[issue23680] Sporadic freeze in test_interrupted_write_retry_text

Gerrit Holl added the comment: I experience this problem when trying to build/test Python 3.6 on the [JASMIN Analysis Platform](http://www.jasmin.ac.uk/services/jasmin-analysis-platform/) which runs Red Hat Enterprise Linux Server release 6.8 on a machine with 48 × Intel(R) Xeon(R) CPU E7-4860

[issue23680] Sporadic freeze in test_interrupted_write_retry_text

Martin Panter added the comment: Here is my suggested change to use setitimer(). I also closed the pipe, which means there is no need to time out the read, speeding the test up a bit more. -- keywords: +patch stage: -> patch review versions: +Python 2.7 Added file: http://bugs.python.o

[issue23680] Sporadic freeze in test_interrupted_write_retry_text

Martin Panter added the comment: I found a relatively recent case of this failing on a buildbot: http://buildbot.python.org/all/builders/AMD64%20FreeBSD%209.x%203.x/builds/4339/steps/test/logs/stdio == CPython 3.6.0a1+ (default:eaee5aed6fbc, May 27 2016, 05:52:21) [GCC 4.2.1 20070831 patched [F

[issue23680] Sporadic freeze in test_interrupted_write_retry_text

Roundup Robot added the comment: New changeset 10acab2d4a88 by Victor Stinner in branch 'default': Issue #23680: Reduce risk of race condition in check_interrupted_write() of https://hg.python.org/cpython/rev/10acab2d4a88 -- ___ Python tracker

[issue23680] Sporadic freeze in test_interrupted_write_retry_text

STINNER Victor added the comment: > Also, I think there another issue in that test. It uses `N = > support.PIPE_MAX_SIZE` but I think you want `N = support.PIPE_MAX_SIZE + 1` > if you want the raw write() to block. PIPE_MAX_SIZE is much larger than the effictive size of a pipe on Linux. It's d

[issue23680] Sporadic freeze in test_interrupted_write_retry_text

STINNER Victor added the comment: 2015-03-16 17:41 GMT+01:00 Antoine Pitrou : > Note that PIPE_MAX_SIZE can be large. Perhaps move the memory allocation > (i.e. `item * (support.PIPE_MAX_SIZE // len(item) + 1)`) before the alarm(1) > call? Good idea, the first strace shows that SIGALRM was rec

[issue23680] Sporadic freeze in test_interrupted_write_retry_text

Antoine Pitrou added the comment: Also, I think there another issue in that test. It uses `N = support.PIPE_MAX_SIZE` but I think you want `N = support.PIPE_MAX_SIZE + 1` if you want the raw write() to block. There is another test that can have a race condition: check_interrupted_write(). ---

[issue23680] Sporadic freeze in test_interrupted_write_retry_text

Roundup Robot added the comment: New changeset a18f7508649b by Victor Stinner in branch 'default': Issue #23680: Reduce risk of race condition in check_interrupted_write_retry() https://hg.python.org/cpython/rev/a18f7508649b -- nosy: +python-dev ___ P

[issue23680] Sporadic freeze in test_interrupted_write_retry_text

Antoine Pitrou added the comment: Note that PIPE_MAX_SIZE can be large. Perhaps move the memory allocation (i.e. `item * (support.PIPE_MAX_SIZE // len(item) + 1)`) before the alarm(1) call? Or what if alarm_interrupt is simply set up to retrigger the signal? e.g. instead of: def alarm_int

[issue23680] Sporadic freeze in test_interrupted_write_retry_text

STINNER Victor added the comment: > The system is rather slow (a shared VPS instance). In the trace you can > SIGALRM is triggered before the first write() call (or so it seems). Yep. It remembers me my old idea to make "sleep configurable" in tests: issue #20910. Most of the time, 1 second i

[issue23680] Sporadic freeze in test_interrupted_write_retry_text

Antoine Pitrou added the comment: The system is rather slow (a shared VPS instance). In the trace you can SIGALRM is triggered before the first write() call (or so it seems). -- ___ Python tracker

[issue23680] Sporadic freeze in test_interrupted_write_retry_text

STINNER Victor added the comment: The bug only occurs in Python 3.5, right? > Sometimes the test suite freezes in test_interrupted_write_retry_text > (test.test_io.CSignalsTest). How do you run the test suite? Is your system heavily loaded? Is it "fast"? This is an obvious race condition in t

[issue23680] Sporadic freeze in test_interrupted_write_retry_text

New submission from Antoine Pitrou: Sometimes the test suite freezes in test_interrupted_write_retry_text (test.test_io.CSignalsTest). Corresponding strace is: write(1, "test_interrupted_write_retry_tex"..., 66) = 66 rt_sigaction(SIGALRM, {0x00, [], SA_RESTORER, 0x7fd1ada2d0a0}, {SIG_DFL,

[issue23680] Sporadic freeze in test_interrupted_write_retry_text

Changes by Antoine Pitrou : -- components: +Tests type: -> behavior ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscr