diff --git a/src/plugins/checkpoint/aix/checkpoint_aix.c b/src/plugins/checkpoint/aix/checkpoint_aix.c index 2f0c7944a35ffde47eca37cfe1c50c62b0bc6589..0e7cc856946f1ed2d507ffd3ab89ff8a11dde05d 100644 --- a/src/plugins/checkpoint/aix/checkpoint_aix.c +++ b/src/plugins/checkpoint/aix/checkpoint_aix.c @@ -256,7 +256,9 @@ extern int slurm_ckpt_comp ( struct step_record * step_ptr, time_t event_time, check_ptr->error_msg = xstrdup(error_msg); } - if (++check_ptr->reply_cnt == check_ptr->node_cnt) { + /* We need a reply from each compute node, + * plus POE itself */ + if (check_ptr->reply_cnt++ == check_ptr->node_cnt) { info("Checkpoint complete for job %u.%u", step_ptr->job_ptr->job_id, step_ptr->step_id); check_ptr->time_stamp = time(NULL);