Hello there! I'm trying to work in a fault-tolerance project, and I immediately got interested in the ULFM extension of OpenMPI. Although I've managed to compile and install OpenMPI-1.8.6 without a problem, I have been failing so far to compile ULFM.
I am trying to follow the instructions in http://fault-tolerance.org/ulfm/ulfm-setup/ but it seems that it is not enough. I'm using m4-1.4.12 autoconf-2.69 automake-1.14.1 libtool-2.4.2 gcc-4.9.3 and I'm trying the mercurial version of ULFM. I am configuring with the exact command provided in the above link. When I try to make, I get the following error: make[2]: Leaving directory `/hydra/u/lrafa/local/ulfm/build/orte/mca/plm/slurm' Making all in mca/plm/poe make[2]: Entering directory `/hydra/u/lrafa/local/ulfm/build/orte/mca/plm/poe' CC plm_poe_component.lo CC plm_poe.lo ../../../../../orte/mca/plm/poe/plm_poe.c: In function 'spawn': ../../../../../orte/mca/plm/poe/plm_poe.c:135:34: error: 'orte_rmaps_t' has no member named 'get_job_map' if (NULL == (map = orte_rmaps.get_job_map(jdata->jobid))) { ^ ../../../../../orte/mca/plm/poe/plm_poe.c:222:42: error: 'orte_job_map_t' has no member named 'policy' if (ORTE_MAPPING_BYNODE & map->policy) { ^ make[2]: *** [plm_poe.lo] Error 1 make[2]: Leaving directory `/hydra/u/lrafa/local/ulfm/build/orte/mca/plm/poe' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/hydra/u/lrafa/local/ulfm/build/orte' make: *** [all-recursive] Error 1 Has anybody experienced that? Thanks in advance! Slainte Rafael Lago
