APSStrictParseArguments

[Xavier Nuel]

Hey there,

I am experiencing some trouble while testing the mogaRing examples. Specifically, at step 3 of the GuideToUse

geneticOptimizer -input optim1.sdds -reduce 1

reports in a file called "o" (I imagine that's a product of changing -j for -eo, see below):

invalid command name "APSStrictParseArguments"
while executing
"APSStrictParseArguments {valueList tagList rootname }"
(file "./runJob1" line 1)

As a consequence, all 40 jobs crash immediately after entering the queue.

I have modified the geneticOptimizer.local file to comply with torque queueing system that runs on our cluster. The modifications were minimal:

catch {exec cat $tmpFile | qsub -l walltime=10000:00:00 -q mono -l nodes=1:ppn=1 -V -o [pwd] -eo -N [file root [file tail $input]] } result

that is, we specified the queue where the job should be submitted, the number of nodes, the walltime and change the -j option to -eo to have combined error/output.

I also changed the type of shell I am using from csh to bash, so that the line

puts $fd "source ~/.bashrc"

The behavior of the script is the following: Upon launching geneticOptimizer -input optim1.sdds -reduce 1

Input file read and checked.
0
Submitting script: runJob1 for optim1-000000
Script submission returns: 31269.isei-oscar1
1
Submitting script: runJob1 for optim1-000001
Script submission returns: 31270.isei-oscar1
...
39
Submitting script: runJob1 for optim1-000039
Script submission returns: 31308.isei-oscar1

After that it keeps sending the following message indefinitely, even after all the queued jobs are dead, until I kill the process:

Updating job counts
jobs: current=40, done=0, proc'd=0, toProc=0, running=40
Wed May 15 13:44:46 CEST 2013: 40 jobs running

When I see the .o file, the result is this error:
[gavaldax@isei mogaRing]$more o
invalid command name "APSStrictParseArguments"
while executing
"APSStrictParseArguments {valueList tagList rootname }"
(file "./runJob1" line 1)
isei098.hpc: Connection refused

What does it means connection refused?

Have you seen this before? I suspect a tclTk problem but I am not sure. Thanks in advance for your help,

Xavier and Alejandro

[Xavier Nuel]

Hi,

Finally we resolved the problem: We saw that some of the oag package were installed in /usr/lib and /usr/bin. Then, we reinstalled it in the path /usr/local/ of our cluster. The second step was export the PATH in ~/.bashrc :

export PATH=$PATH:/usr/local/applications/elegant:/usr/local/applications/elegant/sdds:/usr/local/oag/apps/lib/linux-x86_64:/usr/local/oag/apps/src/tcltkapp/oagapp:/usr/local/oag/apps/bin:/usr/bin:

Thanks

Xavier and Alejandro