The problem was OpenMPI compiled with PMI, since it was refering to old
SLURM libraries.
Thanks,
Marco

Il giorno ven 16 set 2016 alle ore 17:58 Carlos Fenoy <mini...@gmail.com>
ha scritto:

> Marco,
>
> It seems that your application is reading the slurm.conf, if it is an
> ompss application maybe you need to recompile the application or the
> runtime with your modifications to slurm in order for it to understand the
> new parameter.
>
> Regards,
> Carlos
>
> On Fri, Sep 16, 2016 at 2:05 PM, Marco D'Amico <marco.damic...@gmail.com>
> wrote:
>
>> Hi Sergio,
>> yes the slurm.h and slurm.h.in are ok.
>>
>> Il giorno ven 16 set 2016 alle ore 08:42 Sergio Iserte <sise...@uji.es>
>> ha scritto:
>>
>>> Marco,
>>> could you check that files src/slurm/slurm.h.in and src/slurm/slurm.h have
>>> the parameter?
>>>
>>> I'm not sure, I'm just speculating..., but the second file is
>>> autogenerated when configuring.
>>>
>>> Sergio.
>>>
>>>
>>>
>>> 2016-09-15 17:59 GMT+02:00 Marco D'Amico <marco.damic...@gmail.com>:
>>>
>>>> Same problem also as a normal slurm.conf parameter, and also changing
>>>> slurm version (from 15.08.10 to 16.05.4)
>>>>
>>>> Marco
>>>>
>>>>
>>>> Il giorno gio 15 set 2016 alle ore 12:45 Marco D'Amico <
>>>> marco.damic...@gmail.com> ha scritto:
>>>>
>>>>> Hi,
>>>>> I am trying to add a new parameter in slurm.conf in the partition part
>>>>> of the script, so I edited all structures and functions in order to work
>>>>> with it.
>>>>> When slurm daemons read the slurm.conf I have no parsing problems and
>>>>> it looks everything is going fine, but then when the application is
>>>>> launched I have the following error:
>>>>>
>>>>> mpi_ompss_pils: error: Parsing error at unrecognized key: SharingFactor
>>>>> mpi_ompss_pils: error: Parse error in file /......../slurm.conf line
>>>>> 207: " SharingFactor=0.3"
>>>>> srun: error: slurm_receive_msg: Zero Bytes were transmitted or received
>>>>> srun: error: slurm_receive_msg[10.2.1.32]: Zero Bytes were transmitted
>>>>> or received
>>>>> srun: error: slurm_receive_msg: Zero Bytes were transmitted or received
>>>>> srun: error: slurm_receive_msg[10.2.1.32]: Zero Bytes were transmitted
>>>>> or received
>>>>> srun: error: s01r1b32: tasks 0-1: Segmentation fault
>>>>>
>>>>> Why my application is reading slurm.conf again? And why now the
>>>>> parameter is not recognized?
>>>>> Any help is appreciated,
>>>>> Thanks
>>>>>
>>>>
>>>
>>>
>>> --
>>> Sergio Iserte
>>> High Performance Computing & Architectures (HPCA)
>>> Department of Computer Science and Engineering (DICC)
>>> Universitat Jaume I (UJI)
>>>
>>>
>
>
> --
> --
> Carles Fenoy
>

Reply via email to