[mpich-discuss] 答复: MPICH2 Between Linux/Windows

[Jayesh Krishna]

Hi,
 
# This definitely looks like a firewall/network_connectivity issue. Are you sure that you can ping from
   192.168.1.1 --> 192.168.1.2
   192.168.1.2 --> 192.168.1.1

  (Make sure that you can ping each machine from the other)

# Are you sure that the firewalls are off on both linux and Windows?
# Do you get the same error with sock channel ("-channel sock" on Windows, "--with-device=ch3:sock" on Unix)?

Regards,
Jayesh

----- Original Message -----
From: "YOSOQOO" <yosoqoo at 163.com>
To: "Jayesh Krishna" <jayesh at mcs.anl.gov>
Sent: Friday, April 13, 2012 8:58:28 PM
Subject: 答复: [mpich-discuss] MPICH2 Between Linux/Windows

Hi
I'm trying to establish a connection between a process on linux (redhat  2.6.32-71.el6.i686) and a process on a WinXp machine(32-bit system). MPICH2 1.4.1p1 on both machines. The WinXP with installer, Linux from sources. 
I configured the linux side with 
./configure --with-pm=smpd --with-device=ch3:nemesis 
 
I have tested the windows between 32-bit system is OK. but mpich2 between Linux/windows is error. The contents of the machinefile.txt  is:
 192.168.1.1
 192.168.1.2
 
The run results:
C:\mpi>mpiexec -n 4 -machinefile machinefile.txt -channel nemesis -path "c:\mpi;/home/administrator/mpich2-1.4.1p1/examp
les" cpi
Process 2 of 4 is on yangsq
Process 0 of 4 is on yangsq
Process 3 of 4 is on yangsq-redhat32
Process 1 of 4 is on yangsq-redhat32
Fatal error in PMPI_Bcast: Other MPI error, error stack:
PMPI_Bcast(1478)......................: MPI_Bcast(buf=0022FF28, count=1, MPI_INT, root=0, MPI_COMM_WORLD) failed
MPIR_Bcast_impl(1321).................:
MPIR_Bcast_intra(1119)................:
MPIR_Bcast_scatter_ring_allgather(962):
MPIR_Bcast_binomial(213)..............: Failure during collective
MPIR_Bcast_scatter_ring_allgather(955):
MPIR_Bcast_binomial(189)..............:
MPIC_Send(66).........................:
MPIC_Wait(540)........................:
MPIDI_CH3I_Progress(402)..............:
MPID_nem_mpich2_blocking_recv(905)....:
MPID_nem_newtcp_module_poll(37).......:
MPID_nem_newtcp_module_connpoll(2656).:
gen_cnting_fail_handler(1739).........: connect failed - 远程系统拒绝网络连接。
 (errno 1225)
job aborted:
rank: node: exit code[: error message]
0: 192.168.1.1: 1: process 0 exited without calling finalize
1: 192.168.1.2: -2
2: 192.168.1.1: 123
3: 192.168.1.2: -2
 
 
thanks.


-----邮件原件-----
发件人: Jayesh Krishna [mailto:jayesh at mcs.anl.gov] 
发送时间: 2012年4月13日 22:30
收件人: mpich-discuss at mcs.anl.gov
抄送: yosoqoo at 163.com
主题: Re: [mpich-discuss] MPICH2 Between Linux/Windows

Hi,
 Did you try running your job with nemesis and the firewalls turned off?

Regards,
Jayesh

----- Original Message -----
From: "杨尚琴" <yosoqoo at 163.com>
To: mpich-discuss at mcs.anl.gov
Sent: Thursday, April 12, 2012 4:21:55 AM
Subject: [mpich-discuss] MPICH2 Between Linux/Windows




Hi, 
I have make ture turn off firewalls on both the machines (Windows firewall & any firewalls on Linux) . 

And then I try another build with : 
./configure --with-pm=smpd --with-device=ch3:sock 

then I run it ,the results follows, it can print the hostname,but the program dosen't end normally. 

C:\mpi>mpiexec -n 2 -machinefile machinefile.txt -channel sock -path "c:\mpi;/home/administrator/mpich2-1.4.1p1/examples 
" cpi 
Process 0 of 2 is on yangsq 
Process 1 of 2 is on yangsq-redhat32 

thanks. 




_______________________________________________
mpich-discuss mailing list     mpich-discuss at mcs.anl.gov
To manage subscription options or unsubscribe:
https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss