From: Koichi Sasada <ko1@...>
Date: 2018-07-13T12:39:12+09:00
Subject: [ruby-core:87932] Re: [ruby-alerts:8102] failure alert on trunk-mjit@silicon-docker (NG (r63961))

Yes, this machine running many tests at a moment.

 > top - 12:36:03 up 133 days, 21:49,  1 user,  load average: 62.74, 
57.58, 51.04

on 32 cores.

 > $ docker ps | wc
 >     15     162    2106

each container runs parallel tests.


`RUBY_TEST_SUBPROCESS_TIMEOUT_SCALE` multiply waiting time and some 
tests (with asserts) set `RUBY_TEST_SUBPROCESS_TIMEOUT_SCALE=10`.

 > trunk-mjit:           --configure_opts=cppflags="-DMJIT_FORCE_ENABLE" 
                             --add-env=RUBY_TEST_SUBPROCESS_TIMEOUT_SCALE=10

mjit also set this scale option.


On 2018/07/13 12:19, Eric Wong wrote:
> ko1c-failure@atdot.net wrote:
>> Alert on trunk-mjit@silicon-docker
>> rsult : NG (r63961)
>> detail: https://gist.github.com/dce079536bdaab29c141314e63c69c73
>>
>> desc:
>> rev: 63961
>> test-all : 20213 tests, 4538727 assertions, 0 failures, 1 errors, 102 skips
>> :   1)
>> : TestSignal#test_sigchld_ignore:
>> : Timeout::Error: execution expired
> 
> I suspect this was due to high load on the system with "-j6".
> Increased timeout with r63962...
> Can you dump /proc/loadavg on test failures?
> 
> I don't know if you have any other monitoring on the machine,
> either (we could check swap-in/out and iostat)
> 
> Thanks.
> 

-- 
// SASADA Koichi at atdot dot net

Unsubscribe: <mailto:ruby-core-request@ruby-lang.org?subject=unsubscribe>
<http://lists.ruby-lang.org/cgi-bin/mailman/options/ruby-core>