Re: gcc problems on stalder
Hi,
please try to build the package on sompek, which is exactly the same hardware.
Years ago on solaris and US IIi machines building firefox and friends also at
random times failed if you tried to compile it with -j2 or more, but as the
machines were used a lot and under heavy workload, and we never had a problem
with them, we never believed that the hardware was faulty.
stadler or somepek is missing one fan which might be the issue in case the cpu
is overheating, but I doubt that. Please try sompek and see if the issue shows
up there, too, and if not, I'll investigate if there is some hiddne hardware
issue on stadler.
cheers,
bernd
On 07/20/2013 03:33 AM, Kurt Roeckx wrote:
> So I'm still seeing this issue. I see it on both stalder and
> sompek. As far as I know they're both T2000's. I never see
> those issues on the other buildds which I think are all
> UltraSparc IIIi.
>
> Stadler moved to a 3.2 kernel in the mean time, and I'm still
> seeing the same problem.
>
> The problem gets worse depending on the number of cores I use, and
> I'm basicly using only 1 of the 32 cores because otherwise I just
> have too many problems.
>
>
> Kurt
>
> On Mon, Apr 02, 2012 at 07:39:57PM +0200, Kurt Roeckx wrote:
>> Hi,
>>
>> I've had a few packages fail to build on the new buildd stalder.
>> 2 example logs are icewaesel:
>> https://buildd.debian.org/status/logs.php?pkg=iceweasel&arch=sparc&ver=10.0.3esr-3
>>
>> It includes messages like: "The bug is not reproducible, so it is
>> likely a hardware or OS problem."
>>
>> It has always build fine on the other buildds, so I'm assuming
>> that as gcc tells me, it's a hardware or OS problem.
>>
>> Those builds where first with -j32, then with -j8, that is
>> setting DEB_BUILD_OPTIONS to parallel=32.
>>
>> All other sparc buildds run with parallel=2 and I've never seen
>> them having a problem. All others also only have 2 CPUs.
>>
>> stadler has this in /proc/cpuinfo:
>> cpu : UltraSparc T1 (Niagara)
>> fpu : UltraSparc T1 integrated FPU
>> pmu : niagara
>> prom : OBP 4.30.3 2009/06/08 13:29
>> type : sun4v
>> ncpus probed : 32
>> ncpus active : 32
>>
>> So my question is, is this some known problem, like for instance a
>> kernel problem that randomly does something wrong that's triggered
>> by having many proceses using much cpu time?
>>
>> Or maybe a hardware problem, for instance I'm heating things up
>> too much
>> using so many processes?
>>
>> Or something else?
>>
>>
>> PS: Please CC me, I'm not on the list.
>>
>> Kurt
>>
>>
>>
>
>
--
Bernd Zeimetz Debian GNU/Linux Developer
http://bzed.de http://www.debian.org
GPG Fingerprint: ECA1 E3F2 8E11 2432 D485 DD95 EB36 171A 6FF9 435F
Reply to: