Add squeue filtering for jobs or steps with job array ID values

doc/html/job_array.shtml

@@ -33,20 +33,6 @@ $ sbatch --array=1,3,5,7 -N1 tmp
 $ sbatch --array=1-7:2   -N1 tmp
 </pre>
 
-<p>Slurm support for job arrays at this time does not use a meta-job data
-structure, but creates a separate job record for each element of the array.
-Two additional fields were added to Slurm's job record for managing job arrays.
-The first new field is internally called "array_job_id" and is the job ID of
-the first job in the array.
-Subsequent elements of the job array will have a unique Slurm "job_id", but
-all will have the same "array_job_id" value.
-Some Slurm commands interpret the array_job_id as representing all elements of
-the job array, while other commands use the unique job_id assigned to each.
-Support for Slurm job arrays can be expected to improve in later releases.
-The second new field is called "array_task_id" which is the job array index
-value of the job array element.
-More details about these new fields follow.</p>
-
 <h2>Job ID and Environment Variables</h2>
 
 <p>Job arrays will have two additional environment variable set.
@@ -147,7 +133,17 @@ $ squeue
 1088_4         debug   tmp   mac   R  0:03      1 tux3
 </pre>
 
-<p>Two additional job output formats have been added to squeue:<br>
+<p>The squeue --step/-s and --job/-j options can accept job or step
+specifications of the same format.</p>
+
+<pre>
+$ squeue -j 1234_2,1234_3
+...
+$ squeue -s 1234_2.0,1234_3.0
+...
+</pre>
+
+<p>Two additional job output format field options have been added to squeue:<br>
 <b>%F</b> prints the array_job_id value<br>
 <b>%K</b> prints the array_task_id value<br>
 (all of the obvious letters to use were already assigned to other job fields).</p>
@@ -198,6 +194,19 @@ The default value of MaxArraySize is 1001. Be mindful about the value of
 MaxArraySize as job arrays offer an easy way for users to submit large numbers
 of jobs very quickly.</p>
 
+<p>Slurm support for job arrays at this time does not use a meta-job data
+structure, but creates a separate job record for each element of the array.
+Two additional fields were added to Slurm's job record for managing job arrays.
+The first new field is internally called "array_job_id" and is the job ID of
+the first job in the array.
+Subsequent elements of the job array will have a unique Slurm "job_id", but
+all will have the same "array_job_id" value.
+Some Slurm commands interpret the array_job_id as representing all elements of
+the job array, while other commands use the unique job_id assigned to each.
+Support for Slurm job arrays can be expected to improve in later releases.
+The second new field is called "array_task_id" which is the job array index
+value of the job array element.</p>
+
 <p style="text-align:center;">Last modified 25 January 2013</p>
 
 <!--#include virtual="footer.txt"-->

doc/man/man1/squeue.1

@@ -51,7 +51,7 @@ By default, prints a time stamp with the header.
 
 .TP
 \fB\-j <job_id_list>\fR, \fB\-\-jobs=<job_id_list>\fR
-Requests a comma separated list of job ids to display.  Defaults to all jobs.
+Requests a comma separated list of job IDs to display.  Defaults to all jobs.
 The \fB\-\-jobs=<job_id_list>\fR option may be used in conjunction with the
 \fB\-\-steps\fR option to print step information about specific jobs.
 Note: If a list of job IDs is provided, the jobs are displayed even if
@@ -59,6 +59,7 @@ they are on hidden partitions. Since this option's argument is optional,
 for proper parsing the single letter option must be followed immediately
 with the value and not include a space between them. For example "\-j1008"
 and not "\-j 1008".
+The job ID format is "job_id[_array_id]".
 Performance of the command can be measurably improved for systems with large
 numbers of jobs when a single job ID is specified.
 
@@ -373,7 +374,7 @@ Specify the reservation of the jobs to view.
 \fB\-s\fR, \fB\-\-steps\fR
 Specify the job steps to view.  This flag indicates that a comma separated list
 of job steps to view follows without an equal sign (see examples).
-The job step format is "job_id.step_id". Defaults to all job
+The job step format is "job_id[_array_id].step_id". Defaults to all job
 steps. Since this option's argument is optional, for proper parsing
 the single letter option must be followed immediately with the value
 and not include a space between them. For example "\-s1008.0" and not

src/squeue/opts.c

@@ -904,7 +904,6 @@ _print_options(void)
 	uint32_t *user;
 	enum job_states *state_id;
 	squeue_job_step_t *job_step_id;
-	uint32_t *job_id;
 	char hostlist[8192];
 
 	if (params.nodes) {
@@ -936,8 +935,15 @@ _print_options(void)
 	if ((params.verbose > 1) && params.job_list) {
 		i = 0;
 		iterator = list_iterator_create( params.job_list );
-		while ( (job_id = list_next( iterator )) ) {
-			printf( "job_list[%d] = %u\n", i++, *job_id);
+		while ( (job_step_id = list_next( iterator )) ) {
+			if (job_step_id->array_id == (uint16_t) NO_VAL) {
+				printf( "job_list[%d] = %u\n", i++,
+					job_step_id->job_id );
+			} else {
+				printf( "job_list[%d] = %u_%u\n", i++,
+					job_step_id->job_id,
+					job_step_id->array_id );
+			}
 		}
 		list_iterator_destroy( iterator );
 	}
@@ -975,8 +981,16 @@ _print_options(void)
 		i = 0;
 		iterator = list_iterator_create( params.step_list );
 		while ( (job_step_id = list_next( iterator )) ) {
-			printf( "step_list[%d] = %u.%u\n", i++,
-				job_step_id->job_id, job_step_id->step_id );
+			if (job_step_id->array_id == (uint16_t) NO_VAL) {
+				printf( "step_list[%d] = %u.%u\n", i++,
+					job_step_id->job_id,
+					job_step_id->step_id );
+			} else {
+				printf( "step_list[%d] = %u_%u.%u\n", i++,
+					job_step_id->job_id,
+					job_step_id->array_id,
+					job_step_id->step_id );
+			}
 		}
 		list_iterator_destroy( iterator );
 	}
@@ -1003,9 +1017,9 @@ static List
 _build_job_list( char* str )
 {
 	List my_list;
-	char *job = NULL, *tmp_char = NULL, *my_job_list = NULL;
-	int i;
-	uint32_t *job_id = NULL;
+	char *end_ptr, *job = NULL, *tmp_char = NULL, *my_job_list = NULL;
+	int job_id, array_id;
+	squeue_job_step_t *job_step_id;
 
 	if ( str == NULL )
 		return NULL;
@@ -1013,14 +1027,20 @@ _build_job_list( char* str )
 	my_job_list = xstrdup( str );
 	job = strtok_r( my_job_list, ",", &tmp_char );
 	while (job) {
-		i = strtol( job, (char **) NULL, 10 );
-		if (i <= 0) {
+		job_id = strtol( job, &end_ptr, 10 );
+		if (end_ptr[0] == '_')
+			array_id = strtol( end_ptr + 1, &end_ptr, 10 );
+		else
+			array_id = (uint16_t) NO_VAL;
+		if (job_id <= 0) {
 			error( "Invalid job id: %s", job );
 			exit( 1 );
 		}
-		job_id = xmalloc( sizeof( uint32_t ) );
-		*job_id = (uint32_t) i;
-		list_append( my_list, job_id );
+
+		job_step_id = xmalloc( sizeof( squeue_job_step_t ) );
+		job_step_id->job_id   = (uint32_t) job_id;
+		job_step_id->array_id = (uint16_t) array_id;
+		list_append( my_list, job_step_id );
 		job = strtok_r (NULL, ",", &tmp_char);
 	}
 	return my_list;
@@ -1113,16 +1133,16 @@ _build_all_states_list( void )
 
 /*
  * _build_step_list- build a list of job/step_ids
- * IN str - comma separated list of job_id.step_ids
+ * IN str - comma separated list of job_id[array_id].step_id values
  * RET List of job/step_ids (structure of uint32_t's)
  */
 static List
 _build_step_list( char* str )
 {
 	List my_list;
-	char *step = NULL, *tmp_char = NULL, *tmps_char = NULL;
+	char *end_ptr, *step = NULL, *tmp_char = NULL, *tmps_char = NULL;
 	char *job_name = NULL, *step_name = NULL, *my_step_list = NULL;
-	int i, j;
+	int job_id, array_id, step_id;
 	squeue_job_step_t *job_step_id = NULL;
 
 	if ( str == NULL)
@@ -1130,25 +1150,29 @@ _build_step_list( char* str )
 	my_list = list_create( NULL );
 	my_step_list = xstrdup( str );
 	step = strtok_r( my_step_list, ",", &tmp_char );
-	while (step)
-	{
+	while (step) {
 		job_name = strtok_r( step, ".", &tmps_char );
 		step_name = strtok_r( NULL, ".", &tmps_char );
-		i = strtol( job_name, (char **) NULL, 10 );
+		job_id = strtol( job_name, &end_ptr, 10 );
+		if (end_ptr[0] == '_')
+			array_id = strtol( end_ptr + 1, &end_ptr, 10 );
+		else
+			array_id = (uint16_t) NO_VAL;
 		if (step_name == NULL) {
 			error ( "Invalid job_step id: %s.??",
 				 job_name );
 			exit( 1 );
 		}
-		j = strtol( step_name, (char **) NULL, 10 );
-		if ((i <= 0) || (j < 0)) {
+		step_id = strtol( step_name, &end_ptr, 10 );
+		if ((job_id <= 0) || (step_id < 0)) {
 			error( "Invalid job_step id: %s.%s",
 				job_name, step_name );
 			exit( 1 );
 		}
 		job_step_id = xmalloc( sizeof( squeue_job_step_t ) );
-		job_step_id->job_id  = (uint32_t) i;
-		job_step_id->step_id = (uint32_t) j;
+		job_step_id->job_id   = (uint32_t) job_id;
+		job_step_id->array_id = (uint16_t) array_id;
+		job_step_id->step_id  = (uint32_t) step_id;
 		list_append( my_list, job_step_id );
 		step = strtok_r( NULL, ",", &tmp_char);
 	}

src/squeue/print.c

@@ -1474,17 +1474,19 @@ static int _filter_job(job_info_t * job)
 {
 	int filter;
 	ListIterator iterator;
-	uint32_t *job_id, *user;
+	uint32_t *user;
 	uint16_t *state_id;
 	char *account, *part, *qos, *name;
+	squeue_job_step_t *job_step_id;
 
 	if (params.job_list) {
 		filter = 1;
 		iterator = list_iterator_create(params.job_list);
-		while ((job_id = list_next(iterator))) {
-			if ((*job_id == job->job_id) ||
-			    ((job->array_task_id != (uint16_t) NO_VAL) &&
-			     (*job_id == job->array_job_id))) {
+		while ((job_step_id = list_next(iterator))) {
+			if (((job_step_id->array_id == (uint16_t) NO_VAL) &&
+			     (job_step_id->job_id   == job->array_job_id)) ||
+			    ((job_step_id->array_id == job->array_task_id) &&
+			     (job_step_id->job_id   == job->array_job_id))) {
 				filter = 0;
 				break;
 			}
@@ -1622,7 +1624,7 @@ static int _filter_step(job_step_info_t * step)
 {
 	int filter;
 	ListIterator iterator;
-	uint32_t *job_id, *user;
+	uint32_t *user;
 	char *part;
 	squeue_job_step_t *job_step_id;
 
@@ -1632,10 +1634,11 @@ static int _filter_step(job_step_info_t * step)
 	if (params.job_list) {
 		filter = 1;
 		iterator = list_iterator_create(params.job_list);
-		while ((job_id = list_next(iterator))) {
-			if ((*job_id == step->job_id) ||
-			    ((step->array_task_id != (uint16_t) NO_VAL) &&
-			     (*job_id == step->array_job_id))) {
+		while ((job_step_id = list_next(iterator))) {
+			if (((job_step_id->array_id == (uint16_t) NO_VAL) &&
+			     (job_step_id->job_id   == step->array_job_id)) ||
+			    ((job_step_id->array_id == step->array_task_id) &&
+			     (job_step_id->job_id   == step->array_job_id))) {
 				filter = 0;
 				break;
 			}
@@ -1663,8 +1666,12 @@ static int _filter_step(job_step_info_t * step)
 		filter = 1;
 		iterator = list_iterator_create(params.step_list);
 		while ((job_step_id = list_next(iterator))) {
-			if ((job_step_id->job_id == step->job_id) &&
-			    (job_step_id->step_id == step->step_id)) {
+			if (job_step_id->step_id != step->step_id)
+				continue;
+			if (((job_step_id->array_id == (uint16_t) NO_VAL) &&
+			     (job_step_id->job_id   == step->array_job_id)) ||
+			    ((job_step_id->array_id == step->array_task_id) &&
+			     (job_step_id->job_id   == step->array_job_id))) {
 				filter = 0;
 				break;
 			}

src/squeue/squeue.h

@@ -70,11 +70,11 @@
 #include "src/common/slurmdb_defs.h"
 #include "src/squeue/print.h"
 
-struct job_step {
+typedef struct job_step {
 	uint32_t job_id;
+	uint16_t array_id;
 	uint32_t step_id;
-};
-typedef struct job_step squeue_job_step_t;
+} squeue_job_step_t;
 
 struct squeue_parameters {
 	bool all_flag;