[mpich-discuss] HYDRA and kill process

Torquil Macdonald Sørensen torquil at gmail.com
Fri Mar 25 03:37:48 CDT 2011


Hi!

Unfortunately, the error still exists in nightly snapshot r8281.

Just tried r8281 from here:

http://www.mcs.anl.gov/research/projects/mpich2/downloads/tarballs/nightly/trunk/

I started a job consisting of 8 processes, 4 on hostA and 4 on hostB using 
"mpiexec -n 8 progfile". Hitting CTRL-c on hostA kills its four processes, but 
the remaining four still run on hostB, so I am forced to log in there and kill 
them myself.

Best regards
Torquil Sørensen

On 24/03/11 22:25, Pavan Balaji wrote:
>
> This should work correctly. The problem that was fixed earlier was with
> multiple process groups (i.e., using Comm_spawn).
>
> Can you try the nightly snapshot in any case?
>
> -- Pavan
>
> On 03/24/2011 02:16 PM, Jain, Rohit wrote:
>> I have seen same problem and reported earlier. It is not just ctrl-c,
>> even if your process on one host crashes, it leaves other parallel
>> host processes.
>>
>> I think Pavan has already fixed it in newer version. Right, Pavan?
>>
>> Regards,
>> Rohit
>>
>>
>> -----Original Message-----
>> From: mpich-discuss-bounces at mcs.anl.gov
>> [mailto:mpich-discuss-bounces at mcs.anl.gov] On Behalf Of Torquil
>> Macdonald Sørensen
>> Sent: Thursday, March 24, 2011 12:10 PM
>> To: mpich-discuss at mcs.anl.gov
>> Subject: [mpich-discuss] HYDRA and kill process
>>
>> Hi!
>>
>> With mpich-1.3.2, say I run a parallel job consisting of two processes
>> running a
>> hostA and hostB. They are started by running mpiexec on hostA.
>>
>> If I (on hostA) hit CTRL-c to kill my job, the process on hostB keeps
>> running...
>>
>> With earlier MPICH2 versions, without HYDRA, the behaviour was
>> different. The
>> whole job (all processes on all hosts) were killed when I hist CTRL-c
>> on the
>> host where mpiexec was run.
>>
>> Any pointers on how to "restore" the old behaviour when using the
>> newest MPICH2
>> and HYDRA?
>>
>> Or is it a bug?
>>
>> Best regards
>> Torquil Sørensen
>> _______________________________________________
>> mpich-discuss mailing list
>> mpich-discuss at mcs.anl.gov
>> https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
>> _______________________________________________
>> mpich-discuss mailing list
>> mpich-discuss at mcs.anl.gov
>> https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
>



More information about the mpich-discuss mailing list