From e3e3acb8f31c83dd4852bed9a02803ca18d84939 Mon Sep 17 00:00:00 2001 From: Ralph Castain Date: Mon, 5 Feb 2018 19:31:43 -0800 Subject: [PATCH] Ensure we fail if remote nodes cannot find executable Signed-off-by: Ralph Castain (cherry picked from commit ce901ba24746e569e0935f513677d34c7f053d38) --- orte/mca/errmgr/default_orted/errmgr_default_orted.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/orte/mca/errmgr/default_orted/errmgr_default_orted.c b/orte/mca/errmgr/default_orted/errmgr_default_orted.c index 7d131559a91..6a66e35d127 100644 --- a/orte/mca/errmgr/default_orted/errmgr_default_orted.c +++ b/orte/mca/errmgr/default_orted/errmgr_default_orted.c @@ -586,6 +586,14 @@ static void proc_errors(int fd, short args, void *cbdata) * so that we send a consolidated error report * back to the HNP */ + if (jdata->num_local_procs == jdata->num_terminated) { + /* let the state machine know */ + if (ORTE_PROC_STATE_FAILED_TO_START == state) { + ORTE_ACTIVATE_JOB_STATE(jdata, ORTE_JOB_STATE_FAILED_TO_START); + } else { + ORTE_ACTIVATE_JOB_STATE(jdata, ORTE_JOB_STATE_FAILED_TO_LAUNCH); + } + } goto cleanup; }