1Slurm(3) User Contributed Perl Documentation Slurm(3)
2
6 Slurm - Perl API for libslurm
7
9 use Slurm;
10 my $slurm = Slurm::new();
12 $nodes = $slurm->load_node();
13 unless($nodes) {
14 die "failed to load node info: " . $slurm->strerror();
15 }
16
18 The Slurm class provides Perl interface of the SLURM API functions in
19 "<slurm/slurm.h>", with some extra frequently used functions exported
20 by libslurm.
21 METHODS
23 To use the API, first create a Slurm object:
24 $slurm = Slurm::new($conf);
26 Then call the desired functions:
28 $resp = $slurm->load_jobs();
30 In the following "METHODS" section, if a parameter is omitted, it will
32 be listed as "param=val" , where "val" is the default value of the
33 parameter.
34 DATA STRUCTURES
36 Typically, C structures are converted to (maybe blessed) Perl hash
37 references, with field names as hash keys. Arrays in C are converted to
38 arrays in Perl. For example, there is a structure "job_info_msg_t":
39 typedef struct job_info_msg {
41 time_t last_update; /* time of latest info */
42 uint32_t record_count; /* number of records */
43 job_info_t *job_array; /* the job records */
44 } job_info_msg_t;
45 This will be converted to a hash reference with the following
47 structure:
48 {
50 last_update => 1285847672,
51 job_array => [ {account => 'test', alloc_node => 'ln0', alloc_sid => 1234, ...},
52 {account => 'debug', alloc_node => 'ln2', alloc_sid => 5678, ...},
53 ...
54 ]
55 }
56 Note the missing of the "record_count" field in the hash. It can be
58 derived from the number of elements in array "job_array".
59 To pass parameters to the API functions, use the corresponding hash
61 references, for example:
62 $rc = $slurm->update_node({node_names => 'node[0-7]', node_state => NODE_STATE_DRAIN});
64 Please see "<slurm/slurm.h>" for the definition of the structures.
66 CONSTANTS
68 The enumerations and macro definitions are available in the Slurm
69 package. If ':constant' is given when using the Slurm package, the
70 constants will be exported to the calling package.
71 Please see Slurm::Constant for the available constants.
73
75 CONSTRUCTOR/DESTRUCTOR
76 $slurm = Slurm::new($conf_file=undef);
77 Create a Slurm object. For now the object is just a hash reference with
79 no members.
80 · IN $conf_file: the SLURM configuration file. If omitted, the default
82 SLURM configuration file will be used (file specified by environment
83 variable SLURM_CONF or the file slurm.conf under directroy specified
84 in compile time).
85 · RET: blessed opaque Slurm object. On error "undef" is returned.
87 ERROR INFORMATION FUNCTIONS
89 $errno = $slurm->get_errno();
90 Get the error number associated with last operation.
92 · RET: error number associated with last operation.
94 $str = $slurm->strerror($errno=0)
96 Get the string describing the specified error number.
98 · IN $errno: error number. If omitted or 0, the error number returned
100 by "$slurm-"get_errno()> will be used.
101 · RET: error string.
103 ENTITY STATE/REASON/FLAG/TYPE STRING FUNCTIONS
105 $str = $slurm->preempt_mode_string($mode_num);
106 Get the string describing the specified preemt mode number.
108 · IN $mode_num: preempt mode number.
110 · RET: preempt mode string.
112 $num = $slurm->preempt_mode_num($mode_str);
114 Get the preempt mode number of the specified preempt mode string.
116 · IN $mode_str: preempt mode string.
118 · RET: preempt mode number.
120 $str = $slurm->job_reason_string($num);
122 Get the string representation of the specified job state reason number.
124 · IN $num: job reason number.
126 · RET: job reason string.
128 $str = $slurm->job_state_string($num);
130 Get the string representation of the specified job state number.
132 · IN $num: job state number.
134 · RET: job state string.
136 $str = $slurm->job_state_string_compact($num);
138 Get the compact string representation of the specified job state
140 number.
141 · IN $num: job state number.
143 · RET: compact job state string.
145 $num = $slurm->job_state_num($str);
147 Get the job state number of the specified (compact) job state string.
149 · IN $str: job state string.
151 · RET: job state number.
153 $str = $slurm->reservation_flags_string($flags);
155 Get the string representation of the specified reservation flags.
157 · IN $num: reservation flags number.
159 · RET: reservation flags string.
161 $str = $slurm->node_state_string($num);
163 Get the string representation of the specified node state number.
165 · IN $num: node state number.
167 · RET: node state string.
169 $str = $slurm->node_state_string_compact($num);
171 Get the compact string representation of the specified node state
173 number.
174 · IN $num: node state number.
176 · RET: compact node state string.
178 $str = $slurm->private_data_string($num);
180 Get the string representation of the specified private data type.
182 · IN $num: private data type number.
184 · RET: private data type string.
186 $str = $slurm->accounting_enforce_string($num);
188 Get the string representation of the specified accounting enforce type.
190 · IN $num: accounting enforce type number.
192 · RET: accounting enforce type string.
194 $str = $slurm->node_use_string($num);
196 Get the string representation of the specified node usage type.
198 · IN $num: node usage type number.
200 · RET: node usage type string.
202 RESOURCE ALLOCATION FUNCTIONS
204 $resp = $slurm->allocate_resources($job_desc);
205 Allocate resources for a job request. If the requested resources are
207 not immediately available, the slurmctld will send the
208 job_alloc_resp_msg to the sepecified node and port.
209 · IN $job_desc: description of resource allocation request, with
211 structure of "job_desc_msg_t".
212 · RET: response to request, with structure of
214 "resource_allocation_response_msg_t". This only represents a job
215 allocation if resources are immediately available. Otherwise it just
216 contains the job id of the enqueued job request. On failure "undef"
217 is returned.
218 $resp = $slurm->allocate_resources_blocking($job_desc, $timeout=0,
220 $pending_callbacks=undef);
221 Allocate resources for a job request. This call will block until the
223 allocation is granted, or the specified timeout limit is reached.
224 · IN $job_desc: description of resource allocation request, with
226 structure of "job_desc_msg_t".
227 · IN $timeout: amount of time, in seconds, to wait for a response
229 before giving up. A timeout of zero will wait indefinitely.
230 · IN $pending_callbacks: If the allocation cannot be granted
232 immediately, the controller will put the job in the PENDING state.
233 If pending callback is given, it will be called with the job id of
234 the pending job as the sole parameter.
235 · RET: allcation response, with structure of
237 "resource_allocation_response_msg_t". On failure "undef" is returned,
238 with errno set.
239 $resp = $slurm->allocatiion_lookup($job_id);
241 Retrieve info for an existing resource allocation.
243 · IN $job_id: job allocation identifier.
245 · RET: job allocation info, with structure of
247 "resource_allocation_response_msg_t". On failure "undef" is returned
248 with errno set.
249 $str = $slurm->read_hostfile($filename, $n);
251 Read a specified SLURM hostfile. The file must contain a list of SLURM
253 NodeNames, one per line.
254 · IN $filename: name of SLURM hostlist file to be read.
256 · IN $n: number of NodeNames required.
258 · RET: a string representing the hostlist. Returns NULL if there are
260 fewer than $n hostnames in the file, or if an error occurs.
261 $msg_thr = $slurm->allocation_msg_thr_create($port, $callbacks);
263 Startup a message handler talking with the controller dealing with
265 messages from the controller during an allocation.
266 · OUT $port: port we are listening for messages on from the controller.
268 · IN $callbacks: callbacks for different types of messages, with
270 structure of "slurm_allocation_callbacks_t".
271 · RET: opaque object of "allocation_msg_thread_t *", or NULL on
273 failure.
274 $slurm->allocation_msg_thr_destroy($msg_thr);
276 Shutdown the message handler talking with the controller dealing with
278 messages from the controller during an allocation.
279 · IN $msg_thr: opaque object of "allocation_msg_thread_t" pointer.
281 $resp = $slurm->submit_batch_job($job_desc_msg);
283 Issue RPC to submit a job for later execution.
285 · IN $job_desc_msg: description of batch job request, with structure of
287 "job_desc_msg_t".
288 · RET: 0 on success, otherwise return -1 and set errno to indicate the
290 error.
291 $rc = $slurm->job_will_run($job_desc_msg);
293 Determine if a job would execute immediately if submitted now.
295 · IN $job_desc_msg: description of resource allocation request, with
297 structure of "job_desc_msg_t".
298 · RET: 0 on success, otherwise return -1 and set errno to indicate the
300 error.
301 $resp = $slurm->sbcast_lookup($job_id);
303 Retrieve info for an existing resource allocation including a
305 credential needed for sbcast.
306 · IN $jobid: job allocation identifier.
308 · RET: job allocation information includeing a credential for sbcast,
310 with structure of "job_sbcast_cred_msg_t". On failure "undef" is
311 returned with errno set.
312 JOB/STEP SIGNALING FUNCTIONS
314 $rc = $slurm->kill_job($job_id, $signal, $batch_flag=0);
315 Send the specified signal to all steps of an existing job.
317 · IN $job_id: the job's id.
319 · IN $signal: signal number.
321 · IN $batch_flag: 1 to signal batch shell only, otherwise 0.
323 · RET: 0 on success, otherwise return -1 and set errno to indicate the
325 error.
326 $rc = $slurm->kill_job_step($job_id, $step_id, $signal);
328 Send the specified signal to an existing job step.
330 · IN $job_id: the job's id.
332 · IN $step_id: the job step's id.
334 · IN $signal: signal number.
336 · RET: 0 on success, otherwise return -1 and set errno to indicate the
338 error.
339 $rc = $slurm->signal_job($job_id, $signal);
341 Send the specified signal to all steps of an existing job.
343 · IN $job_id: the job's id.
345 · IN $signal: signal number.
347 · RET: 0 on success, otherwise return -1 and set errno to indicate the
349 error.
350 $rc = $slurm->signal_job_step($job_id, $step_id, $signal);
352 Send the specified signal to an existing job step.
354 · IN $job_id: the job's id.
356 · IN $step_id: the job step's id.
358 · IN $signal: signal number.
360 · RET: 0 on success, otherwise return -1 and set errno to indicate the
362 error.
363 JOB/STEP COMPLETION FUNCTIONS
365 $rc = $slurm->complete_job($job_id, $job_rc=0);
366 Note the completion of a job and all of its steps.
368 · IN $job_id: the job's id.
370 · IN $job_rc: the highest exit code of any task of the job.
372 · RET: 0 on success, otherwise return -1 and set errno to indicate the
374 error.
375 $rc = $slurm->terminate_job_step($job_id, $step_id);
377 Terminates a job step by sending a REQUEST_TERMINATE_TASKS rpc to all
379 slurmd of a job step, and then calls slurm_complete_job_step() after
380 verifying that all nodes in the job step no longer have running tasks
381 from the job step. (May take over 35 seconds to return.)
382 · IN $job_id: the job's id.
384 · IN $step_id: the job step's id - use SLURM_BATCH_SCRIPT as the
386 step_id to terminate a job's batch script.
387 · RET: 0 on success, otherwise return -1 and set errno to indicate the
389 error.
390 SLURM TASK SPAWNING FUNCTIONS
392 $ctx = $slurm->step_ctx_create($params);
393 Create a job step and its context.
395 · IN $params: job step parameters, with structure of
397 "slurm_step_ctx_params_t".
398 · RET: the step context. On failure "undef" is returned with errno set.
400 $ctx = $slurm->step_ctx_create_no_alloc($params);
402 Create a job step and its context without getting an allocation.
404 · IN $params: job step parameters, with structure of
406 "slurm_step_ctx_params_t"..
407 · IN $step_id: fake job step id.
409 · RET: the step context. On failure "undef" is returned with errno set.
411 SLURM CONTROL CONFIGURATION READ/PRINT/UPDATE FUNCTIONS
413 ($major, $minor, $micro) = $slurm->api_version();
414 Get the SLURM API's version number.
416 · RET: a three element list of the major, minor, and micro version
418 number.
419 $resp = $slurm->load_ctl_conf($update_time=0);
421 Issue RPC to get SLURM control configuration information if changed.
423 · IN $update_time: time of current configuration data.
425 · RET: SLURM configuration data, with structure of "slurm_ctl_conf_t".
427 On failure "undef" is returned with errno set.
428 $slurm->print_ctl_conf($out, $conf);
430 Output the contents of SLURM control configuration message as loaded
432 using "load_ctl_conf()".
433 · IN $out: file to write to.
435 · IN $conf: SLURM control configuration, with structure of
437 "slurm_ctl_conf_t".
438 $list = $slurm->ctl_conf_2_key_pairs($conf);
440 Put the SLURM configuration data into a List of opaque data type
442 "config_key_pair_t".
443 · IN $conf: SLURM control configuration, with structure of
445 "slurm_ctl_conf_t".
446 · RET: List of opaque data type "config_key_pair_t".
448 $resp = $slurm->load_slurmd_status();
450 Issue RPC to get the status of slurmd daemon on this machine.
452 · RET: slurmd status info, with structure of "slurmd_status_t". On
454 failure "undef" is returned with errno set.
455 $slurm->print_slurmd_status($out, $slurmd_status);
457 Output the contents of slurmd status message as loaded using
459 "load_slurmd_status()".
460 · IN $out: file to write to.
462 · IN $slurmd_status: slurmd status info, with structure of
464 "slurmd_status_t".
465 $slurm->print_key_pairs($out, $key_pairs, $title);
467 Output the contents of key_pairs which is a list of opaque data type
469 "config_key_pair_t".
470 · IN $out: file to write to.
472 · IN $key_pairs: List containing key pairs to be printed.
474 · IN $title: title of key pair list.
476 $rc = $slurm->update_step($step_msg);
478 Update the time limit of a job step.
480 · IN $step_msg: step update messasge descriptor, with structure of
482 "step_update_request_msg_t".
483 · RET: 0 or -1 on error.
485 SLURM JOB RESOURCES READ/PRINT FUNCTIONS
487 $num = $slurm->job_cpus_allocated_on_node_id($job_res, $node_id);
488 Get the number of cpus allocated to a job on a node by node id.
490 · IN $job_res: job resources data, with structure of "job_resources_t".
492 · IN $node_id: zero-origin node id in allocation.
494 · RET: number of CPUs allocated to job on this node or -1 on error.
496 $num = $slurm->job_cpus_allocated_on_node($job_res, $node_name);
498 Get the number of cpus allocated to a job on a node by node name.
500 · IN $job_res: job resources data, with structure of "job_resources_t".
502 · IN $node_name: name of node.
504 · RET: number of CPUs allocated to job on this node or -1 on error.
506 SLURM JOB CONFIGURATION READ/PRINT/UPDATE FUNCTIONS
508 $time = $slurm->get_end_time($job_id);
509 Get the expected end time for a given slurm job.
511 · IN $jobid: SLURM job id.
513 · RET: scheduled end time for the job. On failure "undef" is returned
515 with errno set.
516 $secs = $slurm->get_rem_time($job_id);
518 Get the expected time remaining for a given job.
520 · IN $jobid: SLURM job id.
522 · RET: remaining time in seconds or -1 on error.
524 $rc = $slurm->job_node_ready($job_id);
526 Report if nodes are ready for job to execute now.
528 · IN $job_id: SLURM job id.
530 · RET:
532 · READY_JOB_FATAL: fatal error
534 · READY_JOB_ERROR: ordinary error
536 · READY_NODE_STATE: node is ready
538 · READY_JOB_STATE: job is ready to execute
540 $resp = $slurm->load_job($job_id, $show_flags=0);
542 Issue RPC to get job information for one job ID.
544 · IN $job_id: ID of job we want information about.
546 · IN $show_flags: job filtering options.
548 · RET: job information, with structure of "job_info_msg_t". On failure
550 "undef" is returned with errno set.
551 $resp = $slurm->load_jobs($update_time=0, $show_flags=0);
553 Issue RPC to get all SLURM job information if changed.
555 · IN $update_time: time of current job information data.
557 · IN $show_flags: job filtering options.
559 · RET: job information, with structure of "job_info_msg_t". On failure
561 "undef" is returned with errno set.
562 $rc = $slurm->notify_job($job_id, $message);
564 Send message to the job's stdout, usable only by user root.
566 · IN $job_id: SLURM job id or 0 for all jobs.
568 · IN $message: arbitrary message.
570 · RET: 0 or -1 on error.
572 $job_id = $slurm->pid2jobid($job_pid);
574 Issue RPC to get the SLURM job ID of a given process ID on this
576 machine.
577 · IN $job_pid: process ID of interest on this machine.
579 · RET: corresponding job ID. On failure "undef" is returned.
581 $slurm->print_job_info($out, $job_info, $one_liner=0);
583 Output information about a specific SLURM job based upon message as
585 loaded using "load_jobs()".
586 · IN $out: file to write to.
588 · IN $job_info: an individual job information record, with structure of
590 "job_info_t".
591 · IN $one_liner: print as a single line if true.
593 $slurm->print_job_info_msg($out, $job_info_msg, $one_liner=0);
595 Output information about all SLURM jobs based upon message as loaded
597 using "load_jobs()".
598 · IN $out: file to write to.
600 · IN $job_info_msg: job information message, with structure of
602 "job_info_msg_t".
603 · IN $one_liner: print as a single line if true.
605 $str = $slurm->sprint_job_info($job_info, $one_liner=0);
607 Output information about a specific SLURM job based upon message as
609 loaded using "load_jobs()".
610 · IN $job_info: an individual job information record, with structure of
612 "job_info_t".
613 · IN $one_liner: print as a single line if true.
615 · RET: string containing formatted output.
617 $rc = $slurm->update_job($job_info);
619 Issue RPC to a job's configuration per request only usable by user root
621 or (for some parameters) the job's owner.
622 · IN $job_info: description of job updates, with structure of
624 "job_desc_msg_t".
625 · RET: 0 on success, otherwise return -1 and set errno to indicate the
627 error.
628 SLURM JOB STEP CONFIGURATION READ/PRINT/UPDATE FUNCTIONS
630 $resp = $slurm->get_job_steps($update_time=0, $job_id=NO_VAL,
631 $step_id=NO_VAL, $show_flags=0);
632 Issue RPC to get specific slurm job step configuration information if
634 changed since update_time.
635 · IN $update_time: time of current configuration data.
637 · IN $job_id: get information for specific job id, NO_VAL for all jobs.
639 · IN $step_id: get information for specific job step id, NO_VAL for all
641 job steps.
642 · IN $show_flags: job step filtering options.
644 · RET: job step information, with structure of
646 "job_step_info_response_msg_t". On failure "undef" is returned with
647 errno set.
648 $slurm->print_job_step_info_msg($out, $step_info_msg, $one_liner);
650 Output information about all SLURM job steps based upon message as
652 loaded using "get_job_steps()".
653 · IN $out: file to write to.
655 · IN $step_info_msg: job step information message, with structure of
657 "job_step_info_response_msg_t".
658 · IN $one_liner: print as a single line if true.
660 $slurm->print_job_step_info($out, $step_info, $one_liner);
662 Output information about a specific SLURM job step based upon message
664 as loaded using "get_job_steps()".
665 · IN $out: file to write to.
667 · IN $step_info: job step information, with structure of
669 "job_step_info_t".
670 · IN $one_liner: print as a single line if true.
672 $str = $slurm->sprint_job_step_info($step_info, $one_liner);
674 Output information about a specific SLURM job step based upon message
676 as loaded using "get_job_steps()".
677 · IN $step_info: job step information, with structure of
679 "job_step_info_t".
680 · IN $one_liner: print as a single line if true.
682 · RET: string containing formatted output.
684 $layout = $slurm->job_step_layout_get($job_id, $step_id);
686 Get the layout structure for a particular job step.
688 · IN $job_id: SLURM job ID.
690 · IN $step_id: SLURM step ID.
692 · RET: layout of the job step, with structure of "slurm_step_layout_t".
694 On failure "undef" is returned with errno set.
695 $resp = $slurm->job_step_stat($job_id, $step_id, $nodelist=undef);
697 Get status of a current step.
699 · IN $job_id : SLURM job ID.
701 · IN $step_id: SLURM step ID.
703 · IN $nodelist: nodes to check status of step. If omitted, all nodes in
705 step are used.
706 · RET: response of step status, with structure of
708 "job_step_stat_response_msg_t". On failure "undef" is returned.
709 $resp = $slurm->job_step_get_pids($job_id, $step_id, $nodelist);
711 Get the complete list of pids for a given job step.
713 · IN $job_id: SLURM job ID.
715 · IN $step_id: SLURM step ID.
717 · IN $nodelist: nodes to check pids of step. If omitted, all nodes in
719 step are used.
720 · RET: response of pids information, with structure of
722 "job_step_pids_response_msg_t". On failure "undef" is returned.
723 SLURM NODE CONFIGURATION READ/PRINT/UPDATE FUNCTIONS
725 $resp = $slurm->load_node($update_time=0, $show_flags=0);
726 Issue RPC to get all node configuration information if changed.
728 · IN $update_time: time of current configuration data.
730 · IN $show_flags: node filtering options.
732 · RET: response hash reference with structure of "node_info_msg_t". On
734 failure "undef" is returned with errno set.
735 $slurm->print_node_info_msg($out, $node_info_msg, $one_liner=0);
737 Output information about all SLURM nodes based upon message as loaded
739 using "load_node()".
740 · IN $out: FILE handle to write to.
742 · IN $node_info_msg: node information message to print, with structure
744 of "node_info_msg_t".
745 · IN $one_liner: if true, each node info will be printed as a single
747 line.
748 $slurm->print_node_table($out, $node_info, $one_liner=0);
750 Output information about a specific SLURM node based upon message as
752 loaded using "load_node()".
753 · IN $out: FILE handle to write to.
755 · IN $node_info: an individual node information record with structure
757 of "node_info_t".
758 · IN $one_liner: whether to print as a single line.
760 $str = $slurm->sprint_node_table($node_info, $one_liner=0);
762 Output information about a specific SLURM node based upon message as
764 loaded using "load_node".
765 · IN $node_info: an individual node information record with structure
767 of "node_info_t".
768 · IN $one_liner: whether to print as a single line.
770 · RET: string containing formatted output on success, "undef" on
772 failure.
773 $rc = $slurm->update_node($node_info);
775 Issue RPC to modify a node's configuration per request, only usable by
777 user root.
778 · IN $node_info: description of node updates, with structure of
780 "update_node_msg_t".
781 · RET: 0 on success, -1 on failure with errno set.
783 SLURM SWITCH TOPOLOGY CONFIGURATION READ/PRINT FUNCTIONS
785 $resp = $slurm->load_topo();
786 Issue RPC to get all switch topology configuration information.
788 · RET: response hash reference with structure of
790 "topo_info_response_msg_t". On failure "undef" is returned with errno
791 set.
792 $slurm->print_topo_info_msg($out, $topo_info_msg, $one_liner=0);
794 Output information about all switch topology configuration information
796 based upon message as loaded using "load_topo()".
797 · IN $out: FILE handle to write to.
799 · IN $topo_info_msg: swith topology information message, with structure
801 of "topo_info_response_msg_t".
802 · IN $one_liner: print as a single line if not zero.
804 $slurm->print_topo_record($out, $topo_info, $one_liner);
806 Output information about a specific SLURM topology record based upon
808 message as loaded using "load_topo()".
809 · IN $out: FILE handle to write to.
811 · IN $topo_info: an individual switch information record, with
813 structure of "topo_info_t".
814 · IN $one_liner: print as a single line if not zero.
816 SLURM SELECT READ/PRINT/UPDATE FUNCTIONS
818 $rc = $slurm->get_select_nodeinfo($nodeinfo, $data_type, $state,
819 $data);
820 Get data from a select node credential.
822 · IN $nodeinfo: select node credential to get data from.
824 · IN $data_type: type of data to get.
826 · TODO: enumerate data type and returned value.
828 · IN $state: state of node query.
830 · OUT $data: the data got.
832 SLURM PARTITION CONFIGURATION READ/PRINT/UPDATE FUNCTIONS
834 $resp = $slurm->load_partitions($update_time=0, $show_flags=0);
835 Issue RPC to get all SLURM partition configuration information if
837 changed.
838 · IN $update_time: time of current configuration data.
840 · IN $show_flags: partitions filtering options.
842 · RET: response hash reference with structure of
844 "partition_info_msg_t".
845 $slurm->print_partition_info_msg($out, $part_info_msg, $one_liner=0);
847 Output information about all SLURM partitions based upon message as
849 loaded using "load_partitions()".
850 · IN $out: FILE handle to write to.
852 · IN $part_info_msg: partitions information message, with structure of
854 "partition_info_msg_t".
855 · IN $one_liner: print as a single line if true.
857 $slurm->print_partition_info($out, $part_info, $one_liner=0);
859 Output information about a specific SLURM partition based upon message
861 as loaded using "load_partitions()".
862 · IN $out: FILE handle to write to.
864 · IN $part_info: an individual partition information record, with
866 structure of "partition_info_t".
867 · IN $one_liner: print as a single ine if true.
869 $str = $slurm->sprint_partition_info($part_info, $one_liner=0);
871 Output information about a specific SLURM partition based upon message
873 as loaded using "load_reservations()".
874 · IN $part_info: an individual partition information record, with
876 structure of "partition_info_t".
877 · IN $one_liner: print as a single line if true.
879 · RET: string containing formatted output. On failure "undef" is
881 returned.
882 $rc = $slurm->create_partition($part_info);
884 Create a new partition, only usable by user root.
886 · IN $part_info: description of partition configuration with structure
888 of "update_part_msg_t".
889 · RET: 0 on success, -1 on failure with errno set.
891 $rc = $slurm->update_partition($part_info);
893 Issue RPC to update a partition's configuration per request, only
895 usable by user root.
896 · IN $part_info: description of partition updates with structure of
898 "update_part_msg_t".
899 · RET: 0 on success, -1 on failure with errno set.
901 $rc = $slurm->delete_partition($part_info)
903 Issue RPC to delete a partition, only usable by user root.
905 · IN $part_info: description of partition to delete, with structure of
907 "delete_part_msg_t".
908 · RET: 0 on success, -1 on failure with errno set.
910 SLURM RESERVATION CONFIGURATION READ/PRINT/UPDATE FUNCTIONS
912 $name = $slurm->create_reservation($resv_info);
913 Create a new reservation, only usable by user root.
915 · IN $resv_info: description of reservation, with structure of
917 "resv_desc_msg_t".
918 · RET: name of reservation created. On failure "undef" is returned with
920 errno set.
921 $rc = $slurm->update_reservation($resv_info);
923 Modify an existing reservation, only usable by user root.
925 · IN $resv_info: description of reservation, with structure of
927 "resv_desc_msg_t".
928 · RET: error code.
930 $rc = $slurm->delete_reservation($resv_info);
932 Issue RPC to delete a reservation, only usable by user root.
934 · IN $resv_info: description of reservation to delete, with structure
936 of "reservation_name_msg_t".
937 · RET: error code
939 $resp = $slurm->load_reservations($update_time=0);
941 Issue RPC to get all SLURM reservation configuration information if
943 changed.
944 · IN $update_time: time of current configuration data.
946 · RET: response of reservation information, with structure of
948 "reserve_info_msg_t". On failure "undef" is returned with errno set.
949 $slurm->print_reservation_info_msg($out, $resv_info_msg, $one_liner=0);
951 Output information about all SLURM reservations based upon message as
953 loaded using "load_reservation()".
954 · IN $out: FILE handle to write to.
956 · IN $resv_info_msg: reservation information message, with structure of
958 "reserve_info_msg_t".
959 · IN $one_liner: print as a single line if true.
961 $slurm->print_reservation_info($out, $resv_info, $one_liner=0);
963 Output information about a specific SLURM reservation based upon
965 message as loaded using "load_reservation()".
966 · IN $out: FILE handle to write to.
968 · IN $resv_info: an individual reservation information record, with
970 structure of "reserve_info_t".
971 · IN $one_liner: print as a single line if true.
973 $str = $slurm->sprint_reservation_info($resv_info, $one_liner=0);
975 Output information about a specific SLURM reservation based upon
977 message as loaded using "load_reservations()".
978 · IN $resv_info: an individual reservation information record, with
980 structure of "reserve_info_t".
981 · IN $one_liner: print as a single line if true.
983 · RET: string containing formatted output. On failure "undef" is
985 returned.
986 SLURM PING/RECONFIGURE/SHUTDOWN FUNCTIONS
988 $rc = $slurm->ping($primary);
989 Issue RPC to ping Slurm controller (slurmctld).
991 · IN primary: 1 for primary controller, 2 for secondary controller.
993 · RET: error code.
995 $rc = $slurm->reconfigure()
997 Issue RPC to have Slurm controller (slurmctld) reload its configuration
999 file.
1000 · RET: error code.
1002 $rc = $slurm->shutdown($options);
1004 Issue RPC to have Slurm controller (slurmctld) cease operations, both
1006 the primary and backup controller are shutdown.
1007 · IN $options:
1009 · 0: all slurm daemons are shutdown.
1011 · 1: slurmctld generates a core file.
1013 · 2: only the slurmctld is shutdown (no core file).
1015 · RET: error code.
1017 $rc = $slurm->takeover();
1019 Issue RPC to have Slurm backup controller take over the primary
1021 controller. REQUEST_CONTROL is sent by the backup to the primary
1022 controller to take control.
1023 · RET: error code.
1025 $rc = $slurm->set_debug_level($debug_level)
1027 Issue RPC to set slurm controller debug level.
1029 · IN $debug_level: requested debug level.
1031 · RET: 0 on success, -1 on error with errno set.
1033 $rc = $slurm->set_schedlog_level($schedlog_level);
1035 Issue RPC to set slurm scheduler log level.
1037 · schedlog_level: requested scheduler log level.
1039 · RET: 0 on success, -1 on error with errno set.
1041 SLURM JOB SUSPEND FUNCTIONS
1043 $rc = $slurm->suspend($job_id);
1044 Suspend execution of a job.
1046 · IN $job_id: job on which top perform operation.
1048 · RET: error code.
1050 $rc = $slurm->resume($job_id);
1052 Resume execution of a previously suspended job.
1054 · IN $job_id: job on which to perform operation.
1056 · RET: error code.
1058 $rc = $slurm->requeue($job_id);
1060 Re-queue a batch job, if already running then terminate it first.
1062 · IN $job_id: job on which to perform operation.
1064 · RET: error code.
1066 SLURM JOB CHECKPOINT FUNCTIONS
1068 $rc = $slurm->checkpoint_able($job_id, $step_id, $start_time);
1069 Determine if the specified job step can presently be checkpointed.
1071 · IN $job_id: job on which to perform operation.
1073 · IN $step_id: job step on which to perform operation.
1075 · OUT $start_time: time at which checkpoint request was issued.
1077 · RET: 0 (can be checkpoined) or a slurm error code.
1079 $rc = $slurm->checkpoint_disable($job_id, $step_id);
1081 Disable checkpoint requests for some job step.
1083 · IN $job_id: job on which to perform operation.
1085 · IN $step_id: job step on which to perform operation.
1087 · RET: error code.
1089 $rc = $slurm->checkpoint_enable($job_id, $step_id);
1091 Enable checkpoint requests for some job step.
1093 · IN $job_id: job on which to perform operation.
1095 · IN $step_id: job step on which to perform operation.
1097 · RET: error code.
1099 $rc = $slurm->checkpoint_create($job_id, $step_id, $max_wait,
1101 $image_dir);
1102 Initiate a checkpoint requests for some job step. The job will continue
1104 execution after the checkpoint operation completes.
1105 · IN $job_id: job on which to perform operation.
1107 · IN $step_id: job step on which to perform operation.
1109 · IN $max_wait: maximum wait for operation to complete, in seconds.
1111 · IN $image_dir: directory to store image files.
1113 · RET: error code.
1115 $rc = $slurm->checkpoint_vacate($job_id, $step_id, $max_wait,
1117 $image_dir);
1118 Initiate a checkpoint requests for some job step. The job will
1120 terminate after the checkpoint operation completes.
1121 · IN $job_id: job on which to perform operation.
1123 · IN $step_id: job step on which to perform operation.
1125 · IN $max_wait: maximum wait for operation to complete, in seconds.
1127 · IN $image_dir: directory to store image files.
1129 · RET: error code.
1131 $rc = $slurm->checkpoint_restart($job_id, $step_id, $stick, $image_dir)
1133 Restart execution of a checkpointed job step.
1135 · IN $job_id: job on which to perform operation.
1137 · IN $step_id: job step on which to perform operation.
1139 · IN $stick: if true, stick to nodes previously running on.
1141 · IN $image_dir: directory to find checkpoint image files.
1143 · RET: error code.
1145 $rc = $slurm->checkpoint_complete($job_id, $step_id, $begin_time,
1147 $error_code, $error_msg);
1148 Note the completion of a job step's checkpoint operation.
1150 · IN $job_id: job on which to perform operation.
1152 · IN $step_id: job step on which to perform operation.
1154 · IN $begin_time: time at which checkpoint began.
1156 · IN $error_code: error code, highest value for all complete calls is
1158 preserved.
1159 · IN $error_msg: error message, preserved for highest error_code.
1161 · RET: error code.
1163 checkpoint_task_complete($job_id, $step_id, $task_id, $begin_time,
1165 $error_code, $error_msg);
1166 Note the completion of a task's checkpoint operation.
1168 · IN $job_id: job on which to perform operation.
1170 · IN $step_id: job step on which to perform operation.
1172 · IN $task_id: task which completed the operation.
1174 · IN $begin_time: time at which checkpoint began.
1176 · IN $error_code: error code, highest value for all complete calls is
1178 preserved.
1179 · IN $error_msg: error message, preserved for highest error_code.
1181 · RET: error code.
1183 $rc = $slurm->checkpoint_error($job_id, $step_id, $error_code,
1185 $error_msg);
1186 Gather error information for the last checkpoint operation for some job
1188 step.
1189 · IN $job_id: job on which to perform operation.
1191 · IN $step_id: job step on which to perform operation.
1193 · OUT $error_code: error number associated with the last checkpoint
1195 operation.
1196 · OUT $error_msg: error message associated with the last checkpoint
1198 operation.
1199 · RET: error code.
1201 $rc = $slurm->checkpoint_tasks($job_id, $step_id, $image_dir,
1203 $max_wait, $nodelist);
1204 Send checkoint request to tasks of specified job step.
1206 · IN $job_id: job on which to perform operation.
1208 · IN $step_id: job step on which to perform operation.
1210 · IN $image_dir: location to store checkpoint image files.
1212 · IN $max_wait: seconds to wait for the operation to complete.
1214 · IN $nodelist: nodes to send the request.
1216 · RET: 0 on success, non-zero on failure with errno set.
1218 SLURM TRIGGER FUNCTIONS
1220 $rc = $slurm->set_trigger($trigger_info);
1221 Set an event trigger.
1223 · IN $trigger_info: hash reference of specification of trigger to
1225 create, with structure of "trigger_info_t".
1226 · RET: error code.
1228 $rc = $slurm->clear_trigger($trigger_info);
1230 Clear an existing event trigger.
1232 · IN $trigger_info: hash reference of specification of trigger to
1234 remove, with structure of "trigger_info_t".
1235 · RET: error code.
1237 $resp = $slurm->get_triggers();
1239 Get all event trigger information.
1241 · RET: hash reference with structure of "trigger_info_msg_t". On
1243 failure "undef" is returned with errno set.
1244 JOB/NODE STATE TESTING FUNCTIONS
1246 The following are functions to test job/node state, based on the macros
1247 defined in src/common/slurm_protocol_defs.h. The functions take a
1248 parameter of a hash reference of a job/node, and return a boolean
1249 value. For job, $job->{job_state} is tested. For node,
1250 $node->{node_state} is tested.
1251 $cond = IS_JOB_PENDING($job);
1253 $cond = IS_JOB_RUNNING($job);
1255 $cond = IS_JOB_SUSPENDED($job);
1257 $cond = IS_JOB_COMPLETE($job);
1259 $cond = IS_JOB_CANCELLED($job);
1261 $cond = IS_JOB_FAILED($job);
1263 $cond = IS_JOB_TIMEOUT($job);
1265 $cond = IS_JOB_NODE_FAILED($job);
1267 $cond = IS_JOB_COMPLETING($job);
1269 $cond = IS_JOB_CONFIGURING($job);
1271 $cond = IS_JOB_STARTED($job);
1273 $cond = IS_JOB_FINISHED($job);
1275 $cond = IS_JOB_COMPLETED($job);
1277 $cond = IS_JOB_RESIZING($job);
1279 $cond = IS_NODE_UNKNOWN($node);
1281 $cond = IS_NODE_DOWN($node);
1283 $cond = IS_NODE_IDLE($node);
1285 $cond = IS_NODE_ALLOCATED($node);
1287 $cond = IS_NODE_ERROR($node);
1289 $cond = IS_NODE_MIXED($node);
1291 $cond = IS_NODE_FUTURE($node);
1293 $cond = IS_NODE_DRAIN($node);
1295 $cond = IS_NODE_DRAINING($node);
1297 $cond = IS_NODE_DRAINED($node);
1299 $cond = IS_NODE_COMPLETING($node);
1301 $cond = IS_NODE_NO_RESPOND($node);
1303 $cond = IS_NODE_POWER_SAVE($node);
1305 $cond = IS_NODE_POWER_UP($node);
1307 $cond = IS_NODE_FAIL($node);
1309 $cond = IS_NODE_MAINT($node);
1311
1313 The job/node state testing functions are exported by default.
1314 If ':constant' if specified, all constants are exported.
1316
1318 Slurm::Constant, Slurm::Hostlist, Slurm::Stepctx, Slurm::Bitstr
1319 <slurm/slurm.h> for various hash reference structures.
1321 Home page of SLURM: <http://www.llnl.gov/linux/slurm/>.
1323
1325 This library is created by Hongjia Cao, <hjcao(AT)nudt.edu.cn> and
1326 Danny Auble, <da(AT)llnl.gov>. It is distributed with SLURM.
1327
1329 This library is free software; you can redistribute it and/or modify it
1330 under the same terms as Perl itself, either Perl version 5.8.4 or, at
1331 your option, any later version of Perl 5 you may have available.
1332perl v5.28.2 2019-04-12 Slurm(3)