Skip to main content
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [ptp-user] Failed to parse output of OpenMPI command 
Matthew,
After spending way too long on this, I've come to the conclusion that it is not possible to reliably parse the output from Open MPI 1.3.3 and below. I've implemented a work around, but unfortunately it means that the output you get from the program is wrapped in XML tags, but at least you shouldn't be seeing the parser errors any more. The good news is that Ralph has fixed the problems in Open MPI 1.4, and it looks like this will be back ported to 1.3.4.
I found another problem with RSE interpreting XML escape sequences, which may have also been causing problems when using DStore connections (but not ssh connections). I've implemented a work around for this, but it's really an RSE bug that needs to be fixed.
I've checked in the changes if you'd like to give it a go. Let me know if you still have problems. 

Regards,

Greg

On Aug 13, 2009, at 12:23 PM, Matthew Fotzler wrote:


Hey Greg,

I setup NAMD to see if I would have the same problem.  I'm still
getting the same error, but I also get a little output in the PTP
Runtime perspective.  Here's the debug output in context:

<stdout rank=
RTS job #50331652: "0">Info: 1 NAMD  2.7b1  Linux-x86_64  1
4pcnuggets  rdt-test&#010;</stdout>
<stdout rank="0">Inf
RTS job #50331652: o: Running on 1 processors.&#010;</stdout>
<stdout rank="0">Info: Charm++/Converse parallel runtime
RTS job #50331652: startup completed at 0.00456786 s&#010;</st
dout>
<stdout rank="0">Info: 1.49565 MB of memory in use
RTS job #50331652:  based on CmiMemoryUsage&#010;</stdout>
<stdout rank="0">Info: Changed directory to src&#010;</stdou
RTS job #50331652: t>
<stdout rank="0">Info: Configuration file is alanin&#010;</stdout>
<stdout rank="0">TCL: Suspendi
RTS job #50331652: ng until startup complete.&#010;</stdout>

The last line of output I see in the PTP Runtime view is "Info:
Running on 1 processors."
I updated my PTP from CVS this morning and I'm still using OpenMPI 1.4a1r21763. 

Matthew
On Fri, Aug 7, 2009 at 3:09 PM, Matthew Fotzler<incongruous@xxxxxxxxx> wrote: 
Greg,

I'm using RSE for the connection as well as the resource manager.  Is
there anything I can do to help find the problem?

Matthew
On Fri, Aug 7, 2009 at 2:56 PM, Greg Watson<g.watson@xxxxxxxxxxxx> wrote: 
Hi Matthew,
I don't think the output you're seeing is the problem. I haven't been able 
to reproduce it yet either.

Are you using Remote Tools or RSE for the connection?

Thanks,

Greg

On Aug 6, 2009, at 10:46 AM, Matthew Fotzler wrote:


Hey Greg,
Sure. The output is quite large, so I hope sending as an attachment is okay. debug.log is the debug output eclipse gives me when I try to run the job through PTP, and ompi.log is the output I get when I run 
from a terminal.

Thanks,
Matthew
On Wed, Aug 5, 2009 at 1:32 PM, Greg Watson<g.watson@xxxxxxxxxxxx> wrote: 

Hi Matthew,
If there's a newline between "</std" and "out>" then it is likely to be 
the
problem as XML names cannot contain whitespace. Can you send the actual 
output from POP at this point?

Thanks,

Greg

On Aug 5, 2009, at 12:48 PM, Matthew Fotzler wrote:


Hi,

I'm using PTP 3.0 from CVS and OpenMPI 1.4a1r21763 (yesterday's
nightly build). I'm trying to run the Parallel Ocean Program remotely 
and am receiving the following:
org.eclipse.core.runtime.CoreException: Failed to parse output of Open 
MPI command. Check output for errors.
      at
org .eclipse .ptp .rm .mpi .openmpi .core.OpenMPIPlugin.coreErrorException(OpenMPIPlugin.java:86) 
      at
org .eclipse .ptp .rm .mpi .openmpi .core .rtsystem .OpenMPIRuntimeSystemJob .doWaitExecution(OpenMPIRuntimeSystemJob.java:575) 
      at
org .eclipse .ptp .rm .core .rtsystem .AbstractToolRuntimeSystemJob .run(AbstractToolRuntimeSystemJob.java:258) at org.eclipse.core.internal.jobs.Worker.run(Worker.java: 55) 
Caused by: java.io.IOException: The element type "stdout" must be
terminated by the matching end-tag "</stdout>".
      at
org .eclipse .ptp .rm .mpi .openmpi .core .rtsystem .OpenMPIProcessMapXml13Parser .parse(OpenMPIProcessMapXml13Parser.java:518) 
      at
org .eclipse .ptp.rm.mpi.openmpi.core.rtsystem.OpenMPIRuntimeSystemJob $1.run(OpenMPIRuntimeSystemJob.java:232) 

Looking at the debug output, I see some lines that look like:

RTS job #100663327: <stdout
rank = "0 "> ------------------------------------------------------------------------&#010 ;</std 
RTS job #100663327:

RTS job #100663327: out>
I'm not sure if these are the culprit or not. Any ideas what might be 
causing this?

Thanks,
Matthew Fotzler
_______________________________________________
ptp-user mailing list
ptp-user@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/ptp-user


_______________________________________________
ptp-user mailing list
ptp-user@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/ptp-user
< debug.log><ompi.log>_______________________________________________ 
ptp-user mailing list
ptp-user@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/ptp-user


_______________________________________________
ptp-user mailing list
ptp-user@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/ptp-user




_______________________________________________
ptp-user mailing list
ptp-user@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/ptp-user

Back to the top