-10T00:24:30-08dee51/darwin-arm64-corellium 2020-03-02T15:39:23-12d02e7/darwin-arm64-corellium
Looks like the old failure mode really did stop in October.
2019-10-31T23:37:57-ce49f95/aix-ppc64 2019-10-30T14:06:04-17190de/darwin-amd64-10_12 2019-10-23T07:43:18-ab3f1a2/darwin-amd64-10_12 2019-10-11T16:34:01-ba108c9/darwin-amd64-race 2019-10-08T17:30:14-868de9a/darwin-amd64-10_12 2019-10-02T17:24:11-0000f0b/solaris-amd64-oraclerel 2019-09-24T04:19:31-e29d276/darwin-amd64-race 2019-09-18T20:32:43-bcf6f9f/darwin-amd64-10_12 2019-09-16T22:14:49-115e4c9/darwin-amd64-race 2019-09-06T21:22:26-8a8cf5b/darwin-amd64-race 2019-06-25T17:51:25-a6ad626/darwin-amd64-race 2019-05-13T17:01:25-db2bf15/darwin-amd64-race 2019-04-30T18:29:07-888bac1/darwin-amd64-race
Actually, those new failures seem Corellium-specific. Opening a new issue.
Looks like the hiatus from October 2019 to March 2020 was only temporary.
greplogs --dashboard -md -l -e '^panic: test timed out.*\n(?:.*\n)*net\..*\.Accept.*\n\t.*\nnet\.TestDialParallelSpuriousConnection' --since=2020-01-01
2021-12-21T22:26:48-60f2c12/freebsd-arm64-dmgk 2021-12-02T20:21:16-469f030/netbsd-amd64-9_0-n2d 2021-10-27T08:50:27-bdefb77/openbsd-arm64-jsing 2021-08-26T02:06:43-a6ff433/netbsd-arm-bsiegert 2021-08-20T18:45:25-0f25251/openbsd-arm64-jsing 2021-08-13T20:45:17-98f3d7f/solaris-amd64-oraclerel 2021-06-24T18:58:27-cce6214/openbsd-arm64-jsing 2021-06-18T16:56:48-57aaa19/openbsd-arm64-jsing 2021-05-21T18:55:18-e4d7525/illumos-amd64 2021-05-13T14:30:42-fd4631e/openbsd-arm64-jsing 2021-05-11T02:46:21-326a792/illumos-amd64 2021-05-01T11:42:49-a9db5a7/openbsd-arm64-jsing 2021-04-19T23:02:15-7252e1e/darwin-amd64-10_12 2021-04-14T00:25:33-8dcc071/freebsd-arm64-dmgk 2021-04-09T13:10:36-c432917/netbsd-386-9_0 2021-03-29T08:11:13-565e70f/openbsd-arm64-jsing 2021-01-15T20:37:59-ec94701/darwin-amd64-10_12 2020-12-07T21:01:46-7ad6596/openbsd-arm64-jsing 2020-11-10T09:01:41-1c7650a/darwin-amd64-10_12 2020-06-26T17:07:58-d1015f3/solaris-amd64-oraclerel 2020-04-30T14:41:10-ecdbffd/solaris-amd64-oraclerel
Change https://golang.org/cl/375694 mentions this issue: net: synchronize instead of sleeping in TestDialParallelspuriousconnection
This is a recurring test failure and does not seem to be port-specific, so marking as release-blocker. I've mailed what I believe to be a fix.
From the
darwin-amd64-race
builder (https://build.golang.org/log/96aa6d382799b872cbedeb56e3b9d2076acc6546):It's not obvious to me whether the test was deadlocked or just ran out of time.
If the latter, perhaps some of its tests should be skipped in short mode (and/or when the race detector is enabled)?
CC @mikioh @bradfitz @ianlancetaylor