/*
* This function attempts to find an HNP to connect to.
*/
static int find_hnp(void) {
int ret, exit_status = ORTE_SUCCESS;
opal_list_t hnp_list;
opal_list_item_t *item;
orte_hnp_contact_t *hnpcandidate;
/* get the list of local hnp's available to us and setup
* contact info for them into the RML
*/
OBJ_CONSTRUCT(&hnp_list, opal_list_t);
if (ORTE_SUCCESS != (ret = orte_list_local_hnps(&hnp_list, true) ) ) {
orte_show_help("help-orte-checkpoint.txt", "no_hnps", true,
orte_checkpoint_globals.pid,
orte_process_info.tmpdir_base,
orte_process_info.top_session_dir,
ret, ORTE_ERROR_NAME(ret));
exit_status = ret;
goto cleanup;
}
/* search the list for the desired hnp */
while (NULL != (item = opal_list_remove_first(&hnp_list))) {
hnpcandidate = (orte_hnp_contact_t*)item;
if (hnpcandidate->name.jobid == orte_checkpoint_globals.req_hnp ||
hnpcandidate->pid == orte_checkpoint_globals.pid) {
/* this is the one we want */
orterun_hnp = hnpcandidate;
exit_status = ORTE_SUCCESS;
goto cleanup;
}
}
/* If no match was found, error out */
orte_show_help("help-orte-checkpoint.txt", "no_universe", true,
orte_checkpoint_globals.pid,
orte_process_info.tmpdir_base,
orte_process_info.top_session_dir);
cleanup:
while (NULL != (item = opal_list_remove_first(&hnp_list))) {
OBJ_RELEASE(item);
}
OBJ_DESTRUCT(&hnp_list);
if( NULL == orterun_hnp ) {
return ORTE_ERROR;
} else {
return exit_status;
}
}
/*
* This function attempts to find an HNP to connect to.
*/
static int find_hnp(void) {
int ret, exit_status = ORTE_SUCCESS;
opal_list_t hnp_list;
opal_list_item_t *item;
orte_hnp_contact_t *hnpcandidate;
/* get the list of local hnp's available to us and setup
* contact info for them into the RML
*/
OBJ_CONSTRUCT(&hnp_list, opal_list_t);
if (ORTE_SUCCESS != (ret = orte_list_local_hnps(&hnp_list, true) ) ) {
ORTE_ERROR_LOG(ret);
exit_status = ret;
goto cleanup;
}
/* search the list for the desired hnp */
while (NULL != (item = opal_list_remove_first(&hnp_list))) {
hnpcandidate = (orte_hnp_contact_t*)item;
if( hnpcandidate->pid == orte_migrate_globals.pid) {
/* this is the one we want */
orterun_hnp = hnpcandidate;
exit_status = ORTE_SUCCESS;
goto cleanup;
}
}
cleanup:
while (NULL != (item = opal_list_remove_first(&hnp_list))) {
OBJ_RELEASE(item);
}
OBJ_DESTRUCT(&hnp_list);
if( NULL == orterun_hnp ) {
return ORTE_ERROR;
} else {
return exit_status;
}
}
int
main(int argc, char *argv[])
{
int ret, exit_status = ORTE_SUCCESS;
opal_list_t hnp_list;
opal_list_item_t* item = NULL;
orte_ps_mpirun_info_t hnpinfo;
bool reported = false;
/***************
* Initialize
***************/
OBJ_CONSTRUCT(&hnp_list, opal_list_t);
if (ORTE_SUCCESS != (ret = orte_ps_init(argc, argv))) {
exit_status = ret;
goto cleanup;
}
/*
* Get the directory listing
*/
opal_output_verbose(10, orte_ps_globals.output,
"orte_ps: Acquiring list of HNPs and setting contact info into RML...\n");
if (ORTE_SUCCESS != (ret = orte_list_local_hnps(&hnp_list, true) ) ) {
exit_status = ret;
goto cleanup;
}
opal_output_verbose(10, orte_ps_globals.output,
"orte_ps: Found %d HNPs\n",
(int)opal_list_get_size(&hnp_list));
/*
* For each hnp in the listing
*/
while (NULL != (item = opal_list_remove_first(&hnp_list))) {
orte_hnp_contact_t *hnp = (orte_hnp_contact_t*)item;
hnpinfo.hnp = hnp;
opal_output_verbose(10, orte_ps_globals.output,
"orte_ps: Processing HNP %lu\n",
(unsigned long)hnpinfo.hnp->pid);
if (0 < orte_ps_globals.pid &&
hnpinfo.hnp->pid != orte_ps_globals.pid) {
continue;
}
/*
* Gather the information
*/
opal_output_verbose(10, orte_ps_globals.output,
"orte_ps: Gathering Information for HNP: %s:%d\n",
ORTE_NAME_PRINT(&(hnpinfo.hnp->name)),
hnpinfo.hnp->pid);
if( ORTE_SUCCESS != (ret = gather_information(&hnpinfo)) ) {
/* this could be due to a stale session directory - if so,
* just skip this entry, but don't abort
*/
if (!reported && ORTE_ERR_SILENT == ret) {
orte_show_help("help-orte-ps.txt", "stale-hnp", true,
ORTE_NAME_PRINT(&(hnpinfo.hnp->name)));
reported = true;
continue;
}
goto cleanup;
}
/* Print the information */
if (orte_ps_globals.parseable) {
if (ORTE_SUCCESS != (ret = parseable_print(&hnpinfo))) {
exit_status = ret;
goto cleanup;
}
} else {
if(ORTE_SUCCESS != (ret = pretty_print(&hnpinfo)) ) {
exit_status = ret;
goto cleanup;
}
}
}
/***************
* Cleanup
***************/
cleanup:
orte_finalize();
return exit_status;
}
