From 40cd97d6e8d9e184742a596aa3f3c7baf31a1f9d Mon Sep 17 00:00:00 2001
From: Moe Jette <jette1@llnl.gov>
Date: Wed, 25 Apr 2007 22:12:16 +0000
Subject: [PATCH] Major re-write of jobcomp/script plugin: fix memory leak and
general code clean-up - Mark Grondona.
---
NEWS | 2 +
src/plugins/jobcomp/script/Makefile.am | 2 +-
src/plugins/jobcomp/script/Makefile.in | 5 +-
src/plugins/jobcomp/script/job_record.c | 102 ----
src/plugins/jobcomp/script/job_record.h | 70 ---
src/plugins/jobcomp/script/jobcomp_script.c | 526 ++++++++++++--------
6 files changed, 323 insertions(+), 384 deletions(-)
delete mode 100644 src/plugins/jobcomp/script/job_record.c
delete mode 100644 src/plugins/jobcomp/script/job_record.h
diff --git a/NEWS b/NEWS
index a9f4cb2645b..3546b3bffcc 100644
--- a/NEWS
+++ b/NEWS
@@ -6,6 +6,8 @@ documents those changes that are of interest to users and admins.
-- BLUEGENE - code to make it so you can make a 36x36x36 system.
The wiring isn't correct on the X axis for anything over 8 though.
But that is only in emulation mode.
+ -- Major re-write of jobcomp/script plugin: fix memory leak and
+ general code clean-up.
* Changes in SLURM 1.2.6
========================
diff --git a/src/plugins/jobcomp/script/Makefile.am b/src/plugins/jobcomp/script/Makefile.am
index 9ad7aa90470..eae162b656f 100644
--- a/src/plugins/jobcomp/script/Makefile.am
+++ b/src/plugins/jobcomp/script/Makefile.am
@@ -9,5 +9,5 @@ INCLUDES = -I$(top_srcdir) -I$(top_srcdir)/src/common
pkglib_LTLIBRARIES = jobcomp_script.la
# Text file job completion logging plugin.
-jobcomp_script_la_SOURCES = job_record.c job_record.h jobcomp_script.c
+jobcomp_script_la_SOURCES = jobcomp_script.c
jobcomp_script_la_LDFLAGS = $(SO_LDFLAGS) $(PLUGIN_FLAGS)
diff --git a/src/plugins/jobcomp/script/Makefile.in b/src/plugins/jobcomp/script/Makefile.in
index 871dba84763..0a9a4cfbab6 100644
--- a/src/plugins/jobcomp/script/Makefile.in
+++ b/src/plugins/jobcomp/script/Makefile.in
@@ -78,7 +78,7 @@ am__installdirs = "$(DESTDIR)$(pkglibdir)"
pkglibLTLIBRARIES_INSTALL = $(INSTALL)
LTLIBRARIES = $(pkglib_LTLIBRARIES)
jobcomp_script_la_LIBADD =
-am_jobcomp_script_la_OBJECTS = job_record.lo jobcomp_script.lo
+am_jobcomp_script_la_OBJECTS = jobcomp_script.lo
jobcomp_script_la_OBJECTS = $(am_jobcomp_script_la_OBJECTS)
DEFAULT_INCLUDES = -I. -I$(srcdir) -I$(top_builddir) -I$(top_builddir)/slurm
depcomp = $(SHELL) $(top_srcdir)/auxdir/depcomp
@@ -285,7 +285,7 @@ INCLUDES = -I$(top_srcdir) -I$(top_srcdir)/src/common
pkglib_LTLIBRARIES = jobcomp_script.la
# Text file job completion logging plugin.
-jobcomp_script_la_SOURCES = job_record.c job_record.h jobcomp_script.c
+jobcomp_script_la_SOURCES = jobcomp_script.c
jobcomp_script_la_LDFLAGS = $(SO_LDFLAGS) $(PLUGIN_FLAGS)
all: all-am
@@ -356,7 +356,6 @@ mostlyclean-compile:
distclean-compile:
-rm -f *.tab.c
-@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/job_record.Plo@am__quote@
@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/jobcomp_script.Plo@am__quote@
.c.o:
diff --git a/src/plugins/jobcomp/script/job_record.c b/src/plugins/jobcomp/script/job_record.c
deleted file mode 100644
index 32d987c8f2f..00000000000
--- a/src/plugins/jobcomp/script/job_record.c
+++ /dev/null
@@ -1,102 +0,0 @@
-/******************************************************************************
- * job_record.c - Functions to alter script plugin job record structure.
- ******************************************************************************
- * Produced at Center for High Performance Computing, North Dakota State
- * University
- * Written by Nathan Huff <nhuff@acm.org>
- * UCRL-CODE-226842.
- *
- * This file is part of SLURM, a resource management program.
- * For details, see <http://www.llnl.gov/linux/slurm/>.
- *
- * SLURM is free software; you can redistribute it and/or modify it under
- * the terms of the GNU General Public License as published by the Free
- * Software Foundation; either version 2 of the License, or (at your option)
- * any later version.
- *
- * In addition, as a special exception, the copyright holders give permission
- * to link the code of portions of this program with the OpenSSL library under
- * certain conditions as described in each individual source file, and
- * distribute linked combinations including the two. You must obey the GNU
- * General Public License in all respects for all of the code used other than
- * OpenSSL. If you modify file(s) with this exception, you may extend this
- * exception to your version of the file(s), but you are not obligated to do
- * so. If you do not wish to do so, delete this exception statement from your
- * version. If you delete this exception statement from all source files in
- * the program, then also delete it here.
- *
- * SLURM is distributed in the hope that it will be useful, but WITHOUT ANY
- * WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
- * FOR A PARTICULAR PURPOSE. See the GNU General Public License for more
- * details.
- *
- * You should have received a copy of the GNU General Public License along
- * with SLURM; if not, write to the Free Software Foundation, Inc.,
- * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
-\*****************************************************************************/
-
-#ifdef HAVE_CONFIG_H
-# include "config.h"
-#endif
-
-#if HAVE_STDINT_H
-# include <stdint.h>
-#endif
-#if HAVE_INTTYPES_H
-# include <inttypes.h>
-#endif
-
-
-#include <unistd.h>
-#include <sys/types.h>
-#include <slurm/slurm.h>
-#include "job_record.h"
-#include "src/common/xmalloc.h"
-#include "src/common/xstring.h"
-
-/*
- * Create a new job_record containing the job completion information
- */
-job_record job_record_create(uint32_t job_id, uint32_t user_id, char *job_name,
- char *job_state, char *partition,
- uint32_t limit, time_t start, time_t end, time_t submit,
- uint16_t batch_flag, char *node_list, uint32_t num_procs,
- char *account)
-{
- job_record ret;
-
- ret = xmalloc(sizeof(struct job_record_));
- ret->job_id = job_id;
- ret->user_id = user_id;
- ret->job_name = xstrdup(job_name);
- ret->job_state = xstrdup(job_state);
- ret->partition = xstrdup(partition);
- ret->limit = limit;
- ret->start = start;
- ret->submit = submit;
- ret->batch_flag = batch_flag;
- ret->end = end;
- ret->node_list = xstrdup(node_list);
- ret->num_procs = num_procs;
- if (account)
- ret->account = xstrdup(account);
-
- return ret;
-}
-
-/*
- * Free the memory from a job_record structure
- */
-void job_record_destroy(void *job) {
- job_record j = job;
-
- if (j == NULL)
- return;
-
- xfree(j->job_name);
- xfree(j->partition);
- xfree(j->node_list);
- xfree(j->job_state);
- xfree(j->account);
- xfree(j);
-}
diff --git a/src/plugins/jobcomp/script/job_record.h b/src/plugins/jobcomp/script/job_record.h
deleted file mode 100644
index cfd762739e5..00000000000
--- a/src/plugins/jobcomp/script/job_record.h
+++ /dev/null
@@ -1,70 +0,0 @@
-/*****************************************************************************\
- * job_record.h - Header for job_record structure.
- *****************************************************************************
- * Produced at Center for High Performance Computing, North Dakota State
- * University
- * Written by Nathan Huff <nhuff@geekshanty.com>
- *
- * This file is part of SLURM, a resource management program.
- * For details, see <http://www.llnl.gov/linux/slurm/>.
- *
- * SLURM is free software; you can redistribute it and/or modify it under
- * the terms of the GNU General Public License as published by the Free
- * Software Foundation; either version 2 of the License, or (at your option)
- * any later version.
- *
- * In addition, as a special exception, the copyright holders give permission
- * to link the code of portions of this program with the OpenSSL library under
- * certain conditions as described in each individual source file, and
- * distribute linked combinations including the two. You must obey the GNU
- * General Public License in all respects for all of the code used other than
- * OpenSSL. If you modify file(s) with this exception, you may extend this
- * exception to your version of the file(s), but you are not obligated to do
- * so. If you do not wish to do so, delete this exception statement from your
- * version. If you delete this exception statement from all source files in
- * the program, then also delete it here.
- *
- * SLURM is distributed in the hope that it will be useful, but WITHOUT ANY
- * WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
- * FOR A PARTICULAR PURPOSE. See the GNU General Public License for more
- * details.
- *
- * You should have received a copy of the GNU General Public License along
- * with SLURM; if not, write to the Free Software Foundation, Inc.,
- * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
-\*****************************************************************************/
-
-
-#ifndef _JOB_RECORD_H
-#define _JOB_RECORD_H
-
-/* Structure to contain job completion information */
-struct job_record_ {
- uint32_t job_id;
- uint32_t user_id;
- char *job_name;
- char *job_state;
- char *partition;
- uint32_t limit;
- uint16_t batch_flag;
- uint32_t num_procs;
- time_t submit;
- time_t start;
- time_t end;
- char *node_list;
- char *account;
-};
-
-typedef struct job_record_ * job_record;
-
-/* Create a job record */
-job_record job_record_create(uint32_t job_id, uint32_t user_id, char *job_name,
- char *job_state, char *partition,
- uint32_t limit, time_t start, time_t end, time_t submit,
- uint16_t batch_flag, char *node_list, uint32_t num_procs,
- char *account);
-
-/* Destroy a job record */
-void job_record_destroy(void *j);
-
-#endif /* _JOB_RECORD_H */
diff --git a/src/plugins/jobcomp/script/jobcomp_script.c b/src/plugins/jobcomp/script/jobcomp_script.c
index 1afb8215da9..d2ec5363527 100644
--- a/src/plugins/jobcomp/script/jobcomp_script.c
+++ b/src/plugins/jobcomp/script/jobcomp_script.c
@@ -47,12 +47,13 @@
# include <inttypes.h>
#endif
-#include <stdio.h>
#include <stdlib.h>
#include <string.h>
+#include <stdarg.h>
#include <paths.h>
#include <sys/types.h>
#include <sys/stat.h>
+#include <fcntl.h>
#include <sys/wait.h>
#include <unistd.h>
#include <pthread.h>
@@ -65,7 +66,6 @@
#include "src/common/xstring.h"
#include "src/common/list.h"
#include "src/slurmctld/slurmctld.h"
-#include "job_record.h"
/*
* These variables are required by the generic plugin interface. If they
@@ -100,9 +100,7 @@ const char plugin_name[] = "Job completion logging script plugin";
const char plugin_type[] = "jobcomp/script";
const uint32_t plugin_version = 90;
-static int plugin_errno = SLURM_SUCCESS;
static char * script = NULL;
-static char error_str[256];
static List comp_list = NULL;
static pthread_t script_thread = 0;
@@ -111,222 +109,343 @@ static pthread_mutex_t comp_list_mutex = PTHREAD_MUTEX_INITIALIZER;
static pthread_cond_t comp_list_cond = PTHREAD_COND_INITIALIZER;
static int agent_exit = 0;
+/*
+ * Local plugin errno
+ */
+static int plugin_errno = SLURM_SUCCESS;
+
+static struct jobcomp_errno {
+ int n;
+ const char *descr;
+} errno_table [] = {
+ { 0, "No Error" },
+ { EACCES, "Script access denied" },
+ { EEXIST, "Script does not exist" },
+ { EINVAL, "JocCompLoc invalid" },
+ { -1, "Unknown Error" }
+};
+
+/*
+ * Return string representation of plugin errno
+ */
+static const char * jobcomp_script_strerror (int errnum)
+{
+ struct jobcomp_errno *ep = errno_table;
+
+ while ((ep->n != errnum) && (ep->n != -1))
+ ep++;
+
+ return (ep->descr);
+}
+
+/*
+ * Structure for holding job completion information for later
+ * use by script;
+ */
+struct jobcomp_info {
+ uint32_t jobid;
+ uint32_t uid;
+ uint32_t limit;
+ uint32_t nprocs;
+ uint16_t batch_flag;
+ time_t submit;
+ time_t start;
+ time_t end;
+ char *nodes;
+ char *name;
+ char *partition;
+ char *jobstate;
+ char *account;
+};
+
+struct jobcomp_info * jobcomp_info_create (struct job_record *job)
+{
+ enum job_states state;
+ struct jobcomp_info * j = xmalloc (sizeof (*j));
+
+ j->jobid = job->job_id;
+ j->uid = job->user_id;
+ j->name = xstrdup (job->name);
+
+ /*
+ * Job will typically be COMPLETING when this code is called.
+ * We remove the COMPLETING flag to hopefully get the evenual
+ * completion state: e.g.: JOB_FAILED, TIMEOUT, ....
+ */
+ state = job->job_state & (~JOB_COMPLETING);
+ j->jobstate = xstrdup (job_state_string (state));
+
+ j->partition = xstrdup (job->partition);
+ j->limit = job->time_limit;
+ j->start = job->start_time;
+ j->end = job->end_time;
+ j->submit = job->details ? job->details->submit_time:job->start_time;;
+ j->batch_flag = job->batch_flag;
+ j->nodes = xstrdup (job->nodes);
+ j->nprocs = job->num_procs;
+ j->account = job->account ? xstrdup (job->account) : NULL;
+
+ return (j);
+}
+
+void jobcomp_info_destroy (struct jobcomp_info *j)
+{
+ if (j == NULL)
+ return;
+ xfree (j->name);
+ xfree (j->partition);
+ xfree (j->nodes);
+ xfree (j->jobstate);
+ xfree (j->account);
+ xfree (j);
+}
+
/*
* Check if the script exists and if we can execute it.
*/
static int
-_check_script_permissions(char * location)
+_check_script_permissions(char * path)
{
- struct stat filestat;
- uid_t user;
- gid_t group;
+ struct stat st;
- if(stat(location,&filestat) < 0) {
+ if (stat(path, &st) < 0) {
plugin_errno = errno;
- return error("jobcomp/script script does not exist");
+ return error("jobcomp/script: %s does not exist", path);
}
- if(!(filestat.st_mode&S_IFREG)) {
+ if (!(st.st_mode & S_IFREG)) {
plugin_errno = EACCES;
- return error("jobcomp/script script isn't a regular file");
+ return error("jobcomp/script: %s isn't a regular file", path);
}
-
- user = geteuid();
- group = getegid();
- if(user == filestat.st_uid) {
- if (filestat.st_mode&S_IXUSR) {
- return SLURM_SUCCESS;
- } else {
- plugin_errno = EACCES;
- return error("jobcomp/script script is not executable");
- }
- } else if(group == filestat.st_gid) {
- if (filestat.st_mode&S_IXGRP) {
- return SLURM_SUCCESS;
- } else {
- plugin_errno = EACCES;
- return error("jobcomp/script script is not executable");
- }
- } else if(filestat.st_mode&S_IXOTH) {
- return SLURM_SUCCESS;
+
+ if (access(path, X_OK) < 0) {
+ plugin_errno = EACCES;
+ return error("jobcomp/script: %s is not executable", path);
}
- plugin_errno = EACCES;
- return error("jobcomp/script script is not executable");
+ return SLURM_SUCCESS;
}
-/* Write "name=value" to dest, returns pointer to after that key-pair string,
- * where the next key-pair (if any) should be written. */
-static char *_add_env(char *dest, char *name, char *value)
+static char ** _extend_env (char ***envp)
{
- int len;
+ char **ep;
+ size_t newcnt = (xsize (*envp) / sizeof (char *)) + 1;
+
+ *envp = xrealloc (*envp, newcnt * sizeof (char *));
- len = sprintf(dest, "%s=%s", name, value);
- return dest + len + 1;
+ (*envp)[newcnt - 1] = NULL;
+ ep = &((*envp)[newcnt - 2]);
+
+ /*
+ * Find last non-NULL entry
+ */
+ while (*ep == NULL)
+ --ep;
+
+ return (++ep);
}
-/* Create a new environment pointer containing information from
- * slurm_jobcomp_log_record so that the script can access it.
- */
-static char ** _create_environment(char *job, char *user, char *job_name,
- char *job_state, char *partition, char *limit,
- char* start, char * end, char * submit,
- char * batch, char *node_list, char *procs, char *account)
+static int _env_append (char ***envp, const char *name, const char *val)
{
- int len = 0;
- char ** envptr;
-
- /* Increase ENV_COUNT if new environment variables are added */
-#define ENV_COUNT 15
-
- /* Add strlen() of any sizable fields added to the array */
- len = strlen(job_name) + strlen(partition) + strlen(node_list) +
- strlen(account) + strlen(_PATH_STDPATH) + 1024;
-
- if (!(envptr = (char **)try_xmalloc(len)))
- return NULL;
- envptr[0] = (char *)envptr + (ENV_COUNT * sizeof(char *));
-
-/* NEXT_DEST = ADD_ENV(DEST, NAME=VALUE) */
- envptr[1] = _add_env(envptr[0], "JOBID", job);
- envptr[2] = _add_env(envptr[1], "UID", user);
- envptr[3] = _add_env(envptr[2], "JOBNAME", job_name);
- envptr[4] = _add_env(envptr[3], "JOBSTATE", job_state);
- envptr[5] = _add_env(envptr[4], "PARTITION", partition);
- envptr[6] = _add_env(envptr[5], "LIMIT", limit);
- envptr[7] = _add_env(envptr[6], "START", start);
- envptr[8] = _add_env(envptr[7], "END", end);
- envptr[9] = _add_env(envptr[8], "NODES", node_list);
- envptr[10] = _add_env(envptr[9], "SUBMIT", submit);
- envptr[11] = _add_env(envptr[10], "BATCH", batch);
- envptr[12] = _add_env(envptr[11], "PROCS", procs);
- envptr[13] = _add_env(envptr[12], "ACCOUNT", account);
+ char buf[4096];
+ char *entry;
+ char **ep;
+
+ if (val == NULL)
+ val = "";
+
+ snprintf (buf, sizeof (buf) - 1, "%s=%s", name, val);
+
+ if (!(entry = strdup (buf)))
+ return (-1);
+
+ ep = _extend_env (envp);
+ *ep = entry;
+
+ return (0);
+}
+
+
+static int _env_append_fmt (char ***envp, const char *name,
+ const char *fmt, ...)
+{
+ char val[1024];
+ va_list ap;
+
+ va_start (ap, fmt);
+ vsnprintf (val, sizeof (val) - 1, fmt, ap);
+ va_end (ap);
+
+ return (_env_append (envp, name, val));
+}
+
+static char ** _create_environment (struct jobcomp_info *job)
+{
+ char **env;
+
+ env = xmalloc (1 * sizeof (*env));
+ env[0] = NULL;
+
+ _env_append_fmt (&env, "JOBID", "%u", job->jobid);
+ _env_append_fmt (&env, "UID", "%u", job->uid);
+ _env_append_fmt (&env, "START", "%lu", job->start);
+ _env_append_fmt (&env, "END", "%lu", job->end);
+ _env_append_fmt (&env, "SUBMIT","%lu", job->submit);
+ _env_append_fmt (&env, "PROCS", "%u", job->nprocs);
+
+ _env_append (&env, "BATCH", (job->batch_flag ? "yes" : "no"));
+ _env_append (&env, "NODES", job->nodes);
+ _env_append (&env, "ACCOUNT", job->account);
+ _env_append (&env, "JOBNAME", job->name);
+ _env_append (&env, "JOBSTATE", job->jobstate);
+ _env_append (&env, "PARTITION", job->partition);
+
+ if (job->limit == INFINITE)
+ _env_append (&env, "LIMIT", "UNLIMITED");
+ else
+ _env_append_fmt (&env, "LIMIT", "%u", job->limit);
#ifdef _PATH_STDPATH
- _add_env(envptr[13], "PATH", _PATH_STDPATH);
+ _env_append (&env, "PATH", _PATH_STDPATH);
#else
- envptr[13] = NULL;
+ _env_append (&env, "PATH", "/bin:/usr/bin");
#endif
- /* envptr[14] through envptr[ENV_COUNT] are NULL from xmalloc.
- * Since last entry must be NULL, add new records between
- * "ACCOUNT" and "PATH" above. */
-
- return envptr;
+ return (env);
+}
+
+static int redirect_stdio (void)
+{
+ int devnull;
+ if ((devnull = open ("/dev/null", O_RDWR)) < 0)
+ return error ("jobcomp/script: Failed to open /dev/null: %m");
+ if (dup2 (devnull, STDIN_FILENO) < 0)
+ return error ("jobcomp/script: Failed to redirect stdin: %m");
+ if (dup2 (devnull, STDOUT_FILENO) < 0)
+ return error ("jobcomp/script: Failed to redirect stdout: %m");
+ if (dup2 (devnull, STDERR_FILENO) < 0)
+ return error ("jobcomp/script: Failed to redirect stderr: %m");
+ close (devnull);
+ return (0);
+}
+
+static void jobcomp_child (char * script, char **env)
+{
+ char * args[] = {script, NULL};
+ const char *tmpdir;
+
+#ifdef _PATH_TMP
+ tmpdir = _PATH_TMP;
+#else
+ tmpdir = "/tmp";
+#endif
+
+ /*
+ * Reinitialize log so we can log any errors for
+ * diagnosis
+ */
+ log_reinit ();
+
+ if (redirect_stdio () < 0)
+ exit (1);
+
+ if (chdir (tmpdir) != 0) {
+ error ("jobcomp/script: chdir (%s): %m", _PATH_TMP);
+ exit(1);
+ }
+
+ execve(script, args, env);
+
+ /*
+ * Failure of execve implies error
+ */
+ error ("jobcomp/script: execve(%s): %m\n", script);
+ exit (1);
}
+static int jobcomp_exec_child (char *script, char **env)
+{
+ pid_t pid;
+ int status = 0;
+
+ if (script == NULL || env == NULL)
+ return (-1);
+
+ if ((pid = fork()) < 0) {
+ error ("jobcomp/script: fork: %m");
+ return (-1);
+ }
+
+ if (pid == 0)
+ jobcomp_child (script, env);
+
+ /*
+ * Parent continues
+ */
+
+ if (waitpid(pid, &status, 0) < 0)
+ error ("jobcomp/script: waitpid: %m");
+
+ if (WEXITSTATUS(status))
+ error ("jobcomp/script: script %s exited with status %d\n",
+ script, WEXITSTATUS(status));
+
+ return (0);
+}
+
+
/*
* Thread function that executes a script
*/
-void *script_agent (void *args) {
- pid_t pid = -1;
- char user_id_str[32],job_id_str[32], nodes_cache[1];
- char start_str[32], end_str[32], lim_str[32];
- char submit_str[32], procs_str[32], account_str[128];
- char * argvp[] = {script, NULL};
- int status;
- char ** envp, *nodes, *batch_str;
- job_record job;
-
- while(1) {
+void * script_agent (void *args)
+{
+ info ("jobcomp_script_strerror (137) = %s\n",
+ jobcomp_script_strerror (137));
+
+ while (1) {
+ struct jobcomp_info *job;
+
pthread_mutex_lock(&comp_list_mutex);
- while ((list_is_empty(comp_list) != 0) && (agent_exit == 0)) {
- pthread_cond_wait(&comp_list_cond, &comp_list_mutex);
- }
- if (agent_exit) {
- pthread_mutex_unlock(&comp_list_mutex);
- return NULL;
- }
- job = (job_record)list_pop(comp_list);
- pthread_mutex_unlock(&comp_list_mutex);
- snprintf(user_id_str,sizeof(user_id_str),"%u",job->user_id);
- snprintf(job_id_str,sizeof(job_id_str),"%u",job->job_id);
- snprintf(start_str, sizeof(start_str),"%lu",
- (unsigned long)job->start);
- snprintf(end_str, sizeof(end_str),"%lu",
- (unsigned long)job->end);
- snprintf(submit_str, sizeof(submit_str),"%lu",
- (unsigned long)job->submit);
- if (job->batch_flag)
- batch_str = "yes";
- else
- batch_str = "no";
- nodes_cache[0] = '\0';
-
- if (job->limit == INFINITE) {
- strcpy(lim_str, "UNLIMITED");
- } else {
- snprintf(lim_str, sizeof(lim_str), "%lu",
- (unsigned long) job->limit);
- }
-
- if (job->node_list == NULL) {
- nodes = nodes_cache;
- } else {
- nodes = job->node_list;
- }
+ if (list_is_empty(comp_list) && !agent_exit)
+ pthread_cond_wait(&comp_list_cond, &comp_list_mutex);
- snprintf(procs_str, sizeof(procs_str), "%lu",
- (unsigned long) job->num_procs);
- if (job->account) {
- snprintf(account_str, sizeof(account_str), "%s",
- job->account);
- } else
- account_str[0] = '\0';
-
- /* Setup environment*/
- envp = _create_environment(job_id_str,user_id_str,
- job->job_name, job->job_state,
- job->partition, lim_str,
- start_str, end_str, submit_str,
- batch_str, nodes, procs_str,
- account_str);
-
- if(envp == NULL) {
- plugin_errno = ENOMEM;
- }
+ /*
+ * It is safe to unlock list mutex here. List has its
+ * own internal mutex that protects the comp_list itself
+ */
+ pthread_mutex_unlock(&comp_list_mutex);
- pid = fork();
+ if ((job = list_pop(comp_list))) {
+ char **envp;
- if (pid < 0) {
- xfree(envp);
- job_record_destroy((void *)job);
- plugin_errno = errno;
- } else if (pid == 0) {
- /*Change directory to tmp*/
- if (chdir(_PATH_TMP) != 0) {
- exit(errno);
+ if ((envp = _create_environment (job)) == NULL) {
+ error ("jobcomp/script: create env failed!");
+ jobcomp_info_destroy (job);
+ continue;
}
- /*Redirect stdin, stderr, and stdout to /dev/null*/
- if (freopen(_PATH_DEVNULL, "rb", stdin) == NULL) {
- exit(errno);
- }
- if (freopen(_PATH_DEVNULL, "wb", stdout) == NULL) {
- exit(errno);
- }
- if (freopen(_PATH_DEVNULL, "wb", stderr) == NULL) {
- exit(errno);
- }
-
- /*Exec Script*/
- execve(script,argvp,envp);
- } else {
+ if (jobcomp_exec_child (script, envp) < 0)
+ error ("jobcomp/script: %s failed");
+
xfree(envp);
- job_record_destroy((void *)job);
- waitpid(pid,&status,0);
- if(WIFEXITED(status) && WEXITSTATUS(status)) {
- plugin_errno = WEXITSTATUS(status);
- }
+ jobcomp_info_destroy (job);
}
+
+ /*
+ * Exit if flag is set and we have no more entries to log
+ */
+ if (agent_exit && list_is_empty (comp_list))
+ break;
}
+
+ return NULL;
}
/*
* init() is called when the plugin is loaded, before any other functions
* are called. Put global initialization here.
*/
-int init ( void )
+int init (void)
{
pthread_attr_t attr;
@@ -334,18 +453,17 @@ int init ( void )
pthread_mutex_lock(&thread_flag_mutex);
- comp_list = list_create((ListDelF) job_record_destroy);
- if(comp_list == NULL) {
+ if (!(comp_list = list_create((ListDelF) jobcomp_info_destroy)))
return SLURM_ERROR;
- }
if (script_thread) {
debug2( "Script thread already running, not starting another");
pthread_mutex_unlock(&thread_flag_mutex);
return SLURM_ERROR;
}
+
slurm_attr_init(&attr);
- pthread_attr_setdetachstate( &attr, PTHREAD_CREATE_DETACHED );
+ pthread_attr_setdetachstate(&attr, PTHREAD_CREATE_DETACHED);
pthread_create(&script_thread, &attr, script_agent, NULL);
pthread_mutex_unlock(&thread_flag_mutex);
@@ -355,16 +473,15 @@ int init ( void )
}
/* Set the location of the script to run*/
-int slurm_jobcomp_set_location ( char * location )
+int slurm_jobcomp_set_location (char * location)
{
if (location == NULL) {
plugin_errno = EACCES;
return error("jobcomp/script JobCompLoc needs to be set");
}
- if (_check_script_permissions(location) != SLURM_SUCCESS) {
+ if (_check_script_permissions(location) != SLURM_SUCCESS)
return SLURM_ERROR;
- }
xfree(script);
script = xstrdup(location);
@@ -372,47 +489,34 @@ int slurm_jobcomp_set_location ( char * location )
return SLURM_SUCCESS;
}
-int slurm_jobcomp_log_record ( struct job_record *job_ptr )
+int slurm_jobcomp_log_record (struct job_record *record)
{
- job_record job;
- enum job_states job_state;
- time_t submit;
+ struct jobcomp_info * job;
debug3("Entering slurm_jobcomp_log_record");
- /* Job will typically be COMPLETING when this is called.
- * We remove this flag to get the eventual completion state:
- * JOB_FAILED, JOB_TIMEOUT, etc. */
- job_state = job_ptr->job_state & (~JOB_COMPLETING);
-
- if (job_ptr->details)
- submit = job_ptr->details->submit_time;
- else
- submit = job_ptr->start_time;
-
- job = job_record_create(job_ptr->job_id, job_ptr->user_id, job_ptr->name,
- job_state_string(job_state),
- job_ptr->partition, job_ptr->time_limit,
- job_ptr->start_time, job_ptr->end_time,
- submit, job_ptr->batch_flag, job_ptr->nodes,
- job_ptr->num_procs, job_ptr->account);
+
+ if ((job = jobcomp_info_create (record)))
+ return error ("jobcomp/script: Failed to create job info!");
+
pthread_mutex_lock(&comp_list_mutex);
- list_append(comp_list,(void *)job);
+ list_append(comp_list, job);
pthread_mutex_unlock(&comp_list_mutex);
+
pthread_cond_broadcast(&comp_list_cond);
+
return SLURM_SUCCESS;
}
/* Return the error code of the plugin*/
-int slurm_jobcomp_get_errno( void )
+int slurm_jobcomp_get_errno(void)
{
return plugin_errno;
}
/* Return a string representation of the error */
-char *slurm_jobcomp_strerror( int errnum )
+const char * slurm_jobcomp_strerror(int errnum)
{
- strerror_r(errnum,error_str,sizeof(error_str));
- return error_str;
+ return jobcomp_script_strerror (errnum);
}
static int _wait_for_thread (pthread_t thread_id)
@@ -424,7 +528,13 @@ static int _wait_for_thread (pthread_t thread_id)
return SLURM_SUCCESS;
usleep(1000);
}
- error("Could not kill jobcomp script pthread");
+
+ /*
+ * jobcomp thread appears to be stuck, try harder:
+ */
+ if (pthread_kill(thread_id, SIGKILL))
+ error("Could not kill jobcomp script pthread");
+
return SLURM_ERROR;
}
--
GitLab