diff --git a/NEWS b/NEWS
index 91263d13e49d9ce364dcfe122f679ce9b0eee2ec..6a70c6b987e74cfbcc8f9a87cd025e091faa705d 100644
--- a/NEWS
+++ b/NEWS
@@ -111,6 +111,9 @@ documents those changes that are of interest to users and admins.
together across clusters or within clusters that are not related. Use
the --wckey option in srun, sbatch or salloc or set the SLURM_WCKEY env
var to have this set. Use sreport with the wckey option to view reports.
+ -- Added configuration parameter BatchStartTimeout to control how long to
+ allow for a batch job prolog and environment loading (for Moab) to run.
+ See "man slurm.conf" for details.
* Changes in SLURM 1.3.11
=========================
diff --git a/doc/html/configurator.html.in b/doc/html/configurator.html.in
index 6ec53b9c407183cd3a971544174396e21e4f59a0..41cdf10e56b9bcf51600988063c01fed0f5c382d 100644
--- a/doc/html/configurator.html.in
+++ b/doc/html/configurator.html.in
@@ -191,6 +191,7 @@ function displayfile()
"# <br>" +
"# <br>" +
"# TIMERS <br>" +
+ "#BatchStartTimeout=10 <br>" +
"#EpilogMsgTime=2000 <br>" +
"#GetEnvTimeout=2 <br>" +
"#HealthCheckInterval=0 <br>" +
diff --git a/doc/man/man5/slurm.conf.5 b/doc/man/man5/slurm.conf.5
index 2707e0dfe6cd457e4996fcac39542065e386866d..1f5bddd448c496886df7e604acfa8cf919ee143d 100644
--- a/doc/man/man5/slurm.conf.5
+++ b/doc/man/man5/slurm.conf.5
@@ -156,6 +156,17 @@ This should be a node name without the full domain name (e.g. "lx0002").
While not essential, it is recommended that you specify a backup controller.
See the \fBRELOCATING CONTROLLERS\fR section if you change this.
+.TP
+\fBBatchStartTimeout\fR
+The maximum time (in seconds) that a batch job is permitted for
+launching before being considered missing and releasing the
+allocation. The default value is 10 (seconds). Larger values may
+be required if more time is required to execute the \fBProlog\fR,
+loading user environment variables (for Moab spawned jobs), or the
+slurmd daemon gets paged from memory.
+NOTE: The value will not be reported by "scontrol show config" command
+until SLURM version 1.4.
+
.TP
\fBCacheGroups\fR
If set to 1, the slurmd daemon will cache /etc/groups entries.
diff --git a/slurm/slurm.h.in b/slurm/slurm.h.in
index 3893ae9a8b59f66c265ea173f153ba606e6023f3..bd6814601b0a30274e925bb14e39285411eedbe1 100644
--- a/slurm/slurm.h.in
+++ b/slurm/slurm.h.in
@@ -973,6 +973,7 @@ typedef struct slurm_ctl_conf {
char *authtype; /* authentication type */
char *backup_addr; /* comm path of slurmctld secondary server */
char *backup_controller;/* name of slurmctld secondary server */
+ uint16_t batch_start_timeout; /* max secs for batch job to start */
time_t boot_time; /* time slurmctld last booted */
uint16_t cache_groups; /* cache /etc/groups to avoid initgroups(2) */
char *checkpoint_type; /* checkpoint plugin type */
diff --git a/src/common/read_config.c b/src/common/read_config.c
index c12b5be689402bdf0b8f3b5a09ff29d6d5908d06..dfd2b10a9c1deefb8e51d0fcf9ca3cbc940bd31e 100644
--- a/src/common/read_config.c
+++ b/src/common/read_config.c
@@ -135,6 +135,7 @@ s_p_options_t slurm_conf_options[] = {
{"AuthType", S_P_STRING},
{"BackupAddr", S_P_STRING},
{"BackupController", S_P_STRING},
+ {"BatchStartTimeout", S_P_UINT16},
{"CheckpointType", S_P_STRING},
{"CacheGroups", S_P_UINT16},
{"ClusterName", S_P_STRING},
@@ -1231,6 +1232,7 @@ init_slurm_conf (slurm_ctl_conf_t *ctl_conf_ptr)
xfree (ctl_conf_ptr->authtype);
xfree (ctl_conf_ptr->backup_addr);
xfree (ctl_conf_ptr->backup_controller);
+ ctl_conf_ptr->batch_start_timeout = 0;
ctl_conf_ptr->cache_groups = 0;
xfree (ctl_conf_ptr->checkpoint_type);
xfree (ctl_conf_ptr->cluster_name);
@@ -1546,6 +1548,10 @@ validate_and_set_defaults(slurm_ctl_conf_t *conf, s_p_hashtbl_t *hashtbl)
conf->backup_addr = xstrdup(conf->backup_controller);
}
+ if (!s_p_get_uint16(&conf->batch_start_timeout, "BatchStartTimeout",
+ hashtbl))
+ conf->batch_start_timeout = DEFAULT_BATCH_START_TIMEOUT;
+
s_p_get_string(&conf->cluster_name, "ClusterName", hashtbl);
if (!s_p_get_uint16(&conf->complete_wait, "CompleteWait", hashtbl))
diff --git a/src/common/read_config.h b/src/common/read_config.h
index cdf9b2701c9d4d47243737375e4ff876f4acdcf2..ae5e4583e3646de5daf266cd50af0226cc7d033e 100644
--- a/src/common/read_config.h
+++ b/src/common/read_config.h
@@ -59,6 +59,7 @@ enum {
#define DEFAULT_ACCOUNTING_ENFORCE ACCOUNTING_ENFORCE_NONE
#define DEFAULT_ACCOUNTING_STORAGE_TYPE "accounting_storage/none"
#define DEFAULT_AUTH_TYPE "auth/munge"
+#define DEFAULT_BATCH_START_TIMEOUT 10
#define DEFAULT_CACHE_GROUPS 0
#define DEFAULT_COMPLETE_WAIT 0
#define DEFAULT_CRYPTO_TYPE "crypto/munge"
diff --git a/src/common/slurm_protocol_api.c b/src/common/slurm_protocol_api.c
index 3d524f427f7cdabaf4cb8a9c01297a1639d17fdc..5a3847c874736fb3c4e3a4056458ff9774674d24 100644
--- a/src/common/slurm_protocol_api.c
+++ b/src/common/slurm_protocol_api.c
@@ -114,7 +114,7 @@ int slurm_set_api_config(slurm_protocol_config_t * protocol_conf)
* returns a pointer to the current slurm_protocol_config object
* RET slurm_protocol_config_t - current slurm_protocol_config object
*/
-slurm_protocol_config_t *slurm_get_api_config()
+slurm_protocol_config_t *slurm_get_api_config(void)
{
return proto_conf;
}
@@ -134,7 +134,7 @@ extern void slurm_api_set_conf_file(char *pathname)
* the compiled in default slurm_protocol_config object is initialized
* RET int - return code
*/
-int slurm_api_set_default_config()
+int slurm_api_set_default_config(void)
{
int rc = SLURM_SUCCESS;
slurm_ctl_conf_t *conf;
@@ -206,6 +206,23 @@ uint16_t slurm_get_complete_wait(void)
/* slurm_mutex_lock(&config_lock); */
/* } */
+/* slurm_get_batch_start_timeout
+ * RET BatchStartTimeout value from slurm.conf
+ */
+uint16_t slurm_get_batch_start_timeout(void)
+{
+ uint16_t batch_start_timeout = 0;
+ slurm_ctl_conf_t *conf;
+
+ if(slurmdbd_conf) {
+ } else {
+ conf = slurm_conf_lock();
+ batch_start_timeout = conf->batch_start_timeout;
+ slurm_conf_unlock();
+ }
+ return batch_start_timeout;
+}
+
/* slurm_get_def_mem_per_task
* RET DefMemPerTask value from slurm.conf
*/
diff --git a/src/common/slurm_protocol_api.h b/src/common/slurm_protocol_api.h
index 5086125f26d2f51cd473a94b0502794b4ed05e3b..62274160f5483766fa82ea4d91db94b3dbb93529 100644
--- a/src/common/slurm_protocol_api.h
+++ b/src/common/slurm_protocol_api.h
@@ -95,7 +95,12 @@ int inline slurm_set_api_config(slurm_protocol_config_t * protocol_conf);
* returns a pointer to the current slurm_protocol_config object
* RET slurm_protocol_config_t - current slurm_protocol_config object
*/
-inline slurm_protocol_config_t *slurm_get_api_config();
+inline slurm_protocol_config_t *slurm_get_api_config(void);
+
+/* slurm_get_batch_start_timeout
+ * RET BatchStartTimeout value from slurm.conf
+ */
+uint16_t slurm_get_batch_start_timeout(void);
/* slurm_get_complete_wait
* RET CompleteWait value from slurm.conf
diff --git a/src/slurmctld/job_mgr.c b/src/slurmctld/job_mgr.c
index 0472ed2f06a06ca90d884afd7e7d8acc79cc5fb3..7b8f28581b8e8be2dc91dffba33f77545c04ec45 100644
--- a/src/slurmctld/job_mgr.c
+++ b/src/slurmctld/job_mgr.c
@@ -81,7 +81,6 @@
#include "src/slurmctld/srun_comm.h"
#include "src/slurmctld/trigger_mgr.h"
-#define BATCH_JOB_LAUNCH_TIME 1200 /* seconds for prolog & env var load */
#define DETAILS_FLAG 0xdddd
#define MAX_RETRIES 10
#define SLURM_CREATE_JOB_FLAG_NO_ALLOCATE_0 0
@@ -4929,7 +4928,7 @@ static void _purge_lost_batch_jobs(int node_inx, time_t now)
{
ListIterator job_iterator;
struct job_record *job_ptr;
- time_t recent = now - BATCH_JOB_LAUNCH_TIME;
+ time_t recent = now - slurm_get_batch_start_timeout();
job_iterator = list_iterator_create(job_list);
while ((job_ptr = (struct job_record *) list_next(job_iterator))) {