int orte_ess_base_app_setup(bool db_restrict_local)
{
int ret;
char *error = NULL;
opal_value_t kv;
/*
* stdout/stderr buffering
* If the user requested to override the default setting then do
* as they wish.
*/
if( orte_ess_base_std_buffering > -1 ) {
if( 0 == orte_ess_base_std_buffering ) {
setvbuf(stdout, NULL, _IONBF, 0);
setvbuf(stderr, NULL, _IONBF, 0);
}
else if( 1 == orte_ess_base_std_buffering ) {
setvbuf(stdout, NULL, _IOLBF, 0);
setvbuf(stderr, NULL, _IOLBF, 0);
}
else if( 2 == orte_ess_base_std_buffering ) {
setvbuf(stdout, NULL, _IOFBF, 0);
setvbuf(stderr, NULL, _IOFBF, 0);
}
}
/* if I am an MPI app, we will let the MPI layer define and
* control the opal_proc_t structure. Otherwise, we need to
* do so here */
if (ORTE_PROC_NON_MPI) {
orte_process_info.super.proc_name = *(opal_process_name_t*)ORTE_PROC_MY_NAME;
orte_process_info.super.proc_hostname = strdup(orte_process_info.nodename);
orte_process_info.super.proc_flags = OPAL_PROC_ALL_LOCAL;
orte_process_info.super.proc_arch = opal_local_arch;
opal_proc_local_set(&orte_process_info.super);
}
/* get a separate orte event base */
orte_event_base = opal_start_progress_thread("orte", true);
progress_thread_running = true;
/* open and setup the state machine */
if (ORTE_SUCCESS != (ret = mca_base_framework_open(&orte_state_base_framework, 0))) {
ORTE_ERROR_LOG(ret);
error = "orte_state_base_open";
goto error;
}
if (ORTE_SUCCESS != (ret = orte_state_base_select())) {
ORTE_ERROR_LOG(ret);
error = "orte_state_base_select";
goto error;
}
/* open the errmgr */
if (ORTE_SUCCESS != (ret = mca_base_framework_open(&orte_errmgr_base_framework, 0))) {
ORTE_ERROR_LOG(ret);
error = "orte_errmgr_base_open";
goto error;
}
/* setup my session directory */
if (orte_create_session_dirs) {
OPAL_OUTPUT_VERBOSE((2, orte_ess_base_framework.framework_output,
"%s setting up session dir with\n\ttmpdir: %s\n\thost %s",
ORTE_NAME_PRINT(ORTE_PROC_MY_NAME),
(NULL == orte_process_info.tmpdir_base) ? "UNDEF" : orte_process_info.tmpdir_base,
orte_process_info.nodename));
if (ORTE_SUCCESS != (ret = orte_session_dir(true,
orte_process_info.tmpdir_base,
orte_process_info.nodename, NULL,
ORTE_PROC_MY_NAME))) {
ORTE_ERROR_LOG(ret);
error = "orte_session_dir";
goto error;
}
/* Once the session directory location has been established, set
the opal_output env file location to be in the
proc-specific session directory. */
opal_output_set_output_file_info(orte_process_info.proc_session_dir,
"output-", NULL, NULL);
/* store the session directory location in the database */
OBJ_CONSTRUCT(&kv, opal_value_t);
kv.key = strdup(OPAL_DSTORE_JOB_SDIR);
kv.type = OPAL_STRING;
kv.data.string = strdup(orte_process_info.job_session_dir);
if (OPAL_SUCCESS != (ret = opal_dstore.store(opal_dstore_internal,
ORTE_PROC_MY_NAME,
&kv))) {
ORTE_ERROR_LOG(ret);
OBJ_DESTRUCT(&kv);
error = "opal dstore store";
goto error;
}
OBJ_DESTRUCT(&kv);
OBJ_CONSTRUCT(&kv, opal_value_t);
kv.key = strdup(OPAL_DSTORE_MY_SDIR);
kv.type = OPAL_STRING;
kv.data.string = strdup(orte_process_info.proc_session_dir);
if (OPAL_SUCCESS != (ret = opal_dstore.store(opal_dstore_internal,
ORTE_PROC_MY_NAME,
&kv))) {
ORTE_ERROR_LOG(ret);
OBJ_DESTRUCT(&kv);
error = "opal dstore store";
goto error;
}
OBJ_DESTRUCT(&kv);
}
/* Setup the communication infrastructure */
/*
* OOB Layer
*/
if (ORTE_SUCCESS != (ret = mca_base_framework_open(&orte_oob_base_framework, 0))) {
ORTE_ERROR_LOG(ret);
error = "orte_oob_base_open";
goto error;
}
if (ORTE_SUCCESS != (ret = orte_oob_base_select())) {
ORTE_ERROR_LOG(ret);
error = "orte_oob_base_select";
goto error;
}
/* Runtime Messaging Layer */
if (ORTE_SUCCESS != (ret = mca_base_framework_open(&orte_rml_base_framework, 0))) {
ORTE_ERROR_LOG(ret);
error = "orte_rml_base_open";
goto error;
}
if (ORTE_SUCCESS != (ret = orte_rml_base_select())) {
ORTE_ERROR_LOG(ret);
error = "orte_rml_base_select";
goto error;
}
/* setup the errmgr */
if (ORTE_SUCCESS != (ret = orte_errmgr_base_select())) {
ORTE_ERROR_LOG(ret);
error = "orte_errmgr_base_select";
goto error;
}
/* Routed system */
if (ORTE_SUCCESS != (ret = mca_base_framework_open(&orte_routed_base_framework, 0))) {
ORTE_ERROR_LOG(ret);
error = "orte_routed_base_open";
goto error;
}
if (ORTE_SUCCESS != (ret = orte_routed_base_select())) {
ORTE_ERROR_LOG(ret);
error = "orte_routed_base_select";
goto error;
}
/*
* Group communications
*/
if (ORTE_SUCCESS != (ret = mca_base_framework_open(&orte_grpcomm_base_framework, 0))) {
ORTE_ERROR_LOG(ret);
error = "orte_grpcomm_base_open";
goto error;
}
if (ORTE_SUCCESS != (ret = orte_grpcomm_base_select())) {
ORTE_ERROR_LOG(ret);
error = "orte_grpcomm_base_select";
goto error;
}
/* non-daemon/HNP apps can only have the default proxy PLM
* module open - provide a chance for it to initialize
*/
if (ORTE_SUCCESS != (ret = orte_plm.init())) {
ORTE_ERROR_LOG(ret);
error = "orte_plm_init";
goto error;
}
/* enable communication via the rml */
if (ORTE_SUCCESS != (ret = orte_rml.enable_comm())) {
ORTE_ERROR_LOG(ret);
error = "orte_rml.enable_comm";
goto error;
}
/* setup the routed info */
if (ORTE_SUCCESS != (ret = orte_routed.init_routes(ORTE_PROC_MY_NAME->jobid, NULL))) {
ORTE_ERROR_LOG(ret);
error = "orte_routed.init_routes";
goto error;
}
#if OPAL_ENABLE_FT_CR == 1
/*
* Setup the SnapC
*/
if (ORTE_SUCCESS != (ret = mca_base_framework_open(&orte_snapc_base_framework, 0))) {
ORTE_ERROR_LOG(ret);
error = "orte_snapc_base_open";
goto error;
}
if (ORTE_SUCCESS != (ret = mca_base_framework_open(&orte_sstore_base_framework, 0))) {
ORTE_ERROR_LOG(ret);
error = "orte_sstore_base_open";
goto error;
}
if (ORTE_SUCCESS != (ret = orte_snapc_base_select(ORTE_PROC_IS_HNP, ORTE_PROC_IS_APP))) {
ORTE_ERROR_LOG(ret);
error = "orte_snapc_base_select";
goto error;
}
if (ORTE_SUCCESS != (ret = orte_sstore_base_select())) {
ORTE_ERROR_LOG(ret);
error = "orte_sstore_base_select";
goto error;
}
/* apps need the OPAL CR stuff */
opal_cr_set_enabled(true);
#else
opal_cr_set_enabled(false);
#endif
/* Initalize the CR setup
* Note: Always do this, even in non-FT builds.
* If we don't some user level tools may hang.
*/
if (ORTE_SUCCESS != (ret = orte_cr_init())) {
ORTE_ERROR_LOG(ret);
error = "orte_cr_init";
goto error;
}
/* open the distributed file system */
if (ORTE_SUCCESS != (ret = mca_base_framework_open(&orte_dfs_base_framework, 0))) {
ORTE_ERROR_LOG(ret);
error = "orte_dfs_base_open";
goto error;
}
if (ORTE_SUCCESS != (ret = orte_dfs_base_select())) {
ORTE_ERROR_LOG(ret);
error = "orte_dfs_base_select";
goto error;
}
return ORTE_SUCCESS;
error:
if (!progress_thread_running) {
/* can't send the help message, so ensure it
* comes out locally
*/
orte_show_help_finalize();
}
orte_show_help("help-orte-runtime.txt",
"orte_init:startup:internal-failure",
true, error, ORTE_ERROR_NAME(ret), ret);
return ret;
}
static int native_init(void)
{
char **uri, *srv;
++init_cntr;
if (1 < init_cntr) {
return OPAL_SUCCESS;
}
opal_output_verbose(2, opal_pmix_base_framework.framework_output,
"%s pmix:native init called",
OPAL_NAME_PRINT(OPAL_PROC_MY_NAME));
/* if we don't have a path to the daemon rendezvous point,
* then we need to return an error UNLESS we have been directed
* to allow init prior to having an identified server. This is
* needed for singletons as they will start without a server
* to support them, but may have one assigned at a later time */
if (NULL == mca_pmix_native_component.uri) {
opal_output_verbose(2, opal_pmix_base_framework.framework_output,
"%s pmix:native NULL uri",
OPAL_NAME_PRINT(OPAL_PROC_MY_NAME));
if (NULL != (srv = getenv("PMIX_SERVER_URI"))) {
mca_pmix_native_component.uri = strdup(srv);
mca_pmix_native_component.id = OPAL_PROC_MY_NAME;
} else if (opal_pmix_base_allow_delayed_server) {
/* not ready yet, so decrement our init_cntr so we can come thru
* here again */
--init_cntr;
/* let the caller know that the server isn't available yet */
return OPAL_ERR_SERVER_NOT_AVAIL;
} else {
/* not ready yet, so decrement our init_cntr so we can come thru
* here again */
--init_cntr;
return OPAL_ERROR;
}
}
/* if we have it, setup the path to the daemon rendezvous point */
if (NULL != mca_pmix_native_component.uri) {
opal_output_verbose(2, opal_pmix_base_framework.framework_output,
"%s pmix:native constructing component fields with server %s",
OPAL_NAME_PRINT(OPAL_PROC_MY_NAME),
mca_pmix_native_component.uri);
memset(&mca_pmix_native_component.address, 0, sizeof(struct sockaddr_un));
mca_pmix_native_component.address.sun_family = AF_UNIX;
uri = opal_argv_split(mca_pmix_native_component.uri, ':');
if (2 != opal_argv_count(uri)) {
opal_argv_free(uri);
return OPAL_ERROR;
}
/* if the rendezvous file doesn't exist, that's an error */
if (0 != access(uri[1], R_OK)) {
opal_output_verbose(2, opal_pmix_base_framework.framework_output,
"%s pmix:native rendezvous file %s does not exist",
OPAL_NAME_PRINT(OPAL_PROC_MY_NAME), uri[1]);
opal_argv_free(uri);
return OPAL_ERR_NOT_FOUND;
}
opal_convert_string_to_process_name(&mca_pmix_native_component.server, uri[0]);
snprintf(mca_pmix_native_component.address.sun_path,
sizeof(mca_pmix_native_component.address.sun_path)-1,
"%s", uri[1]);
opal_argv_free(uri);
/* create an event base and progress thread for us */
if (NULL == (mca_pmix_native_component.evbase = opal_start_progress_thread("opal_async", true))) {
return OPAL_ERROR;
}
}
/* we will connect on first send */
return OPAL_SUCCESS;
}
