Add example of child dep termination

Add a simple example/test of terminating child
jobs upon termination of the parent, allowing
user to toggle between that behavior and allowing
the child to continue.

Signed-off-by: Ralph Castain <[email protected]>

.gitignore

@@ -135,6 +135,7 @@ examples/colocate
 examples/pset
 examples/nodeid
 examples/client-threaded
+examples/dynamic-dep
 
 src/sys/powerpc/atomic-32.s
 src/sys/powerpc/atomic-64.s

examples/Makefile

@@ -62,7 +62,8 @@ EXAMPLES = \
 	colocate \
 	pset \
 	nodeid \
-	client-threaded
+	client-threaded \
+	dynamic-dep
 
 all: $(EXAMPLES)
 

examples/Makefile.include

@@ -70,4 +70,5 @@ EXTRA_DIST += \
         examples/colocate.c \
         examples/pset.c \
         examples/nodeid.c \
-        examples/client-threaded.c
+        examples/client-threaded.c \
+        examples/dynamic-dep.c

examples/dynamic-dep.c

@@ -0,0 +1,100 @@
+/*
+ * Copyright (c) 2004-2010 The Trustees of Indiana University and Indiana
+ *                         University Research and Technology
+ *                         Corporation.  All rights reserved.
+ * Copyright (c) 2004-2011 The University of Tennessee and The University
+ *                         of Tennessee Research Foundation.  All rights
+ *                         reserved.
+ * Copyright (c) 2004-2005 High Performance Computing Center Stuttgart,
+ *                         University of Stuttgart.  All rights reserved.
+ * Copyright (c) 2004-2005 The Regents of the University of California.
+ *                         All rights reserved.
+ * Copyright (c) 2006-2013 Los Alamos National Security, LLC.
+ *                         All rights reserved.
+ * Copyright (c) 2009-2012 Cisco Systems, Inc.  All rights reserved.
+ * Copyright (c) 2011      Oak Ridge National Labs.  All rights reserved.
+ * Copyright (c) 2013-2019 Intel, Inc.  All rights reserved.
+ * Copyright (c) 2015      Mellanox Technologies, Inc.  All rights reserved.
+ * Copyright (c) 2016      Research Organization for Information Science
+ *                         and Technology (RIST). All rights reserved.
+ * Copyright (c) 2021-2024 Nanook Consulting  All rights reserved.
+ * $COPYRIGHT$
+ *
+ * Additional copyrights may follow
+ *
+ * $HEADER$
+ *
+ */
+
+#include <stdbool.h>
+
+#define _GNU_SOURCE
+#include <stdio.h>
+#include <stdlib.h>
+#include <sys/param.h>
+#include <time.h>
+#include <unistd.h>
+
+#include "examples.h"
+#include <pmix.h>
+
+static pmix_proc_t myproc;
+
+int main(int argc, char **argv)
+{
+    int rc, exitcode;
+    char nsp2[PMIX_MAX_NSLEN + 1];
+    pmix_proc_t proc;
+    pmix_app_t *app;
+    bool test_dep = true;
+    pmix_info_t info;
+
+    if (1 < argc) {
+        if (0 == strcmp(argv[1], "dep")) {
+            test_dep = false;
+        }
+    }
+
+    /* init us */
+    if (PMIX_SUCCESS != (rc = PMIx_Init(&myproc, NULL, 0))) {
+        fprintf(stderr, "Client ns %s rank %d: PMIx_Init failed: %d\n", myproc.nspace, myproc.rank,
+                rc);
+        exit(rc);
+    }
+
+    /* rank=0 calls spawn */
+    if (0 == myproc.rank) {
+        PMIX_APP_CREATE(app, 1);
+        app->cmd = strdup("sleep");
+        app->maxprocs = 2;
+        PMIX_ARGV_APPEND(rc, app->argv, app->cmd);
+        PMIX_ARGV_APPEND(rc, app->argv, "5");
+
+        PMIX_INFO_LOAD(&info, PMIX_SPAWN_CHILD_SEP, &test_dep, PMIX_BOOL);
+        rc = PMIx_Spawn(&info, 1, app, 1, nsp2);
+        if (PMIX_SUCCESS != rc) {
+            fprintf(stderr, "Client ns %s rank %d: PMIx_Spawn failed: %s(%d)\n", myproc.nspace,
+                    myproc.rank, PMIx_Error_string(rc), rc);
+            exitcode = rc;
+            /* terminate our peers */
+            PMIX_LOAD_PROCID(&proc, myproc.nspace, PMIX_RANK_WILDCARD);
+            PMIx_Abort(rc, "FAILED TO START CHILD JOB", &proc, 1);
+            goto done;
+        } else {
+          fprintf(stderr, "Spawn success.\n");
+        }
+        PMIX_APP_FREE(app, 1);
+
+        // now terminate
+    }
+
+
+done:
+    if (PMIX_SUCCESS != (rc = PMIx_Finalize(NULL, 0))) {
+        fprintf(stderr, "Client ns %s rank %d:PMIx_Finalize failed: %d\n", myproc.nspace,
+                myproc.rank, rc);
+    }
+
+    fflush(stderr);
+    return (0);
+}

src/docs/show-help-files/help-state-base.txt

@@ -48,7 +48,7 @@ its parent:
   Parent: %s
   Child:  %s
 
-This behavior is controlled by setting the PMIX_JOB_CHILD_SEP attribute
+This behavior is controlled by setting the PMIX_SPAWN_CHILD_SEP attribute
 in the job info provided at time of spawn for the child job. When set to
 "true", the runtime will "separate" the child from its parent and allow
 it to continue execution after parent termination. Note that this is only

src/mca/state/dvm/state_dvm.c

@@ -835,6 +835,7 @@ static void check_complete(int fd, short args, void *cbdata)
         PMIX_LIST_FOREACH(jptr, &jdata->children, prte_job_t)
         {
             if (prte_get_attribute(&jptr->attributes, PRTE_JOB_CHILD_SEP, (void**)&sepptr, PMIX_BOOL) && !sep) {
+                pmix_output(0, "TERMINATING CHILD");
                 proc = PMIX_NEW(prte_proc_t);
                 PMIX_LOAD_PROCID(&proc->name, jptr->nspace, PMIX_RANK_WILDCARD);
                 pmix_pointer_array_add(&procs, proc);
@@ -853,6 +854,8 @@ static void check_complete(int fd, short args, void *cbdata)
                     PMIX_RELEASE(proc);
                 }
             }
+        } else {
+            pmix_output(0, "NOT TERMINATING CHILD");
         }
         PMIX_DESTRUCT(&procs);
     }