diff --git a/src/plugins/sched/backfill/backfill.c b/src/plugins/sched/backfill/backfill.c
index ec1fa9061ffb46b74781cbdc2a0c80547e27f66f..49215b16f9c1aa4a433fa86a9fabc5eee0df1cd9 100644
--- a/src/plugins/sched/backfill/backfill.c
+++ b/src/plugins/sched/backfill/backfill.c
@@ -64,6 +64,7 @@
#include "src/common/slurm_protocol_api.h"
#include "src/common/xmalloc.h"
#include "src/common/xstring.h"
+#include "src/slurmctld/job_scheduler.h"
#include "src/slurmctld/locks.h"
#include "src/slurmctld/slurmctld.h"
diff --git a/src/plugins/sched/wiki/job_modify.c b/src/plugins/sched/wiki/job_modify.c
index 35486df25611c6ac45b854a20dd5bb1471fd47d0..19953c2eff65e9166c9e075b3fb91d7129100a16 100644
--- a/src/plugins/sched/wiki/job_modify.c
+++ b/src/plugins/sched/wiki/job_modify.c
@@ -37,6 +37,7 @@
#include "./msg.h"
#include <strings.h>
+#include "src/slurmctld/job_scheduler.h"
#include "src/slurmctld/locks.h"
#include "src/slurmctld/slurmctld.h"
diff --git a/src/plugins/sched/wiki/start_job.c b/src/plugins/sched/wiki/start_job.c
index f7cdd46d189ad8b080c780beb3ac794e7a9ca0ee..015fdf4256a3184860f263ba0b3cc9325921f7d5 100644
--- a/src/plugins/sched/wiki/start_job.c
+++ b/src/plugins/sched/wiki/start_job.c
@@ -39,6 +39,7 @@
#include "src/common/node_select.h"
#include "src/common/slurm_protocol_defs.h"
#include "src/common/xstring.h"
+#include "src/slurmctld/job_scheduler.h"
#include "src/slurmctld/locks.h"
#include "src/slurmctld/slurmctld.h"
#include "src/slurmctld/state_save.h"
diff --git a/src/plugins/sched/wiki2/job_modify.c b/src/plugins/sched/wiki2/job_modify.c
index 25e1b1aaf862ab78f8560283f8a0fff39ee498a7..a29a6012efc4c7484e2db1ca3f425756940f4e2a 100644
--- a/src/plugins/sched/wiki2/job_modify.c
+++ b/src/plugins/sched/wiki2/job_modify.c
@@ -37,6 +37,7 @@
#include "./msg.h"
#include <strings.h>
+#include "src/slurmctld/job_scheduler.h"
#include "src/slurmctld/locks.h"
#include "src/slurmctld/slurmctld.h"
diff --git a/src/plugins/sched/wiki2/start_job.c b/src/plugins/sched/wiki2/start_job.c
index b28c7f196f85d800ec9f891ad6a862f1d5e9d53e..97a4156b71e0843dc03bd2adabdfb9cf78ddcb1e 100644
--- a/src/plugins/sched/wiki2/start_job.c
+++ b/src/plugins/sched/wiki2/start_job.c
@@ -39,6 +39,7 @@
#include "src/common/node_select.h"
#include "src/common/slurm_protocol_defs.h"
#include "src/common/xstring.h"
+#include "src/slurmctld/job_scheduler.h"
#include "src/slurmctld/locks.h"
#include "src/slurmctld/slurmctld.h"
#include "src/slurmctld/state_save.h"
diff --git a/src/slurmctld/Makefile.am b/src/slurmctld/Makefile.am
index 5852f1e58ea7af94af56011527746e519f28f766..39181b917e79f8bda58e36db13a1f0b974ded8be 100644
--- a/src/slurmctld/Makefile.am
+++ b/src/slurmctld/Makefile.am
@@ -20,10 +20,12 @@ slurmctld_SOURCES = \
controller.c \
job_mgr.c \
job_scheduler.c \
+ job_scheduler.h \
locks.c \
locks.h \
node_mgr.c \
- node_scheduler.c node_scheduler.h \
+ node_scheduler.c \
+ node_scheduler.h \
partition_mgr.c \
ping_nodes.c \
ping_nodes.h \
diff --git a/src/slurmctld/Makefile.in b/src/slurmctld/Makefile.in
index 3d73446210f9a7545640f56875f85a722900d30d..363c513e1dfeaf522f4319c498d69bbfea2ad350 100644
--- a/src/slurmctld/Makefile.in
+++ b/src/slurmctld/Makefile.in
@@ -278,10 +278,12 @@ slurmctld_SOURCES = \
controller.c \
job_mgr.c \
job_scheduler.c \
+ job_scheduler.h \
locks.c \
locks.h \
node_mgr.c \
- node_scheduler.c node_scheduler.h \
+ node_scheduler.c \
+ node_scheduler.h \
partition_mgr.c \
ping_nodes.c \
ping_nodes.h \
diff --git a/src/slurmctld/agent.c b/src/slurmctld/agent.c
index e98a488dd7f0b29c8fc1f831cf76f17a6d38abcf..e8bf2fa39f19e040032e5bb237a33971ee3a91bd 100644
--- a/src/slurmctld/agent.c
+++ b/src/slurmctld/agent.c
@@ -89,6 +89,7 @@
#include "src/common/uid.h"
#include "src/common/forward.h"
#include "src/slurmctld/agent.h"
+#include "src/slurmctld/job_scheduler.h"
#include "src/slurmctld/locks.h"
#include "src/slurmctld/ping_nodes.h"
#include "src/slurmctld/slurmctld.h"
diff --git a/src/slurmctld/agent.h b/src/slurmctld/agent.h
index 1ccb6277bd90a6d44443e0556428b681078aedd0..c03b47c15e2ec80f5ccf3789790a68a8194f2196 100644
--- a/src/slurmctld/agent.h
+++ b/src/slurmctld/agent.h
@@ -42,7 +42,6 @@
#ifndef _AGENT_H
#define _AGENT_H
-#include "src/slurmctld/agent.h"
#include "src/slurmctld/slurmctld.h"
#define AGENT_IS_THREAD 1 /* set if agent itself a thread of
diff --git a/src/slurmctld/controller.c b/src/slurmctld/controller.c
index 16b8d2a4bb3c311d6a02cae00bfe1f394550ab88..93ab07c7c1dfcd01b2d4848b9568e3cd28f59e2b 100644
--- a/src/slurmctld/controller.c
+++ b/src/slurmctld/controller.c
@@ -79,6 +79,7 @@
#include "src/common/xstring.h"
#include "src/slurmctld/agent.h"
+#include "src/slurmctld/job_scheduler.h"
#include "src/slurmctld/locks.h"
#include "src/slurmctld/ping_nodes.h"
#include "src/slurmctld/proc_req.h"
diff --git a/src/slurmctld/job_mgr.c b/src/slurmctld/job_mgr.c
index e9b8abd4e7a7f05980756bb4724d8010355dd608..890c0190f7744b7c2e389bbf9d4297ab47e91a28 100644
--- a/src/slurmctld/job_mgr.c
+++ b/src/slurmctld/job_mgr.c
@@ -69,6 +69,7 @@
#include "src/common/slurm_protocol_pack.h"
#include "src/slurmctld/agent.h"
+#include "src/slurmctld/job_scheduler.h"
#include "src/slurmctld/locks.h"
#include "src/slurmctld/node_scheduler.h"
#include "src/slurmctld/proc_req.h"
diff --git a/src/slurmctld/job_scheduler.c b/src/slurmctld/job_scheduler.c
index 72e5b1e109217f28d8012af78d18cf4eb942478e..9cdf7fd6a1a9bf47b7c25f197a1f71f3b3a37fea 100644
--- a/src/slurmctld/job_scheduler.c
+++ b/src/slurmctld/job_scheduler.c
@@ -53,6 +53,7 @@
#include "src/common/xstring.h"
#include "src/slurmctld/agent.h"
+#include "src/slurmctld/job_scheduler.h"
#include "src/slurmctld/locks.h"
#include "src/slurmctld/node_scheduler.h"
#include "src/slurmctld/slurmctld.h"
@@ -61,27 +62,17 @@
#define _DEBUG 0
#define MAX_RETRIES 10
-struct job_queue {
- struct job_record *job_ptr;
- uint32_t job_priority;
- uint16_t part_priority;
-};
-
-static int _build_job_queue(struct job_queue **job_queue);
static void _depend_list_del(void *dep_ptr);
static void _launch_job(struct job_record *job_ptr);
-static void _sort_job_queue(struct job_queue *job_queue,
- int job_queue_size);
static char **_xduparray(uint16_t size, char ** array);
/*
- * _build_job_queue - build (non-priority ordered) list of pending jobs
+ * build_job_queue - build (non-priority ordered) list of pending jobs
* OUT job_queue - pointer to job queue
* RET number of entries in job_queue
- * global: job_list - global list of job records
* NOTE: the buffer at *job_queue must be xfreed by the caller
*/
-static int _build_job_queue(struct job_queue **job_queue)
+extern int build_job_queue(struct job_queue **job_queue)
{
ListIterator job_iterator;
struct job_record *job_ptr = NULL;
@@ -189,14 +180,12 @@ extern void set_job_elig_time(void)
* pending jobs for each partition will be scheduled in priority
* order until a request fails
* RET count of jobs scheduled
- * global: job_list - global list of job records
- * last_job_update - time of last update to job table
* Note: We re-build the queue every time. Jobs can not only be added
* or removed from the queue, but have their priority or partition
* changed with the update_job RPC. In general nodes will be in priority
* order (by submit time), so the sorting should be pretty fast.
*/
-int schedule(void)
+extern int schedule(void)
{
struct job_queue *job_queue;
int i, j, error_code, failed_part_cnt, job_queue_size, job_cnt = 0;
@@ -232,12 +221,12 @@ int schedule(void)
return SLURM_SUCCESS;
}
debug("Running job scheduler");
- job_queue_size = _build_job_queue(&job_queue);
+ job_queue_size = build_job_queue(&job_queue);
if (job_queue_size == 0) {
unlock_slurmctld(job_write_lock);
return SLURM_SUCCESS;
}
- _sort_job_queue(job_queue, job_queue_size);
+ sort_job_queue(job_queue, job_queue_size);
failed_part_cnt = 0;
failed_parts = NULL;
@@ -324,11 +313,11 @@ int schedule(void)
/*
- * _sort_job_queue - sort job_queue in decending priority order
+ * sort_job_queue - sort job_queue in decending priority order
* IN job_queue_size - count of elements in the job queue
* IN/OUT job_queue - pointer to sorted job queue
*/
-static void _sort_job_queue(struct job_queue *job_queue, int job_queue_size)
+extern void sort_job_queue(struct job_queue *job_queue, int job_queue_size)
{
int i, j, top_prio_inx;
struct job_record *tmp_job_ptr;
diff --git a/src/slurmctld/job_scheduler.h b/src/slurmctld/job_scheduler.h
new file mode 100644
index 0000000000000000000000000000000000000000..70268735c4cce3dbd21d647923bb4d953c6b9b71
--- /dev/null
+++ b/src/slurmctld/job_scheduler.h
@@ -0,0 +1,129 @@
+/*****************************************************************************\
+ * job_scheduler.h - data structures and function definitions for scheduling
+ * of pending jobs in priority order
+ *****************************************************************************
+ * Copyright (C) 2002-2006 The Regents of the University of California.
+ * Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
+ * Written by Morris Jette <jette@llnl.gov>, et. al.
+ * Derived from dsh written by Jim Garlick <garlick1@llnl.gov>
+ * 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.
+\*****************************************************************************/
+
+#ifndef _JOB_SCHEDULER_H
+#define _JOB_SCHEDULER_H
+
+#include "src/slurmctld/slurmctld.h"
+
+struct job_queue {
+ struct job_record *job_ptr;
+ uint32_t job_priority;
+ uint16_t part_priority;
+};
+
+/*
+ * build_job_queue - build (non-priority ordered) list of pending jobs
+ * OUT job_queue - pointer to job queue
+ * RET number of entries in job_queue
+ * NOTE: the buffer at *job_queue must be xfreed by the caller
+ */
+extern int build_job_queue(struct job_queue **job_queue);
+
+/*
+ * job_is_completing - Determine if jobs are in the process of completing.
+ * RET - True of any job is in the process of completing
+ * NOTE: This function can reduce resource fragmentation, which is a
+ * critical issue on Elan interconnect based systems.
+ */
+
+extern bool job_is_completing(void);
+
+/* Determine if a pending job will run using only the specified nodes
+ * (in job_desc_msg->req_nodes), build response message and return
+ * SLURM_SUCCESS on success. Otherwise return an error code. Caller
+ * must free response message */
+extern int job_start_data(job_desc_msg_t *job_desc_msg,
+ will_run_response_msg_t **resp);
+
+/*
+ * make_batch_job_cred - add a job credential to the batch_job_launch_msg
+ * IN/OUT launch_msg_ptr - batch_job_launch_msg in which job_id, step_id,
+ * uid and nodes have already been set
+ * RET 0 or error code
+ */
+extern int make_batch_job_cred(batch_job_launch_msg_t *launch_msg_ptr);
+
+/* Print a job's dependency information based upon job_ptr->depend_list */
+extern void print_job_dependency(struct job_record *job_ptr);
+
+/*
+ * schedule - attempt to schedule all pending jobs
+ * pending jobs for each partition will be scheduled in priority
+ * order until a request fails
+ * RET count of jobs scheduled
+ * Note: We re-build the queue every time. Jobs can not only be added
+ * or removed from the queue, but have their priority or partition
+ * changed with the update_job RPC. In general nodes will be in priority
+ * order (by submit time), so the sorting should be pretty fast.
+ */
+extern int schedule(void);
+
+/*
+ * set_job_elig_time - set the eligible time for pending jobs once their
+ * dependencies are lifted (in job->details->begin_time)
+ */
+extern void set_job_elig_time(void);
+
+/*
+ * sort_job_queue - sort job_queue in decending priority order
+ * IN job_queue_size - count of elements in the job queue
+ * IN/OUT job_queue - pointer to sorted job queue
+ */
+extern void sort_job_queue(struct job_queue *job_queue, int job_queue_size);
+
+/*
+ * Determine if a job's dependencies are met
+ * RET: 0 = no dependencies
+ * 1 = dependencies remain
+ * 2 = failure (job completion code not per dependency), delete the job
+ */
+extern int test_job_dependency(struct job_record *job_ptr);
+
+/*
+ * Parse a job dependency string and use it to establish a "depend_spec"
+ * list of dependencies. We accept both old format (a single job ID) and
+ * new format (e.g. "afterok:123:124,after:128").
+ * IN job_ptr - job record to have dependency and depend_list updated
+ * IN new_depend - new dependency description
+ * RET returns an error code from slurm_errno.h
+ */
+extern int update_job_dependency(struct job_record *job_ptr, char *new_depend);
+
+#endif /* !_JOB_SCHEDULER_H */
diff --git a/src/slurmctld/node_scheduler.c b/src/slurmctld/node_scheduler.c
index b51fdea162be11ba14ab917369e45440426b687b..6c06d4d896df22add634d67cd198d4ec699d3ba3 100644
--- a/src/slurmctld/node_scheduler.c
+++ b/src/slurmctld/node_scheduler.c
@@ -1161,7 +1161,7 @@ extern int job_req_node_filter(struct job_record *job_ptr,
multi_core_data_t *mc_ptr;
struct node_record *node_ptr;
struct config_record *config_ptr;
- bitstr_t *feature_bitmap;
+ bitstr_t *feature_bitmap = NULL;
if (detail_ptr == NULL) {
error("job_req_node_filter: job %u has no details",
diff --git a/src/slurmctld/proc_req.c b/src/slurmctld/proc_req.c
index a469e75460ee1f59df1c7db593b242bc4cb7e008..aa23dd722a8fc5af11f3c783fcb8ab98bb5a423b 100644
--- a/src/slurmctld/proc_req.c
+++ b/src/slurmctld/proc_req.c
@@ -71,6 +71,7 @@
#include "src/common/forward.h"
#include "src/slurmctld/agent.h"
+#include "src/slurmctld/job_scheduler.h"
#include "src/slurmctld/locks.h"
#include "src/slurmctld/proc_req.h"
#include "src/slurmctld/read_config.h"
diff --git a/src/slurmctld/read_config.c b/src/slurmctld/read_config.c
index 346b62925eb0da74123fbba52db9a605861617ea..8a8d3956ab1eaa6c63f6af15629a5b2cc499980f 100644
--- a/src/slurmctld/read_config.c
+++ b/src/slurmctld/read_config.c
@@ -68,6 +68,7 @@
#include "src/common/slurm_jobacct_storage.h"
#include "src/common/slurm_rlimits_info.h"
+#include "src/slurmctld/job_scheduler.h"
#include "src/slurmctld/locks.h"
#include "src/slurmctld/node_scheduler.h"
#include "src/slurmctld/proc_req.h"
diff --git a/src/slurmctld/slurmctld.h b/src/slurmctld/slurmctld.h
index 76783be10a2f013dd30b6ef5bf34a9b32d98df13..672273bbf226f0a161c8e219e27822f2fd4ef053 100644
--- a/src/slurmctld/slurmctld.h
+++ b/src/slurmctld/slurmctld.h
@@ -789,14 +789,6 @@ extern int job_end_time(job_alloc_info_msg_t *time_req_msg,
/* job_fini - free all memory associated with job records */
extern void job_fini (void);
-/*
- * job_is_completing - Determine if jobs are in the process of completing.
- * RET - True of any job is in the process of completing
- * NOTE: This function can reduce resource fragmentation, which is a
- * critical issue on Elan interconnect based systems.
- */
-extern bool job_is_completing(void);
-
/*
* job_fail - terminate a job due to initiation failure
* IN job_id - id of the job to be killed
@@ -951,13 +943,6 @@ extern int job_step_signal(uint32_t job_id, uint32_t step_id,
*/
extern void job_time_limit (void);
-/* Determine if a pending job will run using only the specified nodes
- * (in job_desc_msg->req_nodes), build response message and return
- * SLURM_SUCCESS on success. Otherwise return an error code. Caller
- * must free response message */
-extern int job_start_data(job_desc_msg_t *job_desc_msg,
- will_run_response_msg_t **resp);
-
/*
* kill_job_by_part_name - Given a partition name, deallocate resource for
* its jobs and kill them
@@ -1041,14 +1026,6 @@ extern int load_all_part_state ( void );
*/
extern int load_step_state(struct job_record *job_ptr, Buf buffer);
-/*
- * make_batch_job_cred - add a job credential to the batch_job_launch_msg
- * IN/OUT launch_msg_ptr - batch_job_launch_msg in which job_id, step_id,
- * uid and nodes have already been set
- * RET 0 or error code
- */
-extern int make_batch_job_cred(batch_job_launch_msg_t *launch_msg_ptr);
-
/* make_node_alloc - flag specified node as allocated to a job
* IN node_ptr - pointer to node being allocated
* IN job_ptr - pointer to job that is starting
@@ -1204,9 +1181,6 @@ extern void part_filter_set(uid_t uid);
/* part_fini - free all memory associated with partition records */
void part_fini (void);
-/* Print a job's dependency information based upon job_ptr->depend_list */
-extern void print_job_dependency(struct job_record *job_ptr);
-
/*
* purge_old_job - purge old job records.
* the jobs must have completed at least MIN_JOB_AGE minutes ago
@@ -1261,26 +1235,6 @@ extern void run_backup(void);
/* save_all_state - save entire slurmctld state for later recovery */
extern void save_all_state(void);
-/*
- * schedule - attempt to schedule all pending jobs
- * pending jobs for each partition will be scheduled in priority
- * order until a request fails
- * RET count of jobs scheduled
- * global: job_list - global list of job records
- * last_job_update - time of last update to job table
- * Note: We re-build the queue every time. Jobs can not only be added
- * or removed from the queue, but have their priority or partition
- * changed with the update_job RPC. In general nodes will be in priority
- * order (by submit time), so the sorting should be pretty fast.
- */
-extern int schedule (void);
-
-/*
- * set_job_elig_time - set the eligible time for pending jobs once their
- * dependencies are lifted (in job->details->begin_time)
- */
-extern void set_job_elig_time(void);
-
/*
* set_node_down - make the specified node's state DOWN if possible
* (not in a DRAIN state), kill jobs as needed
@@ -1391,14 +1345,6 @@ extern void suspend_job_step(struct job_record *job_ptr);
*/
extern int sync_job_files(void);
-/*
- * Determine if a job's dependencies are met
- * RET: 0 = no dependencies
- * 1 = dependencies remain
- * 2 = failure (job completion code not per dependency), delete the job
- */
-extern int test_job_dependency(struct job_record *job_ptr);
-
/*
* update_job - update a job's parameters per the supplied specifications
* IN job_specs - a job's specification
@@ -1409,16 +1355,6 @@ extern int test_job_dependency(struct job_record *job_ptr);
*/
extern int update_job (job_desc_msg_t * job_specs, uid_t uid);
-/*
- * Parse a job dependency string and use it to establish a "depend_spec"
- * list of dependencies. We accept both old format (a single job ID) and
- * new format (e.g. "afterok:123:124,after:128").
- * IN job_ptr - job record to have dependency and depend_list updated
- * IN new_depend - new dependency description
- * RET returns an error code from slurm_errno.h
- */
-extern int update_job_dependency(struct job_record *job_ptr, char *new_depend);
-
/* Reset nodes_completing field for all jobs */
extern void update_job_nodes_completing(void);