diff --git a/NEWS b/NEWS
index e1bf5638181ad40d62f556091242396f2626365a..c2102e368da10a02f4580f2580372e8c2f24b425 100644
--- a/NEWS
+++ b/NEWS
@@ -12,6 +12,7 @@ documents those changes that are of interest to users and admins.
-- Don't strip binaries, breaks parallel debuggers
-- Fix bug in Munge authentication retry logic
-- Change srun handling of interupts to work properly with TotalView
+ -- Added "reason" field to job info showing why a job is waiting to run
* Changes in SLURM 0.3.7
========================
diff --git a/doc/man/man1/squeue.1 b/doc/man/man1/squeue.1
index 75d203dca93f5dfbb36e27889e98eb6d6564c17f..8a6052bfcf1b645a3a5fa558341e59391b047b37 100644
--- a/doc/man/man1/squeue.1
+++ b/doc/man/man1/squeue.1
@@ -55,10 +55,10 @@ various options are
.RS
.TP 15
.I "default"
-"%.7i %.9P %.8j %.8u %.2t %.9M %.6D %N"
+"%.7i %.9P %.8j %.8u %.2t %.9M %.6D %R"
.TP
.I "-l, --long"
-".7i %.9P %.8j %.8u %.8T %.9M %.9l %.6D %N"
+".7i %.9P %.8j %.8u %.8T %.9M %.9l %.6D %R"
.TP
.I "-s, --steps"
"%10i %.9P %.8u %.9M %N"
@@ -147,17 +147,30 @@ Priority of the job (converted to a floating point number between 0.0 and 1.0
\fB%P\fR
Partition of the job or job step
.TP
+\fB%r\fR
+The reason a job is waiting for execution.
+See the \fBJOB REASON CODES\fR section below for more information.
+.TP
+\fB%R\fR
+For running or completed jobs: the list of allocate nodes.
+For pending jobs: the reason a job is waiting for execution
+is printed within parenthesis.
+See the \fBJOB REASON CODES\fR section below for more information.
+.TP
\fB%S\fR
Start time of the job or job step
.TP
\fB%t\fR
Job state, compact form:
-PD (pending), R (running), CG (completing), CD (completed), F (failed),
-TO (timeout), and NF (node failure).
+PD (pending), R (running), CA (cancelled), CG (completing), CD (completed),
+F (failed), TO (timeout), and NF (node failure).
+See the \fBJOB STATE CODES\fR section below for more information.
.TP
\fB%T\fR
Job state, extended form:
-PENDING, RUNNING, CANCELLED, COMPLETING, COMPLETED, FAILED, TIMEOUT, and NODE_FAIL.
+PENDING, RUNNING, CANCELLED, COMPLETING, COMPLETED, FAILED, TIMEOUT,
+and NODE_FAIL.
+See the \fBJOB STATE CODES\fR section below for more information.
.TP
\fB%u\fR
User name
@@ -210,6 +223,7 @@ jobs are reported. Valid states (in both extended and compact form) include:
PENDING (PD), RUNNING (R), COMPLETING (CG), COMPLETED (CD), CANCELLED (CA),
FAILED (F), TIMEOUT (TO), and NODE_FAIL (NF). Note the \fB<state_list>\fR
supplied is case insensitve ("pd" and "PD" work the same).
+See the \fBJOB STATE CODES\fR section below for more information.
.TP
\fB\-u <user_list>\fR, \fB\-\-user=<user_list>\fR
Specifies a comma separated list of users whose jobs or job steps are to be
@@ -221,8 +235,39 @@ Report details of squeues actions.
\fB\-V\fR , \fB\-\-version\fR
Print version information and exit.
+.SH "JOB REASON CODES"
+These codes identify the reason that a job is waiting for execution.
+A job may be waiting for more than one reason, in which case only
+one of those reasons is displayed.
+.TP 20
+Dependency
+This job is waiting for a dependent job to complete.
+.TP
+None
+No reason is set for this job.
+.TP
+PartitionDown
+The partition required by this job is in a DOWN state.
+.TP
+PartitionNodeLimit
+The number of nodes required by this job is outside of it's
+partitions current limits.
+.TP
+PartitionTimeLimit
+The job's time limit exceeds it's partition's current time limit.
+.TP
+Priority
+One or more higher priority jobs exist for this partition.
+.TP
+Resources
+The job is waiting for resources to become availble.
+
.SH "JOB STATE CODES"
-.TP 17
+Jobs typically pass through several states in the course of their
+execution.
+The typical states are PENDING, RUNNING, COMPLETING, and COMPLETED.
+An explanation of each state follows.
+.TP 20
CA CANCELLED
Job was explicitly cancelled by the user or system administrator.
The job may or may not have been initiated.
@@ -242,6 +287,9 @@ Job terminated due to failure of one or more allocated nodes.
PD PENDING
Job is awaiting resource allocation.
.TP
+R RUNNING
+Job currently has an allocation.
+.TP
TO TIMEOUT
Job terminated upon reaching its time limit.
@@ -291,7 +339,7 @@ Print the job steps in the debug partition sorted by user:
.br
# squeue -s -p debug -S u
.br
- STEPID PARTITION USER TIME_USED NODELIST
+ STEPID PARTITION USER TIME_USED NODELIST(REASON)
.br
65552.1 debug alice 0:23 dev[1-4]
.br
@@ -301,15 +349,17 @@ Print the job steps in the debug partition sorted by user:
.ec
.eo
-Print information only about jobs 12345 and 12346:
+Print information only about jobs 12345,12345, and 12348:
.br
-# squeue --jobs 12345,12346
+# squeue --jobs 12345,12346,12348
.br
- JOBID PARTITION NAME USER ST TIME_USED NODES NODELIST
+ JOBID PARTITION NAME USER ST TIME_USED NODES NODELIST(REASON)
.br
12345 debug job1 dave R 0:21 4 dev[9-12]
.br
- 12346 debug job2 dave PD 0:00 8
+ 12346 debug job2 dave PD 0:00 8 (Resources)
+.br
+ 12348 debug job3 ed PD 0:00 4 (Priority)
.ec
.eo
@@ -317,7 +367,7 @@ Print information only about job step 65552.1:
.br
# squeue --steps 65552.1
.br
- STEPID PARTITION USER TIME_USED NODELIST
+ STEPID PARTITION USER TIME_USED NODELIST(REASON)
.br
65552.1 debug alice 12:49 dev[1-4]
.ec
diff --git a/slurm/slurm.h.in b/slurm/slurm.h.in
index a1016b445814069bdd3da516f401bfb759bd8369..4a33cad7da757946364cab841498399153fe92f0 100644
--- a/slurm/slurm.h.in
+++ b/slurm/slurm.h.in
@@ -133,6 +133,18 @@ enum job_states {
};
#define JOB_COMPLETING (0x8000)
+/* Reason for job to be pending rather than executing. If multiple reasons
+ * exists, only one is given for the sake of system efficiency */
+enum job_wait_reason {
+ WAIT_NO_REASON = 0, /* not set or job not pending */
+ WAIT_PRIORITY, /* higher priority jobs exist */
+ WAIT_DEPENDENCY, /* depedent job has not completed */
+ WAIT_RESOUCES, /* required resources not available */
+ WAIT_PART_NODE_LIMIT, /* request exceeds partition node limit */
+ WAIT_PART_TIME_LIMIT, /* request exceeds partition time limit */
+ WAIT_PART_STATE /* requested partition is down */
+};
+
/* Possible task distributions across the nodes */
enum task_dist_states {
SLURM_DIST_CYCLIC, /* distribute tasks 1 per node, round robin */
@@ -279,6 +291,8 @@ typedef struct job_info {
char *features; /* comma separated list of required features */
uint32_t dependency; /* defer until specified job completes */
char *account; /* charge to specified account */
+ uint16_t wait_reason; /* reason job still pending, see
+ * slurm.h:enum job_wait_reason */
} job_info_t;
typedef struct job_info_msg {
diff --git a/src/api/job_info.c b/src/api/job_info.c
index a85fba2dfcb5f6558c74b4ef11ca7a159a17ccc7..8cd716f72a61188123291572572f62c462a8b8f5 100644
--- a/src/api/job_info.c
+++ b/src/api/job_info.c
@@ -182,8 +182,9 @@ slurm_print_job_info ( FILE* out, job_info_t * job_ptr, int one_liner )
fprintf ( out, "\n ");
/****** Line 9 ******/
- fprintf ( out, "Dependency=%u Account=%s",
- job_ptr->dependency, job_ptr->account);
+ fprintf ( out, "Dependency=%u Account=%s Reason=%s",
+ job_ptr->dependency, job_ptr->account,
+ job_reason_string(job_ptr->wait_reason));
if (one_liner)
fprintf ( out, " ");
else
diff --git a/src/common/slurm_protocol_defs.c b/src/common/slurm_protocol_defs.c
index 58b6dfb447009a2e318d6b0d5cea71766caf732b..ed35a4bb67ab503f26b5520702149338c22a76e2 100644
--- a/src/common/slurm_protocol_defs.c
+++ b/src/common/slurm_protocol_defs.c
@@ -409,6 +409,21 @@ void inline slurm_free_checkpoint_resp_msg(checkpoint_resp_msg_t *msg)
}
}
+/* Given a job's reason for waiting, return a descriptive string */
+extern char *job_reason_string(enum job_wait_reason inx)
+{
+ static char *job_reason_string[] = {
+ "None", /* WAIT_NO_REASON */
+ "Priority",
+ "Dependency",
+ "Resources",
+ "PartitionNodeLimit",
+ "PartitionTimeLimit",
+ "PartitionDown"
+ };
+ return job_reason_string[inx];
+}
+
char *job_state_string(enum job_states inx)
{
static char *job_state_string[] = {
diff --git a/src/common/slurm_protocol_defs.h b/src/common/slurm_protocol_defs.h
index 27dfb6c4c5b0e425887ed8e26a39a3f2479ad981..ab0e6e7a122e28e0b8073a8322183b71f24ddb18 100644
--- a/src/common/slurm_protocol_defs.h
+++ b/src/common/slurm_protocol_defs.h
@@ -501,6 +501,7 @@ void slurm_free_node_info_msg(node_info_msg_t * msg);
void slurm_free_partition_info_msg(partition_info_msg_t * msg);
extern char *job_dist_string(uint16_t inx);
+extern char *job_reason_string(enum job_wait_reason inx);
extern char *job_state_string(enum job_states inx);
extern char *job_state_string_compact(enum job_states inx);
extern char *node_state_string(enum node_states inx);
diff --git a/src/common/slurm_protocol_pack.c b/src/common/slurm_protocol_pack.c
index 61ae7d4d4a3215544aefd9b845b0addd5b61724c..f63d670c8c611b4e3f9fe77411e28126b6d98499 100644
--- a/src/common/slurm_protocol_pack.c
+++ b/src/common/slurm_protocol_pack.c
@@ -1640,6 +1640,7 @@ _unpack_job_info_members(job_info_t * job, Buf buffer)
safe_unpack32(&job->min_procs, buffer);
safe_unpack32(&job->min_memory, buffer);
safe_unpack32(&job->min_tmp_disk, buffer);
+ safe_unpack16(&job->wait_reason, buffer);
safe_unpackstr_xmalloc(&job->req_nodes, &uint16_tmp, buffer);
safe_unpackstr_xmalloc(&node_inx_str, &uint16_tmp, buffer);
diff --git a/src/slurmctld/job_mgr.c b/src/slurmctld/job_mgr.c
index 233844a63aff193fd76bef2228a4713423294a48..de52c7956ebe328e1e7de42578a2b73aa80ae9ba 100644
--- a/src/slurmctld/job_mgr.c
+++ b/src/slurmctld/job_mgr.c
@@ -1178,9 +1178,11 @@ void rehash_jobs(void)
* OUT node_list - list of nodes allocated to the job
* OUT node_cnt - number of allocated nodes
* OUT node_addr - slurm_addr's for the allocated nodes
- * RET 0 or an error code
+ * RET 0 or an error code. If the job would only be able to execute with
+ * some change in partition configuration then
+ * ESLURM_REQUESTED_PART_CONFIG_UNAVAILABLE is returned
* NOTE: If allocating nodes lx[0-7] to a job and those nodes have cpu counts
- * of 4, 4, 4, 4, 8, 8, 4, 4 then num_cpu_groups=3, cpus_per_node={4,8,4}
+ * of 4, 4, 4, 4, 8, 8, 4, 4 then num_cpu_groups=3, cpus_per_node={4,8,4}
* and cpu_count_reps={4,2,2}
* globals: job_list - pointer to global job list
* list_part - global list of partition info
@@ -1500,7 +1502,9 @@ job_complete(uint32_t job_id, uid_t uid, bool requeue,
* IN will_run - job is not to be created, test of validity only
* OUT new_job_id - the job's ID
* OUT job_pptr - pointer to the job (NULL on error)
- * RET 0 on success, otherwise ESLURM error code
+ * RET 0 on success, otherwise ESLURM error code. If the job would only be
+ * able to execute with some change in partition configuration then
+ * ESLURM_REQUESTED_PART_CONFIG_UNAVAILABLE is returned
* globals: job_list - pointer to global job list
* list_part - global list of partition info
* default_part_loc - pointer to default partition
@@ -1512,6 +1516,8 @@ static int _job_create(job_desc_msg_t * job_desc, uint32_t * new_job_id,
struct job_record **job_pptr, uid_t submit_uid)
{
int error_code = SLURM_SUCCESS, i;
+ struct job_details *detail_ptr;
+ enum job_wait_reason fail_reason;
struct part_record *part_ptr;
bitstr_t *req_bitmap = NULL, *exc_bitmap = NULL;
bool super_user = false;
@@ -1669,6 +1675,8 @@ static int _job_create(job_desc_msg_t * job_desc, uint32_t * new_job_id,
/* Insure that requested partition is valid right now,
* otherwise leave job queued and provide warning code */
+ detail_ptr = (*job_pptr)->details;
+ fail_reason= WAIT_NO_REASON;
if ((job_desc->user_id == 0) ||
(job_desc->user_id == slurmctld_conf.slurm_user_id))
super_user = true;
@@ -1677,21 +1685,25 @@ static int _job_create(job_desc_msg_t * job_desc, uint32_t * new_job_id,
info("Job %u requested too many nodes (%d) of "
"partition %s(%d)", *new_job_id, job_desc->min_nodes,
part_ptr->name, part_ptr->max_nodes);
- error_code = ESLURM_REQUESTED_PART_CONFIG_UNAVAILABLE;
+ fail_reason = WAIT_PART_NODE_LIMIT;
} else if ((!super_user) &&
(job_desc->max_nodes != 0) && /* no max_nodes for job */
(job_desc->max_nodes < part_ptr->min_nodes)) {
info("Job %u requested too few nodes (%d) of partition %s(%d)",
*new_job_id, job_desc->max_nodes,
part_ptr->name, part_ptr->min_nodes);
- error_code = ESLURM_REQUESTED_PART_CONFIG_UNAVAILABLE;
+ fail_reason = WAIT_PART_NODE_LIMIT;
} else if (part_ptr->state_up == 0) {
info("Job %u requested down partition %s",
*new_job_id, part_ptr->name);
+ fail_reason = WAIT_PART_STATE;
+ }
+ if (fail_reason != WAIT_NO_REASON) {
error_code = ESLURM_REQUESTED_PART_CONFIG_UNAVAILABLE;
+ (*job_pptr)->priority = 1; /* Move to end of queue */
+ if (detail_ptr)
+ detail_ptr->wait_reason = fail_reason;
}
- if (error_code == ESLURM_REQUESTED_PART_CONFIG_UNAVAILABLE)
- (*job_pptr)->priority = 1; /* Move to end of queue */
cleanup:
FREE_NULL_BITMAP(req_bitmap);
@@ -2494,6 +2506,7 @@ static void _pack_job_details(struct job_details *detail_ptr, Buf buffer)
pack32((uint32_t) detail_ptr->min_procs, buffer);
pack32((uint32_t) detail_ptr->min_memory, buffer);
pack32((uint32_t) detail_ptr->min_tmp_disk, buffer);
+ pack16((uint16_t) detail_ptr->wait_reason, buffer);
packstr(detail_ptr->req_nodes, buffer);
pack_bit_fmt(detail_ptr->req_node_bitmap, buffer);
@@ -2511,6 +2524,7 @@ static void _pack_job_details(struct job_details *detail_ptr, Buf buffer)
pack32((uint32_t) 0, buffer);
pack32((uint32_t) 0, buffer);
pack32((uint32_t) 0, buffer);
+ pack16((uint16_t) 0, buffer);
packstr(NULL, buffer);
packstr(NULL, buffer);
@@ -2748,30 +2762,38 @@ void reset_job_priority(void)
*/
static bool _top_priority(struct job_record *job_ptr)
{
- ListIterator job_iterator;
- struct job_record *job_ptr2;
+ struct job_details *detail_ptr = job_ptr->details;
bool top;
- if (job_ptr->priority == 0) /* held */
- return false;
-
- top = true; /* assume top priority until found otherwise */
- job_iterator = list_iterator_create(job_list);
- while ((job_ptr2 = (struct job_record *) list_next(job_iterator))) {
- if (job_ptr2 == job_ptr)
- continue;
- if (job_ptr2->job_state != JOB_PENDING)
- continue;
- if (!job_independent(job_ptr2))
- continue;
- if ((job_ptr2->priority > job_ptr->priority) &&
- (job_ptr2->part_ptr == job_ptr->part_ptr)) {
- top = false;
- break;
+ if (job_ptr->priority == 0) /* user held */
+ top = false;
+ else {
+ ListIterator job_iterator;
+ struct job_record *job_ptr2;
+
+ top = true; /* assume top priority until found otherwise */
+ job_iterator = list_iterator_create(job_list);
+ while ((job_ptr2 = (struct job_record *)
+ list_next(job_iterator))) {
+ if (job_ptr2 == job_ptr)
+ continue;
+ if (job_ptr2->job_state != JOB_PENDING)
+ continue;
+ if (!job_independent(job_ptr2))
+ continue;
+ if ((job_ptr2->priority > job_ptr->priority) &&
+ (job_ptr2->part_ptr == job_ptr->part_ptr)) {
+ top = false;
+ break;
+ }
}
+ list_iterator_destroy(job_iterator);
}
- list_iterator_destroy(job_iterator);
+ if ((!top) && /* not top prio and */
+ (job_ptr->priority != 1) && /* not system hold */
+ (detail_ptr))
+ detail_ptr->wait_reason = WAIT_PRIORITY;
return top;
}
@@ -3458,6 +3480,7 @@ extern void job_completion_logger(struct job_record *job_ptr)
extern bool job_independent(struct job_record *job_ptr)
{
struct job_record *dep_ptr;
+ struct job_details *detail_ptr = job_ptr->details;
if (job_ptr->dependency == 0)
return true;
@@ -3469,6 +3492,9 @@ extern bool job_independent(struct job_record *job_ptr)
if (((dep_ptr->job_state & JOB_COMPLETING) == 0) &&
(dep_ptr->job_state >= JOB_COMPLETE))
return true;
+
+ if (detail_ptr)
+ detail_ptr->wait_reason = WAIT_DEPENDENCY;
return false; /* job exists and incomplete */
}
diff --git a/src/slurmctld/node_scheduler.c b/src/slurmctld/node_scheduler.c
index 91aeaf65325c56a8dbe29be603efc328c77475e8..d2627a3eedcaad7db7405b8c9cf878fdbc515c7a 100644
--- a/src/slurmctld/node_scheduler.c
+++ b/src/slurmctld/node_scheduler.c
@@ -852,6 +852,7 @@ int select_nodes(struct job_record *job_ptr, bool test_only)
{
int error_code = SLURM_SUCCESS, i, shared, node_set_size = 0;
bitstr_t *req_bitmap = NULL;
+ struct job_details *detail_ptr = job_ptr->details;
struct node_set *node_set_ptr = NULL;
struct part_record *part_ptr = job_ptr->part_ptr;
uint32_t min_nodes, max_nodes, part_node_limit;
@@ -872,16 +873,24 @@ int select_nodes(struct job_record *job_ptr, bool test_only)
/* Confirm that partition is up and has compatible nodes limits */
if ((job_ptr->user_id == 0) || (job_ptr->user_id == getuid()))
super_user = true;
- else if (
- (part_ptr->state_up == 0) ||
- ((job_ptr->time_limit != NO_VAL) &&
- (job_ptr->time_limit > part_ptr->max_time)) ||
- ((job_ptr->details->max_nodes != 0) && /* no node limit */
- (job_ptr->details->max_nodes < part_ptr->min_nodes)) ||
- (job_ptr->details->min_nodes > part_ptr->max_nodes)) {
- job_ptr->priority = 1; /* move to end of queue */
- last_job_update = time(NULL);
- return ESLURM_REQUESTED_PART_CONFIG_UNAVAILABLE;
+ else {
+ enum job_wait_reason fail_reason = WAIT_NO_REASON;
+ if (part_ptr->state_up == 0)
+ fail_reason = WAIT_PART_STATE;
+ else if ((job_ptr->time_limit != NO_VAL) &&
+ (job_ptr->time_limit > part_ptr->max_time))
+ fail_reason = WAIT_PART_TIME_LIMIT;
+ else if (((job_ptr->details->max_nodes != 0) &&
+ (job_ptr->details->max_nodes < part_ptr->min_nodes)) ||
+ (job_ptr->details->min_nodes > part_ptr->max_nodes))
+ fail_reason = WAIT_PART_NODE_LIMIT;
+ if (fail_reason != WAIT_NO_REASON) {
+ if (detail_ptr)
+ detail_ptr->wait_reason = fail_reason;
+ job_ptr->priority = 1; /* sys hold, move to end of queue */
+ last_job_update = time(NULL);
+ return ESLURM_REQUESTED_PART_CONFIG_UNAVAILABLE;
+ }
}
/* build sets of usable nodes based upon their configuration */
@@ -932,7 +941,11 @@ int select_nodes(struct job_record *job_ptr, bool test_only)
job_ptr->details->contiguous,
shared, part_node_limit);
if (error_code) {
+ if (detail_ptr)
+ detail_ptr->wait_reason = WAIT_RESOUCES;
if (error_code == ESLURM_REQUESTED_PART_CONFIG_UNAVAILABLE) {
+ /* Required nodes are down or
+ * too many nodes requested */
debug3("JobId=%u not runnable with present config",
job_ptr->job_id);
job_ptr->priority = 1; /* Move to end of queue */
@@ -948,6 +961,8 @@ int select_nodes(struct job_record *job_ptr, bool test_only)
}
/* assign the nodes and stage_in the job */
+ if (detail_ptr)
+ detail_ptr->wait_reason = WAIT_NO_REASON;
job_ptr->nodes = bitmap2node_name(req_bitmap);
job_ptr->node_bitmap = req_bitmap;
job_ptr->details->shared = shared;
diff --git a/src/slurmctld/slurmctld.h b/src/slurmctld/slurmctld.h
index fc2b04282635fc357ed814b99b8ebff9dc58efef..4947687bb73ae37385f3248893f06f40d55b8470 100644
--- a/src/slurmctld/slurmctld.h
+++ b/src/slurmctld/slurmctld.h
@@ -244,6 +244,8 @@ struct job_details {
uint16_t req_tasks; /* required number of tasks */
uint16_t shared; /* set node can be shared*/
uint16_t contiguous; /* set if requires contiguous nodes */
+ uint16_t wait_reason; /* reason job still pending, see
+ * slurm.h:enum job_wait_reason */
uint32_t min_procs; /* minimum processors per node */
uint32_t min_memory; /* minimum memory per node, MB */
uint32_t min_tmp_disk; /* minimum tempdisk per node, MB */
@@ -610,9 +612,11 @@ extern bool is_node_resp (char *name);
* OUT node_list - list of nodes allocated to the job
* OUT node_cnt - number of allocated nodes
* OUT node_addr - slurm_addr's for the allocated nodes
- * RET 0 or an error code
+ * RET 0 or an error code. If the job would only be able to execute with
+ * some change in partition configuration then
+ * ESLURM_REQUESTED_PART_CONFIG_UNAVAILABLE is returned
* NOTE: If allocating nodes lx[0-7] to a job and those nodes have cpu counts
- * of 4, 4, 4, 4, 8, 8, 4, 4 then num_cpu_groups=3, cpus_per_node={4,8,4}
+ * of 4, 4, 4, 4, 8, 8, 4, 4 then num_cpu_groups=3, cpus_per_node={4,8,4}
* and cpu_count_reps={4,2,2}
* globals: job_list - pointer to global job list
* list_part - global list of partition info
diff --git a/src/squeue/opts.c b/src/squeue/opts.c
index 8459d4b101123cba4dddfb011d03774f79f1d08c..e8c0b3a7771bb9c449e32e4c21181d022ff91504 100644
--- a/src/squeue/opts.c
+++ b/src/squeue/opts.c
@@ -482,6 +482,16 @@ extern int parse_format( char* format )
field_size,
right_justify,
suffix );
+ else if (field[0] == 'r')
+ job_format_add_reason( params.format_list,
+ field_size,
+ right_justify,
+ suffix );
+ else if (field[0] == 'R')
+ job_format_add_reason_list( params.format_list,
+ field_size,
+ right_justify,
+ suffix );
else if (field[0] == 'S')
job_format_add_time_start( params.format_list,
field_size,
diff --git a/src/squeue/print.c b/src/squeue/print.c
index cd139f5b104d5d92729a01d5c960b80695854c9a..382277eeccd9207c083737bb860bb9f52fdb6bc3 100644
--- a/src/squeue/print.c
+++ b/src/squeue/print.c
@@ -311,10 +311,10 @@ int _print_job_job_id(job_info_t * job, int width, bool right, char* suffix)
int _print_job_partition(job_info_t * job, int width, bool right, char* suffix)
{
- char id[FORMAT_STRING_SIZE];
if (job == NULL) /* Print the Header instead */
_print_str("PARTITION", width, right, true);
else {
+ char id[FORMAT_STRING_SIZE];
snprintf(id, FORMAT_STRING_SIZE, "%s", job->partition);
_print_str(id, width, right, true);
}
@@ -330,6 +330,21 @@ int _print_job_prefix(job_info_t * job, int width, bool right, char* suffix)
return SLURM_SUCCESS;
}
+int _print_job_reason(job_info_t * job, int width, bool right, char* suffix)
+{
+ if (job == NULL) /* Print the Header instead */
+ _print_str("REASON", width, right, true);
+ else {
+ char id[FORMAT_STRING_SIZE];
+ snprintf(id, FORMAT_STRING_SIZE, "%s",
+ job_reason_string(job->wait_reason));
+ _print_str(id, width, right, true);
+ }
+ if (suffix)
+ printf("%s", suffix);
+ return SLURM_SUCCESS;
+}
+
int _print_job_name(job_info_t * job, int width, bool right, char* suffix)
{
if (job == NULL) /* Print the Header instead */
@@ -510,7 +525,7 @@ int _print_job_priority(job_info_t * job, int width, bool right, char* suffix)
int _print_job_nodes(job_info_t * job, int width, bool right, char* suffix)
{
- if (job == NULL) /* Print the Header instead */
+ if (job == NULL) /* Print the Header instead */
_print_str("NODELIST", width, right, false);
else
_print_nodes(job->nodes, width, right, false);
@@ -520,6 +535,24 @@ int _print_job_nodes(job_info_t * job, int width, bool right, char* suffix)
return SLURM_SUCCESS;
}
+int _print_job_reason_list(job_info_t * job, int width, bool right,
+ char* suffix)
+{
+ if (job == NULL) /* Print the Header instead */
+ _print_str("NODELIST(REASON)", width, right, false);
+ else if (job->job_state == JOB_PENDING) {
+ char id[FORMAT_STRING_SIZE];
+ snprintf(id, FORMAT_STRING_SIZE, "(%s)",
+ job_reason_string(job->wait_reason));
+ _print_str(id, width, right, true);
+ } else
+ _print_nodes(job->nodes, width, right, false);
+
+ if (suffix)
+ printf("%s", suffix);
+ return SLURM_SUCCESS;
+}
+
int _print_job_node_inx(job_info_t * job, int width, bool right, char* suffix)
{
if (job == NULL) /* Print the Header instead */
diff --git a/src/squeue/print.h b/src/squeue/print.h
index 44532ae6741982c565a96fc2f7d5bc6ee285abb3..68aea253eea63726e3efbb70305377c9b8da5c50 100644
--- a/src/squeue/print.h
+++ b/src/squeue/print.h
@@ -74,6 +74,10 @@ int job_format_add_function(List list, int width, bool right_justify,
job_format_add_function(list,wid,right,suffix,_print_job_partition)
#define job_format_add_prefix(list,wid,right,prefix) \
job_format_add_function(list,0,0,prefix,_print_job_prefix)
+#define job_format_add_reason(list,wid,right,prefix) \
+ job_format_add_function(list,0,0,prefix,_print_job_reason)
+#define job_format_add_reason_list(list,wid,right,prefix) \
+ job_format_add_function(list,0,0,prefix,_print_job_reason_list)
#define job_format_add_name(list,wid,right,suffix) \
job_format_add_function(list,wid,right,suffix,_print_job_name)
#define job_format_add_user_name(list,wid,right,suffix) \
@@ -142,6 +146,10 @@ int _print_job_partition(job_info_t * job, int width, bool right_justify,
char* suffix);
int _print_job_prefix(job_info_t * job, int width, bool right_justify,
char* suffix);
+int _print_job_reason(job_info_t * job, int width, bool right_justify,
+ char* suffix);
+int _print_job_reason_list(job_info_t * job, int width, bool right_justify,
+ char* suffix);
int _print_job_name(job_info_t * job, int width, bool right_justify,
char* suffix);
int _print_job_user_id(job_info_t * job, int width, bool right_justify,
diff --git a/src/squeue/squeue.c b/src/squeue/squeue.c
index 790906f1c754a5cf1b8a6e0d073c9cb548c38ea4..e840e7a642a8cd67b2eb117bc80e2733f2c07521 100644
--- a/src/squeue/squeue.c
+++ b/src/squeue/squeue.c
@@ -145,9 +145,9 @@ _print_job ( void )
if (params.format == NULL) {
if (params.long_list)
params.format = "%.7i %.9P %.8j %.8u %.8T %.10M %.9l "
- "%.6D %N";
+ "%.6D %R";
else
- params.format = "%.7i %.9P %.8j %.8u %.2t %.10M %.6D %N";
+ params.format = "%.7i %.9P %.8j %.8u %.2t %.10M %.6D %R";
}
if (params.format_list == NULL)
parse_format(params.format);