Bug #34711 for Time-HiRes: Occasional failures of test 35 or 36 (FreeBSD)

Sat Apr 05 17:21:38 2008 SREZIC [...] cpan.org - Ticket created

Subject:

Occasional failures of test 35 or 36 (FreeBSD)

(version 1.9713) The tests 35 and 36 seem to fail sometimes on my FreeBSD machine (you already got a FAIL report). Attached is a log of a loop testing Time-HiRes running about 20 minutes under heavy load (load average 8-10) until the first failure hit. Here's a list of all r values for the test 36 from the attached log file: $ grep -B 1 'ok 36' /tmp/timehires.log |grep '# r' # r = 1.00059097463434 # r = 1.00590781732039 # r = 1.01816675879739 # r = 1.0113263130188 # r = 1.00092356855219 # r = 1.41417860984802 # r = 1.00736184553667 # r = 1.0004723072052 # r = 1.00633946332065 # r = 1.00755182179538 # r = 1.00052226673473 # r = 1.14257855848833 # r = 1.00050676952709 # r = 1.05191685936668 # r = 1.00051912394437 # r = 1.00086450576782 # r = 1.01550774140791 # r = 1.00059769370339 # r = 1.00049051371488 # r = 1.00913134488192 # r = 1.00039872256192 # r = 1.00589275360107 # r = 1.00050135092302 # r = 1.36329369111495 # r = 1.0049709406766 # r = 1.00041324442083 # r = 1.02450825951316 # r = 1.00414091890508 # r = 1.00646582516757 # r = 1.00047317418185 # r = 1.68967051939531 Let me know if you need more data. Regards, Slaven

Subject:

timehires.log

Message body is not shown because it is too large.

Sat Apr 05 20:32:30 2008 jhi [...] iki.fi - Correspondence added

Subject:	Re: [rt.cpan.org #34711] Occasional failures of test 35 or 36 (FreeBSD)
Date:	Sat, 05 Apr 2008 20:31:19 -0400
To:	bug-Time-HiRes [...] rt.cpan.org
From:	Jarkko Hietaniemi <jhi [...] iki.fi>

Slaven_Rezic via RT wrote: Show quoted text

> Sat Apr 05 17:21:38 2008: Request 34711 was acted upon. > Transaction: Ticket created by SREZIC > Queue: Time-HiRes > Subject: Occasional failures of test 35 or 36 (FreeBSD) > Broken in: (no value) > Severity: (no value) > Owner: Nobody > Requestors: SREZIC@cpan.org > Status: new > Ticket <URL: http://rt.cpan.org/Ticket/Display.html?id=34711 > > > > (version 1.9713) > > The tests 35 and 36 seem to fail sometimes on my FreeBSD machine (you > already got a FAIL report). Attached is a log of a loop testing > Time-HiRes running about 20 minutes under heavy load (load average 8-10) > until the first failure hit.

It is well known that under heavy load some of the Time::HiRes tests fail, but I don't know that I can do much about this without making up some rule about "heavy load" and then retrying the tests (how many times?) Show quoted text

> Here's a list of all r values for the test 36 from the attached log file: > $ grep -B 1 'ok 36' /tmp/timehires.log |grep '# r' > # r = 1.00059097463434 > # r = 1.00590781732039 > # r = 1.01816675879739 > # r = 1.0113263130188 > # r = 1.00092356855219 > # r = 1.41417860984802 > # r = 1.00736184553667 > # r = 1.0004723072052 > # r = 1.00633946332065 > # r = 1.00755182179538 > # r = 1.00052226673473 > # r = 1.14257855848833 > # r = 1.00050676952709 > # r = 1.05191685936668 > # r = 1.00051912394437 > # r = 1.00086450576782 > # r = 1.01550774140791 > # r = 1.00059769370339 > # r = 1.00049051371488 > # r = 1.00913134488192 > # r = 1.00039872256192 > # r = 1.00589275360107 > # r = 1.00050135092302 > # r = 1.36329369111495 > # r = 1.0049709406766 > # r = 1.00041324442083 > # r = 1.02450825951316 > # r = 1.00414091890508 > # r = 1.00646582516757 > # r = 1.00047317418185 > # r = 1.68967051939531 > > Let me know if you need more data. > > Regards, > Slaven >

Sat Apr 05 20:32:41 2008 The RT System itself - Status changed from 'new' to 'open'

Tue Apr 08 17:11:19 2008 SREZIC [...] cpan.org - Correspondence added

On Sat Apr 05 20:32:30 2008, jhi@iki.fi wrote: Show quoted text

> Slaven_Rezic via RT wrote:

> > Sat Apr 05 17:21:38 2008: Request 34711 was acted upon. > > Transaction: Ticket created by SREZIC > > Queue: Time-HiRes > > Subject: Occasional failures of test 35 or 36 (FreeBSD) > > Broken in: (no value) > > Severity: (no value) > > Owner: Nobody > > Requestors: SREZIC@cpan.org > > Status: new > > Ticket <URL: http://rt.cpan.org/Ticket/Display.html?id=34711 > > > > > > > (version 1.9713) > > > > The tests 35 and 36 seem to fail sometimes on my FreeBSD machine (you > > already got a FAIL report). Attached is a log of a loop testing > > Time-HiRes running about 20 minutes under heavy load (load average 8-10) > > until the first failure hit.

> > It is well known that under heavy load some of the Time::HiRes tests > fail, but I don't know that I can do much about this without making up > some rule about "heavy load" and then retrying the tests (how many times?) >

Maybe you can always retry the failing tests, regardless whether there's heavy load or not (I doubt there's some good rule to determine if there's heavy load --- a high load average number may be completely harmless, but a low load average at the beginning or end of the test does not tell anything about some peaks in the middle) And I think you can retry the tests many times --- why not 10 or 100 times? As for failing tests, here's a list of failures (with the total count of failures in the first column) while running the Time-HiRes test more than 1000 times: 48 not ok 35 2 not ok 36 9 not ok 6 2 not ok 8 Regards, Slaven

Tue Apr 08 20:11:56 2008 jhi [...] iki.fi - Correspondence added

Subject:	Re: [rt.cpan.org #34711] Occasional failures of test 35 or 36 (FreeBSD)
Date:	Tue, 08 Apr 2008 20:09:33 -0400
To:	bug-Time-HiRes [...] rt.cpan.org
From:	Jarkko Hietaniemi <jhi [...] iki.fi>

I now uploaded 1.9715 which does a little bit of retrying (though not 100 times, that would be silly) for some of the tests.

Wed Apr 09 12:20:32 2008 jhi [...] iki.fi - Correspondence added

Subject:	Re: [rt.cpan.org #34711] Occasional failures of test 35 or 36 (FreeBSD)
Date:	Wed, 9 Apr 2008 12:20:03 -0400
To:	bug-Time-HiRes [...] rt.cpan.org
From:	"Jarkko Hietaniemi" <jhi [...] iki.fi>

On Tue, Apr 8, 2008 at 8:12 PM, jhi@iki.fi via RT <bug-Time-HiRes@rt.cpan.org> wrote: Show quoted text

> > Queue: Time-HiRes > Ticket <URL: http://rt.cpan.org/Ticket/Display.html?id=34711 > > > I now uploaded 1.9715 which does a little bit of retrying (though not > 100 times, that would be silly) for some of the tests.

If you can get the 34...37 now to fail, I have failed :-) Show quoted text

> >

-- There is this special biologist word we use for 'stable'. It is 'dead'. -- Jack Cohen

Thu Apr 10 15:17:07 2008 SREZIC [...] cpan.org - Correspondence added

Thanks, you've got six PASSes from me. No FAILs anymore. Regards, Slaven

Thu Apr 10 15:17:10 2008 SREZIC [...] cpan.org - Status changed from 'open' to 'resolved'

Thu Apr 10 21:37:14 2008 jhi [...] iki.fi - Correspondence added

Subject:	Re: [rt.cpan.org #34711] Occasional failures of test 35 or 36 (FreeBSD)
Date:	Thu, 10 Apr 2008 21:36:54 -0400
To:	bug-Time-HiRes [...] rt.cpan.org
From:	Jarkko Hietaniemi <jhi [...] iki.fi>

Slaven_Rezic via RT wrote: Show quoted text

> Queue: Time-HiRes > Ticket <URL: http://rt.cpan.org/Ticket/Display.html?id=34711 > > > Thanks, you've got six PASSes from me. No FAILs anymore. > > Regards, > Slaven > >

I didn't yet cover all the failures you saw in the retry wrapping, but hopefully this will help...

Thu Apr 10 21:37:15 2008 The RT System itself - Status changed from 'resolved' to 'open'

Tue Apr 29 22:15:32 2008 JHI [...] cpan.org - Status changed from 'open' to 'resolved'