[mvapich-discuss] No route to host in MVAPICH-Virt

Maksym Planeta mplaneta at os.inf.tu-dresden.de
Wed Jul 5 05:00:04 EDT 2017



On 07/04/2017 07:31 PM, Xiaoyi Lu wrote:
> Hi,
>
> In addition to Jie's comments, please try to stop your firewall and
> rerun the commands.

This actually helped. I ran systemctl stop firewalld. And now it seems 
to work. Thank you.

>
> Thanks,
> Xiaoyi
>
> Sent from my iPhone
>
> On Jul 4, 2017, at 12:05 PM, Jie Zhang <zhanjie at cse.ohio-state.edu
> <mailto:zhanjie at cse.ohio-state.edu>> wrote:
>
>> Hi,  Maksym,
>>
>> This should be a network setup issue.
>> Can you please try a few things:
>>
>> 1. Do you have correct IP/hostname mapping for both hosts in the file,
>> such as /etc/hosts
>> 2. What is the output if you run
>> /opt/mvapich2/virt/2.2/bin/mpirun_rsh  -np 2 host11 host12
>> MV2_VIRT_USE_IVSHMEM=1 hostname
>>
>>
>> On Tue, Jul 4, 2017 at 11:01 AM Maksym Planeta
>> <mplaneta at os.inf.tu-dresden.de <mailto:mplaneta at os.inf.tu-dresden.de>>
>> wrote:
>>
>>     Hello,
>>
>>     I have setup several virtual machines and I'm trying to run simple
>>     MPI programs in there.
>>
>>     My current test is following:
>>
>>     /opt/mvapich2/virt/2.2/bin/mpirun_rsh  -export-all -n 2 host11
>>     host12 MV2_VIRT_USE_IVSHMEM=1 bash -c 'netstat -l | grep
>>     $(printenv MPISPAWN_CHECKIN_PORT)'
>>     tcp        0      0 0.0.0.0:49379 <http://0.0.0.0:49379>
>>      0.0.0.0:*               LISTEN
>>     connect [mt_checkin]: No route to host
>>     [host11:mpirun_rsh][child_handler] Error in init phase, aborting!
>>     (1/2 mpispawn connections)
>>
>>     So I'm trying to look if the checkin port is open. But the remote
>>     host connection fails with a message "no route to host".
>>
>>     Showing the command which is about to be executed gives me this:
>>
>>     $ /opt/mvapich2/virt/2.2/bin/mpirun_rsh -show  -export-all -n 2
>>     host11 host12 MV2_VIRT_USE_IVSHMEM=1 bash -c 'netstat -l | grep
>>     $(printenv MPISPAWN_CHECKIN_PORT)'
>>     /usr/bin/ssh -q host12 cd /home/user; /usr/bin/env
>>     MPISPAWN_MPIRUN_MPD=0 USE_LINEAR_SSH=1 MPISPAWN_MPIRUN_HOST=host11
>>     MPISPAWN_MPIRUN_HOSTIP=141.76.84.15 MPIRUN_RSH_LAUNCH=1
>>     MPISPAWN_CHECKIN_PORT=52056 MPISPAWN_MPIRUN_PORT=52056
>>     MPISPAWN_NNODES=2 MPISPAWN_GLOBAL_NPROCS=2 MPISPAWN_MPIRUN_ID=5190
>>     MPISPAWN_ARGC=3 MPDMAN_KVS_TEMPLATE=kvs_506_host11_5190
>>     MPISPAWN_LOCAL_NPROCS=1 MPISPAWN_ARGV_0='bash'
>>     MPISPAWN_ARGV_1='-c' MPISPAWN_ARGV_2='netstat -l | grep $(printenv
>>     MPISPAWN_CHECKIN_PORT)' MPISPAWN_ARGC=3
>>     MPISPAWN_GENERIC_ENV_COUNT=1
>>     MPISPAWN_GENERIC_NAME_0=MV2_VIRT_USE_IVSHMEM
>>     MPISPAWN_GENERIC_VALUE_0=1 MPISPAWN_ID=1
>>     MPISPAWN_WORKING_DIR=/home/user MPISPAWN_MPIRUN_RANK_0=1
>>     /opt/mvapich2/virt/2.2/bin/mpispawn 0
>>
>>     /usr/bin/bash -c cd /home/user; /usr/bin/env
>>     MPISPAWN_MPIRUN_MPD=0 USE_LINEAR_SSH=1 MPISPAWN_MPIRUN_HOST=host11
>>     MPISPAWN_MPIRUN_HOSTIP=141.76.84.15 MPIRUN_RSH_LAUNCH=1
>>     MPISPAWN_CHECKIN_PORT=52056 MPISPAWN_MPIRUN_PORT=52056
>>     MPISPAWN_NNODES=2 MPISPAWN_GLOBAL_NPROCS=2 MPISPAWN_MPIRUN_ID=5190
>>     MPISPAWN_ARGC=3 MPDMAN_KVS_TEMPLATE=kvs_506_host11_5190
>>     MPISPAWN_LOCAL_NPROCS=1 MPISPAWN_ARGV_0='bash'
>>     MPISPAWN_ARGV_1='-c' MPISPAWN_ARGV_2='netstat -l | grep $(printenv
>>     MPISPAWN_CHECKIN_PORT)' MPISPAWN_ARGC=3
>>     MPISPAWN_GENERIC_ENV_COUNT=1
>>     MPISPAWN_GENERIC_NAME_0=MV2_VIRT_USE_IVSHMEM
>>     MPISPAWN_GENERIC_VALUE_0=1 MPISPAWN_ID=0
>>     MPISPAWN_WORKING_DIR=/home/user MPISPAWN_MPIRUN_RANK_0=0
>>     /opt/mvapich2/virt/2.2/bin/mpispawn 0
>>
>>
>>     I can reach both hosts from each other and they see the same IP
>>     addresses, and I can ssh to each of the two hosts from each of the
>>     two hosts.
>>
>>
>>     Could you help me to identify the issue?
>>
>>     The MVAPICH-Virt version is the last accessible on the mvapich
>>     download page
>>
>>     P. S.
>>
>>     You say to allow reading for the user in section 5 of the Virt
>>     user manual, but actually the library asks for write access as
>>     well, so I set up rights to 766.
>>
>>     --
>>     Regards,
>>     Maksym Planeta
>>
>>     _______________________________________________
>>     mvapich-discuss mailing list
>>     mvapich-discuss at cse.ohio-state.edu
>>     <mailto:mvapich-discuss at cse.ohio-state.edu>
>>     http://mailman.cse.ohio-state.edu/mailman/listinfo/mvapich-discuss
>>
>> --
>> Thanks,
>> -Jie
>>
>> --------------------------------------------
>> Department of Computer Science & Engineering
>> Erwin E. Dreese Laboratories, 2015 Neil Avenue
>> Ohio State University
>> _______________________________________________
>> mvapich-discuss mailing list
>> mvapich-discuss at cse.ohio-state.edu
>> <mailto:mvapich-discuss at cse.ohio-state.edu>
>> http://mailman.cse.ohio-state.edu/mailman/listinfo/mvapich-discuss

-- 
Regards,
Maksym Planeta

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5174 bytes
Desc: S/MIME Cryptographic Signature
URL: <http://mailman.cse.ohio-state.edu/pipermail/mvapich-discuss/attachments/20170705/87dd076c/attachment-0001.p7s>


More information about the mvapich-discuss mailing list