Re: [OMPI devel] Error after ompi-restart

2008-11-10 Thread Josh Hursey
Thanks for the patch. I tested and applied it in r19961 of the Open MPI Trunk. Sorry it took me so long to do so. Thanks again, Josh On Nov 4, 2008, at 11:10 AM, Leonardo Fialho wrote: Josh, It works fine for me. I think that it is the error. Leonardo Josh Hursey escribió: Leonardo, Sor

Re: [OMPI devel] Error after ompi-restart

2008-11-04 Thread Leonardo Fialho
Josh, It works fine for me. I think that it is the error. Leonardo Josh Hursey escribió: Leonardo, Sorry I have been really slow in replying, I have been pretty swamped lately. What version of the trunk are you using? I've been seeing C/R failures starting around r19872, but I haven't had

Re: [OMPI devel] Error after ompi-restart

2008-11-04 Thread Josh Hursey
Leonardo, Sorry I have been really slow in replying, I have been pretty swamped lately. What version of the trunk are you using? I've been seeing C/R failures starting around r19872, but I haven't had time to focus on trying to find out what is going wrong. You may be right in your asse

Re: [OMPI devel] Error after ompi-restart

2008-11-04 Thread Leonardo Fialho
I´m not sure, but I think that line 659 on file orte/mca/ess/env/ess_env_module.c should contain if (ORTE_SUCCESS != (ret = orte_ess_base_build_nidmap(orte_process_info.sync_buf, &nidmap, *jmap*))) { But actually it contains if (ORTE_SUCCESS != (ret = orte_ess_base_build_nidmap(orte_process

[OMPI devel] Error after ompi-restart

2008-11-03 Thread Leonardo Fialho
Hi All, I think that exists an error in the trunk version while trying to restore a checkpoint. The function orte_util_decode_pidmap while attempts to execute the following code /* store the data */ for (i=0; i < num_procs; i++) { pmap.node = nodes[i]; pmap.local_rank =