New submission from STINNER Victor: The bpo-31151 (socketserver.ForkingMixIn.server_close() leaks zombie processes) was fixed in the master branch, but 2.7 and 3.6 branches were not changed.
I proposed to modify test_socketserver to explicitly waits until child processes complete. See attached PR. Example of leaked child processes the on "x86 Gentoo Refleaks 3.6" buildbot: http://buildbot.python.org/all/builders/x86%20Gentoo%20Refleaks%203.6/builds/102/steps/test/logs/stdio 0:26:11 load avg: 5.72 [ 62/405] test_socketserver passed -- running: test_multiprocessing_forkserver (594 sec) beginning 6 repetitions 123456 Warning -- reap_children() reaped child process 6891 .Warning -- reap_children() reaped child process 6976 Warning -- reap_children() reaped child process 6981 .Warning -- reap_children() reaped child process 7059 .Warning -- reap_children() reaped child process 7136 ..Warning -- reap_children() reaped child process 7306 . ---------- messages: 303040 nosy: haypo priority: normal severity: normal status: open title: test_socketserver ForkingMixIn tests leaks child processes versions: Python 3.6 _______________________________________ Python tracker <rep...@bugs.python.org> <https://bugs.python.org/issue31593> _______________________________________ _______________________________________________ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
