Skip to content

Several fixes related to session directories: #2042

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 4 commits into from
Sep 7, 2016
Merged

Several fixes related to session directories: #2042

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
81195ab
Several fixes related to session directories:
artpol84 Sep 1, 2016
dc0ab67
Add PMIx key to provide RM with ability to indicate that it will cleanup
artpol84 Sep 1, 2016
74a11d7
Fix session dir cleanup code.
artpol84 Sep 1, 2016
a9a7f39
ess/pmi: fix the comments about MCA/PMIx setting conflict resolution.
artpol84 Sep 7, 2016
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions opal/mca/pmix/pmix_types.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -75,6 +75,7 @@ BEGIN_C_DECLS
#define OPAL_PMIX_TMPDIR "pmix.tmpdir" // (char*) top-level tmp dir assigned to session
#define OPAL_PMIX_NSDIR "pmix.nsdir" // (char*) sub-tmpdir assigned to namespace
#define OPAL_PMIX_PROCDIR "pmix.pdir" // (char*) sub-nsdir assigned to proc
#define OPAL_PMIX_TDIR_RMCLEAN "pmix.tdir.rmclean" // (bool) Resource Manager will clean session directories

/* information about relative ranks as assigned by the RM */
#define OPAL_PMIX_JOBID "pmix.jobid" // (uint32_t) jobid assigned by scheduler
Expand Down
28 changes: 1 addition & 27 deletions orte/mca/ess/base/ess_base_std_app.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -136,10 +136,7 @@ int orte_ess_base_app_setup(bool db_restrict_local)
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,
ORTE_PROC_MY_NAME))) {
if (ORTE_SUCCESS != (ret = orte_session_dir(true, ORTE_PROC_MY_NAME))) {
ORTE_ERROR_LOG(ret);
error = "orte_session_dir";
goto error;
Expand All @@ -149,29 +146,6 @@ int orte_ess_base_app_setup(bool db_restrict_local)
proc-specific session directory. */
opal_output_set_output_file_info(orte_process_info.proc_session_dir,
"output-", NULL, NULL);
/* store the session directory location */
OBJ_CONSTRUCT(&kv, opal_value_t);
kv.key = strdup(OPAL_PMIX_NSDIR);
kv.type = OPAL_STRING;
kv.data.string = strdup(orte_process_info.job_session_dir);
if (OPAL_SUCCESS != (ret = opal_pmix.store_local(ORTE_PROC_MY_NAME, &kv))) {
ORTE_ERROR_LOG(ret);
OBJ_DESTRUCT(&kv);
error = "opal pmix put job sessiondir";
goto error;
}
OBJ_DESTRUCT(&kv);
OBJ_CONSTRUCT(&kv, opal_value_t);
kv.key = strdup(OPAL_PMIX_PROCDIR);
kv.type = OPAL_STRING;
kv.data.string = strdup(orte_process_info.proc_session_dir);
if (OPAL_SUCCESS != (ret = opal_pmix.store_local(ORTE_PROC_MY_NAME, &kv))) {
ORTE_ERROR_LOG(ret);
OBJ_DESTRUCT(&kv);
error = "opal pmix put proc sessiondir";
goto error;
}
OBJ_DESTRUCT(&kv);
}
/* Setup the communication infrastructure */
/*
Expand Down
17 changes: 4 additions & 13 deletions orte/mca/ess/base/ess_base_std_orted.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -237,10 +237,7 @@ int orte_ess_base_orted_setup(char **hosts)
/* take a pass thru the session directory code to fillin the
* tmpdir names - don't create anything yet
*/
if (ORTE_SUCCESS != (ret = orte_session_dir(false,
orte_process_info.tmpdir_base,
orte_process_info.nodename,
ORTE_PROC_MY_NAME))) {
if (ORTE_SUCCESS != (ret = orte_session_dir(false, ORTE_PROC_MY_NAME))) {
ORTE_ERROR_LOG(ret);
error = "orte_session_dir define";
goto error;
Expand All @@ -250,10 +247,7 @@ int orte_ess_base_orted_setup(char **hosts)
*/
orte_session_dir_cleanup(ORTE_JOBID_WILDCARD);
/* now actually create the directory tree */
if (ORTE_SUCCESS != (ret = orte_session_dir(true,
orte_process_info.tmpdir_base,
orte_process_info.nodename,
ORTE_PROC_MY_NAME))) {
if (ORTE_SUCCESS != (ret = orte_session_dir(true, ORTE_PROC_MY_NAME))) {
ORTE_ERROR_LOG(ret);
error = "orte_session_dir";
goto error;
Expand All @@ -277,11 +271,8 @@ int orte_ess_base_orted_setup(char **hosts)
/* define a log file name in the session directory */
snprintf(log_file, PATH_MAX, "output-orted-%s-%s.log",
jobidstring, orte_process_info.nodename);
log_path = opal_os_path(false,
orte_process_info.tmpdir_base,
orte_process_info.top_session_dir,
log_file,
NULL);
log_path = opal_os_path(false, orte_process_info.top_session_dir,
log_file, NULL);

fd = open(log_path, O_RDWR|O_CREAT|O_TRUNC, 0640);
if (fd < 0) {
Expand Down
7 changes: 3 additions & 4 deletions orte/mca/ess/base/ess_base_std_tool.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -145,10 +145,9 @@ int orte_ess_base_tool_setup(void)
* tmp base where any other session directories on
* this node might be located
*/
if (ORTE_SUCCESS != (ret = orte_session_dir_get_name(NULL,
&orte_process_info.tmpdir_base,
&orte_process_info.top_session_dir,
orte_process_info.nodename, NULL))) {

ret = orte_session_setup_base(NULL);
if (ORTE_SUCCESS != ret ) {
ORTE_ERROR_LOG(ret);
error = "define session dir names";
goto error;
Expand Down
36 changes: 15 additions & 21 deletions orte/mca/ess/hnp/ess_hnp_module.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -138,7 +138,7 @@ static int rte_init(void)
{
int ret;
char *error = NULL;
char *contact_path, *jobfam_dir;
char *contact_path;
orte_job_t *jdata;
orte_node_t *node;
orte_proc_t *proc;
Expand Down Expand Up @@ -294,10 +294,7 @@ static int rte_init(void)
/* take a pass thru the session directory code to fillin the
* tmpdir names - don't create anything yet
*/
if (ORTE_SUCCESS != (ret = orte_session_dir(false,
orte_process_info.tmpdir_base,
orte_process_info.nodename,
ORTE_PROC_MY_NAME))) {
if (ORTE_SUCCESS != (ret = orte_session_dir(false, ORTE_PROC_MY_NAME))) {
error = "orte_session_dir define";
goto error;
}
Expand All @@ -307,10 +304,7 @@ static int rte_init(void)
orte_session_dir_cleanup(ORTE_JOBID_WILDCARD);

/* now actually create the directory tree */
if (ORTE_SUCCESS != (ret = orte_session_dir(true,
orte_process_info.tmpdir_base,
orte_process_info.nodename,
ORTE_PROC_MY_NAME))) {
if (ORTE_SUCCESS != (ret = orte_session_dir(true, ORTE_PROC_MY_NAME))) {
error = "orte_session_dir";
goto error;
}
Expand Down Expand Up @@ -586,9 +580,12 @@ static int rte_init(void)
opal_output_set_output_file_info(orte_process_info.proc_session_dir,
"output-", NULL, NULL);
/* save my contact info in a file for others to find */
jobfam_dir = opal_dirname(orte_process_info.job_session_dir);
contact_path = opal_os_path(false, jobfam_dir, "contact.txt", NULL);
free(jobfam_dir);
if( NULL == orte_process_info.jobfam_session_dir ){
/* has to be set here! */
ORTE_ERROR_LOG(ORTE_ERR_BAD_PARAM);
goto error;
}
contact_path = opal_os_path(false, orte_process_info.jobfam_session_dir, "contact.txt", NULL);
OPAL_OUTPUT_VERBOSE((2, orte_debug_output,
"%s writing contact file %s",
ORTE_NAME_PRINT(ORTE_PROC_MY_NAME),
Expand Down Expand Up @@ -758,10 +755,9 @@ static int rte_init(void)
true, error, ORTE_ERROR_NAME(ret), ret);
}
/* remove my contact info file, if we have session directories */
if (NULL != orte_process_info.job_session_dir) {
jobfam_dir = opal_dirname(orte_process_info.job_session_dir);
contact_path = opal_os_path(false, jobfam_dir, "contact.txt", NULL);
free(jobfam_dir);
if (NULL != orte_process_info.jobfam_session_dir) {
contact_path = opal_os_path(false, orte_process_info.jobfam_session_dir,
"contact.txt", NULL);
unlink(contact_path);
free(contact_path);
}
Expand All @@ -775,7 +771,6 @@ static int rte_init(void)
static int rte_finalize(void)
{
char *contact_path;
char *jobfam_dir;

if (signals_set) {
/* Remove the epipe handler */
Expand Down Expand Up @@ -816,10 +811,9 @@ static int rte_finalize(void)
(void) mca_base_framework_close(&opal_pstat_base_framework);

/* remove my contact info file, if we have session directories */
if (NULL != orte_process_info.job_session_dir) {
jobfam_dir = opal_dirname(orte_process_info.job_session_dir);
contact_path = opal_os_path(false, jobfam_dir, "contact.txt", NULL);
free(jobfam_dir);
if (NULL != orte_process_info.jobfam_session_dir) {
contact_path = opal_os_path(false, orte_process_info.jobfam_session_dir,
"contact.txt", NULL);
unlink(contact_path);
free(contact_path);
}
Expand Down
58 changes: 58 additions & 0 deletions orte/mca/ess/pmi/ess_pmi_module.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -94,6 +94,7 @@ static int rte_init(void)
uint16_t u16, *u16ptr;
char **peers=NULL, *mycpuset, **cpusets=NULL;
opal_process_name_t wildcard_rank, pname;
bool bool_val, *bool_ptr = &bool_val, tdir_mca_override = false;
size_t i;

/* run the prolog */
Expand Down Expand Up @@ -242,6 +243,63 @@ static int rte_init(void)
free(string_key);
}

/* retrieve temp directories info */
OPAL_MODEX_RECV_VALUE_OPTIONAL(ret, OPAL_PMIX_TMPDIR, &wildcard_rank, &val, OPAL_STRING);
if (OPAL_SUCCESS == ret && NULL != val) {
/* We want to provide user with ability
* to override RM settings at his own risk
*/
if( NULL == orte_process_info.top_session_dir ){
orte_process_info.top_session_dir = val;
} else {
/* keep the MCA setting */
tdir_mca_override = true;
free(val);
}
val = NULL;
}

if( !tdir_mca_override ){
OPAL_MODEX_RECV_VALUE_OPTIONAL(ret, OPAL_PMIX_NSDIR, &wildcard_rank, &val, OPAL_STRING);
if (OPAL_SUCCESS == ret && NULL != val) {
/* We want to provide user with ability
* to override RM settings at his own risk
*/
if( NULL == orte_process_info.job_session_dir ){
orte_process_info.job_session_dir = val;
} else {
/* keep the MCA setting */
free(val);
tdir_mca_override = true;
}
val = NULL;
}
}

if( !tdir_mca_override ){
OPAL_MODEX_RECV_VALUE_OPTIONAL(ret, OPAL_PMIX_PROCDIR, &wildcard_rank, &val, OPAL_STRING);
if (OPAL_SUCCESS == ret && NULL != val) {
/* We want to provide user with ability
* to override RM settings at his own risk
*/
if( NULL == orte_process_info.proc_session_dir ){
orte_process_info.proc_session_dir = val;
} else {
/* keep the MCA setting */
tdir_mca_override = true;
free(val);
}
val = NULL;
}
}

if( !tdir_mca_override ){
OPAL_MODEX_RECV_VALUE_OPTIONAL(ret, OPAL_PMIX_TDIR_RMCLEAN, &wildcard_rank, &bool_ptr, OPAL_BOOL);
if (OPAL_SUCCESS == ret ) {
orte_process_info.rm_session_dirs = bool_val;
}
}

/* retrieve our topology */
val = NULL;
OPAL_MODEX_RECV_VALUE_OPTIONAL(ret, OPAL_PMIX_LOCAL_TOPO,
Expand Down
Loading