Ray crash: how to debug?

[krobison]

I'm trying to play with the Ray assembler and my first run with 50x2 paired end Illumina data (kmer=21; running on a single core of a large machine) ended ignobly (end of run output below). Any suggestions on how to troubleshoot this?

***
Step: Coverage distribution analysis
Date: Thu Apr 26 22:33:39 2012
Elapsed time: 6 seconds
Since beginning: 1 hours, 23 minutes, 6 seconds
***



Rank 0 is computing vertices & edges [1/25886148]
[actinode.cluster:1932] *** An error occurred in MPI_Iprobe
[actinode.cluster:1932] *** on communicator MPI_COMM_WORLD
[actinode.cluster:1932] *** MPI_ERR_RANK: invalid rank
[actinode.cluster:1932] *** MPI_ERRORS_ARE_FATAL (your MPI job will now abort)
--------------------------------------------------------------------------
mpiexec has exited due to process rank 0 with PID 1932 on
node actinode.cluster exiting without calling "finalize". This may
have caused other processes in the application to be
terminated by signals sent by mpiexec (as reported here).

[severin]

input command?

Please provide the input command.

I suspect you either tried to execute mpirun with a single core or mpi is not installed.

[krobison]

Sorry about that -- obviously I should have posted the command line. Yes, I set the number of processes to 1 (worried about things getting too big), which I suspect was the problem -- I've had a successful run now with the number set much higher.

[seb567]

Hello,

Ray should run fine with one core.

I just fixed a bug that occured sometimes when running Ray with less than 4 cores.

The bug was in RayPlatform.

Sorry about that -- obviously I should have posted the command line. Yes, I set the number of processes to 1 (worried about things getting too big), which I suspect was the problem -- I've had a successful run now with the number set much higher.