On 22/12/16 13:13, Pádraig Brady wrote:
> On 21/12/16 16:04, Dan Horák wrote:
>> On Tue, 20 Dec 2016 16:03:49 +0100
>>
>> Pavel Raiskup <praiskup(a)redhat.com> wrote:
>>> Hi all,
>>>
>>> where is documented what system/hw is used on (primary) Koji builders?
>>> I'm interested in memory, storage, filesystem, host operating system,
>>> guest operating system (if those are VMs), etc.
>>>
>>> The only thing I was able to find is version of mock in the log
>>> output.
>>>
>>> FWIW, I'd like to reproduce hang in gnulib's test-lock [1] test
case.
>>> Unless I'm able to reproduce that somehow, I'll disable the test
for
>>> the time being (done in 'coreutils' and I guess elsewhere, too).
>>
>> oh no, test-lock again :-( A year (or two) ago it turned to be a kernel
>> bug -
https://bugzilla.redhat.com/show_bug.cgi?id=1155291
>
> I saw Kamil recently disabled that test for the coreutils update.
> Note the test itself may have issues as discussed at:
>
https://lists.gnu.org/archive/html/bug-gnulib/2015-07/msg00032.html
Note this test was just changed upstream to use locks instead of volatiles,
which significantly improved performance on a 40 core NUMA system at least:
http://git.savannah.gnu.org/gitweb/?p=gnulib.git;a=commitdiff;h=480d374
Thanks for letting me know! I have applied the above patch on the coreutils
package in Fedora rawhide:
... and it seems to have resolved the issue with hanging Koji builds.
Kamil
> cheers,
> Pádraig