1sinfo(1) Slurm Commands sinfo(1)
2
3
4
6 sinfo - view information about Slurm nodes and partitions.
7
8
10 sinfo [OPTIONS...]
11
13 sinfo is used to view partition and node information for a system run‐
14 ning Slurm.
15
16
18 -a, --all
19 Display information about all partitions. This causes informa‐
20 tion to be displayed about partitions that are configured as
21 hidden and partitions that are unavailable to user's group.
22
23
24 -d, --dead
25 If set only report state information for non-responding (dead)
26 nodes.
27
28
29 -e, --exact
30 If set, do not group node information on multiple nodes unless
31 their configurations to be reported are identical. Otherwise cpu
32 count, memory size, and disk space for nodes will be listed with
33 the minimum value followed by a "+" for nodes with the same par‐
34 tition and state (e.g., "250+").
35
36
37 --federation
38 Show all partitions from the federation if a member of one.
39
40
41 -h, --noheader
42 Do not print a header on the output.
43
44
45 --help Print a message describing all sinfo options.
46
47
48 --hide Do not display information about hidden partitions. By default,
49 partitions that are configured as hidden or are not available to
50 the user's group will not be displayed (i.e. this is the default
51 behavior).
52
53
54 -i <seconds>, --iterate=<seconds>
55 Print the state on a periodic basis. Sleep for the indicated
56 number of seconds between reports. By default, prints a time
57 stamp with the header.
58
59
60 --local
61 Show only jobs local to this cluster. Ignore other clusters in
62 this federation (if any). Overrides --federation.
63
64
65 -l, --long
66 Print more detailed information. This is ignored if the --for‐
67 mat option is specified.
68
69
70 -M, --clusters=<string>
71 Clusters to issue commands to. Multiple cluster names may be
72 comma separated. A value of of 'all' will query to run on all
73 clusters. Note that the SlurmDBD must be up for this option to
74 work properly. This option implicitly sets the --local option.
75
76
77 -n <nodes>, --nodes=<nodes>
78 Print information only about the specified node(s). Multiple
79 nodes may be comma separated or expressed using a node range
80 expression. For example "linux[00-07]" would indicate eight
81 nodes, "linux00" through "linux07." Performance of the command
82 can be measurably improved for systems with large numbers of
83 nodes when a single node name is specified.
84
85
86 --noconvert
87 Don't convert units from their original type (e.g. 2048M won't
88 be converted to 2G).
89
90
91 -N, --Node
92 Print information in a node-oriented format with one line per
93 node and partition. That is, if a node belongs to more than one
94 partition, then one line for each node-partition pair will be
95 shown. If --partition is also specified, then only one line per
96 node in this partition is shown. The default is to print infor‐
97 mation in a partition-oriented format. This is ignored if the
98 --format option is specified.
99
100
101 -o <output_format>, --format=<output_format>
102 Specify the information to be displayed using an sinfo format
103 string. Format strings transparently used by sinfo when running
104 with various options are
105
106 default "%#P %.5a %.10l %.6D %.6t %N"
107
108 --summarize "%#P %.5a %.10l %.16F %N"
109
110 --long "%#P %.5a %.10l %.10s %.4r %.8h %.10g %.6D %.11T
111 %N"
112
113 --Node "%#N %.6D %#P %6t"
114
115 --long --Node "%#N %.6D %#P %.11T %.4c %.8z %.6m %.8d %.6w %.8f
116 %20E"
117
118 --list-reasons "%20E %9u %19H %N"
119
120 --long --list-reasons
121 "%20E %12U %19H %6t %N"
122
123
124 In the above format strings, the use of "#" represents the maxi‐
125 mum length of any partition name or node list to be printed. A
126 pass is made over the records to be printed to establish the
127 size in order to align the sinfo output, then a second pass is
128 made over the records to print them. Note that the literal
129 character "#" itself is not a valid field length specification,
130 but is only used to document this behaviour.
131
132 The field specifications available include:
133
134 %all Print all fields available for this data type with a ver‐
135 tical bar separating each field.
136
137 %a State/availability of a partition
138
139 %A Number of nodes by state in the format "allocated/idle".
140 Do not use this with a node state option ("%t" or "%T") or
141 the different node states will be placed on separate
142 lines.
143
144 %b Features currently active on the nodes, also see %f
145
146 %B The max number of CPUs per node available to jobs in the
147 partition.
148
149 %c Number of CPUs per node
150
151 %C Number of CPUs by state in the format "allo‐
152 cated/idle/other/total". Do not use this with a node state
153 option ("%t" or "%T") or the different node states will be
154 placed on separate lines.
155
156 %d Size of temporary disk space per node in megabytes
157
158 %D Number of nodes
159
160 %e Free memory of a node
161
162 %E The reason a node is unavailable (down, drained, or drain‐
163 ing states).
164
165 %f Features available the nodes, also see %b
166
167 %F Number of nodes by state in the format "allo‐
168 cated/idle/other/total". Note the use of this format
169 option with a node state format option ("%t" or "%T") will
170 result in the different node states being be reported on
171 separate lines.
172
173 %g Groups which may use the nodes
174
175 %G Generic resources (gres) associated with the nodes
176
177 %h Jobs may oversubscribe compute resources (i.e. CPUs),
178 "yes", "no", "exclusive" or "force"
179
180 %H Print the timestamp of the reason a node is unavailable.
181
182 %I Partition job priority weighting factor.
183
184 %l Maximum time for any job in the format "days-hours:min‐
185 utes:seconds"
186
187 %L Default time for any job in the format "days-hours:min‐
188 utes:seconds"
189
190 %m Size of memory per node in megabytes
191
192 %M PreemptionMode
193
194 %n List of node hostnames
195
196 %N List of node names
197
198 %o List of node communication addresses
199
200 %O CPU load of a node
201
202 %p Partition scheduling tier priority.
203
204 %P Partition name followed by "*" for the default partition,
205 also see %R
206
207 %r Only user root may initiate jobs, "yes" or "no"
208
209 %R Partition name, also see %P
210
211 %s Maximum job size in nodes
212
213 %S Allowed allocating nodes
214
215 %t State of nodes, compact form
216
217 %T State of nodes, extended form
218
219 %u Print the user name of who set the reason a node is
220 unavailable.
221
222 %U Print the user name and uid of who set the reason a node
223 is unavailable.
224
225 %v Print the version of the running slurmd daemon.
226
227 %V Print the cluster name if running in a federation
228
229 %w Scheduling weight of the nodes
230
231 %X Number of sockets per node
232
233 %Y Number of cores per socket
234
235 %Z Number of threads per core
236
237 %z Extended processor information: number of sockets, cores,
238 threads (S:C:T) per node
239
240 %.<*> right justification of the field
241
242 %<Number><*>
243 size of field
244
245
246 -O <output_format>, --Format=<output_format>
247 Specify the information to be displayed. Also see the -o <out‐
248 put_format>, --format=<output_format> option described below
249 (which supports greater flexibility in formatting, but does not
250 support access to all fields because we ran out of letters).
251 Requests a comma separated list of job information to be dis‐
252 played.
253
254
255 The format of each field is "type[:[.]size]"
256
257 size is the minimum field size. If no size is specified, 20
258 characters will be allocated to print the information.
259
260 . indicates the output should be right justified and size
261 must be specified. By default, output is left justi‐
262 fied.
263
264
265 Valid type specifications include:
266
267 all Print all fields available in the -o format for this data
268 type with a vertical bar separating each field.
269
270 allocmem
271 Prints the amount of allocated memory on a node.
272
273 allocnodes
274 Allowed allocating nodes.
275
276 available
277 State/availability of a partition.
278
279 cluster
280 Print the cluster name if running in a federation
281
282 cpus Number of CPUs per node.
283
284 cpusload
285 CPU load of a node.
286
287 freemem
288 Free memory of a node.
289
290 cpusstate
291 Number of CPUs by state in the format "allo‐
292 cated/idle/other/total". Do not use this with a node state
293 option ("%t" or "%T") or the different node states will be
294 placed on separate lines.
295
296 cores Number of cores per socket.
297
298 defaulttime
299 Default time for any job in the format "days-hours:min‐
300 utes:seconds".
301
302 disk Size of temporary disk space per node in megabytes.
303
304 features
305 Features available on the nodes. Also see features_act.
306
307 features_act
308 Features currently active on the nodes. Also see features.
309
310 groups
311 Groups which may use the nodes.
312
313 gres Generic resources (gres) associated with the nodes.
314
315 maxcpuspernode
316 The max number of CPUs per node available to jobs in the
317 partition.
318
319 memory
320 Size of memory per node in megabytes.
321
322 nodes Number of nodes.
323
324 nodeaddr
325 List of node communication addresses.
326
327 nodeai
328 Number of nodes by state in the format "allocated/idle".
329 Do not use this with a node state option ("%t" or "%T") or
330 the different node states will be placed on separate
331 lines.
332
333 nodeaiot
334 Number of nodes by state in the format "allo‐
335 cated/idle/other/total". Do not use this with a node
336 state option ("%t" or "%T") or the different node states
337 will be placed on separate lines.
338
339 nodehost
340 List of node hostnames.
341
342 nodelist
343 List of node names.
344
345 oversubscribe
346 Jobs may oversubscribe compute resources (i.e. CPUs),
347 "yes", "no", "exclusive" or "force".
348
349 partition
350 Partition name followed by "*" for the default partition,
351 also see %R.
352
353 partitionname
354 Partition name, also see %P.
355
356 port Node TCP port.
357
358 preemptmode
359 PreemptionMode.
360
361 priorityjobfactor
362 Partition factor used by priority/multifactor plugin in
363 calculating job priority.
364
365 prioritytier or priority
366 Partition scheduling tier priority.
367
368 reason
369 The reason a node is unavailable (down, drained, or drain‐
370 ing states).
371
372 root Only user root may initiate jobs, "yes" or "no".
373
374 size Maximum job size in nodes.
375
376 statecompact
377 State of nodes, compact form.
378
379 statelong
380 State of nodes, extended form.
381
382 sockets
383 Number of sockets per node.
384
385 socketcorethread
386 Extended processor information: number of sockets, cores,
387 threads (S:C:T) per node.
388
389 time Maximum time for any job in the format "days-hours:min‐
390 utes:seconds".
391
392 timestamp
393 Print the timestamp of the reason a node is unavailable.
394
395 threads
396 Number of threads per core.
397
398 user Print the user name of who set the reason a node is
399 unavailable.
400
401 userlong
402 Print the user name and uid of who set the reason a node
403 is unavailable.
404
405 version
406 Print the version of the running slurmd daemon.
407
408 weight
409 Scheduling weight of the nodes.
410
411
412 -p <partition>, --partition=<partition>
413 Print information only about the specified partition(s). Multi‐
414 ple partitions are separated by commas.
415
416
417 -r, --responding
418 If set only report state information for responding nodes.
419
420
421 -R, --list-reasons
422 List reasons nodes are in the down, drained, fail or failing
423 state. When nodes are in these states Slurm supports optional
424 inclusion of a "reason" string by an administrator. This option
425 will display the first 20 characters of the reason field and
426 list of nodes with that reason for all nodes that are, by
427 default, down, drained, draining or failing. This option may be
428 used with other node filtering options (e.g. -r, -d, -t, -n),
429 however, combinations of these options that result in a list of
430 nodes that are not down or drained or failing will not produce
431 any output. When used with -l the output additionally includes
432 the current node state.
433
434
435 -s, --summarize
436 List only a partition state summary with no node state details.
437 This is ignored if the --format option is specified.
438
439
440 -S <sort_list>, --sort=<sort_list>
441 Specification of the order in which records should be reported.
442 This uses the same field specification as the <output_format>.
443 Multiple sorts may be performed by listing multiple sort fields
444 separated by commas. The field specifications may be preceded
445 by "+" or "-" for ascending (default) and descending order
446 respectively. The partition field specification, "P", may be
447 preceded by a "#" to report partitions in the same order that
448 they appear in Slurm's configuration file, slurm.conf. For
449 example, a sort value of "+P,-m" requests that records be
450 printed in order of increasing partition name and within a par‐
451 tition by decreasing memory size. The default value of sort is
452 "#P,-t" (partitions ordered as configured then decreasing node
453 state). If the --Node option is selected, the default sort
454 value is "N" (increasing node name).
455
456
457 -t <states> , --states=<states>
458 List nodes only having the given state(s). Multiple states may
459 be comma separated and the comparison is case insensitive. Pos‐
460 sible values include (case insensitive): ALLOC, ALLOCATED, COMP,
461 COMPLETING, DOWN, DRAIN (for node in DRAINING or DRAINED
462 states), DRAINED, DRAINING, FAIL, FUTURE, FUTR, IDLE, MAINT,
463 MIX, MIXED, NO_RESPOND, NPC, PERFCTRS, POWER_DOWN, POWER_UP,
464 RESV, RESERVED, UNK, and UNKNOWN. By default nodes in the spec‐
465 ified state are reported whether they are responding or not.
466 The --dead and --responding options may be used to filtering
467 nodes by the responding flag.
468
469
470 -T, --reservation
471 Only display information about Slurm reservations.
472
473
474 --usage
475 Print a brief message listing the sinfo options.
476
477
478 -v, --verbose
479 Provide detailed event logging through program execution.
480
481
482 -V, --version
483 Print version information and exit.
484
485
487 AVAIL Partition state: up or down.
488
489 CPUS Count of CPUs (processors) on each node.
490
491 S:C:T Count of sockets (S), cores (C), and threads (T) on these nodes.
492
493 SOCKETS
494 Count of sockets on these nodes.
495
496 CORES Count of cores on these nodes.
497
498 THREADS
499 Count of threads on these nodes.
500
501 GROUPS Resource allocations in this partition are restricted to the
502 named groups. all indicates that all groups may use this parti‐
503 tion.
504
505 JOB_SIZE
506 Minimum and maximum node count that can be allocated to any user
507 job. A single number indicates the minimum and maximum node
508 count are the same. infinite is used to identify partitions
509 without a maximum node count.
510
511 TIMELIMIT
512 Maximum time limit for any user job in days-hours:minutes:sec‐
513 onds. infinite is used to identify partitions without a job
514 time limit.
515
516 MEMORY Size of real memory in megabytes on these nodes.
517
518 NODELIST
519 Names of nodes associated with this configuration/partition.
520
521 NODES Count of nodes with this particular configuration.
522
523 NODES(A/I)
524 Count of nodes with this particular configuration by node state
525 in the form "available/idle".
526
527 NODES(A/I/O/T)
528 Count of nodes with this particular configuration by node state
529 in the form "available/idle/other/total".
530
531 PARTITION
532 Name of a partition. Note that the suffix "*" identifies the
533 default partition.
534
535 PORT Local TCP port used by slurmd on the node.
536
537 ROOT Is the ability to allocate resources in this partition
538 restricted to user root, yes or no.
539
540 OVERSUBSCRIBE
541 Will jobs allocated resources in this partition oversubscribe
542 those compute resources (i.e. CPUs). no indicates resources are
543 never oversubscribed. exclusive indicates whole nodes are dedi‐
544 cated to jobs (equivalent to srun --exclusive option, may be
545 used even with select/cons_res managing individual processors).
546 force indicates resources are always available to be oversub‐
547 scribed. yes indicates resource may be oversubscribed or not
548 per job's resource allocation.
549
550 STATE State of the nodes. Possible states include: allocated, com‐
551 pleting, down, drained, draining, fail, failing, future, idle,
552 maint, mixed, perfctrs, power_down, power_up, reserved, and
553 unknown plus Their abbreviated forms: alloc, comp, down, drain,
554 drng, fail, failg, futr, idle, maint, mix, npc, pow_dn, pow_up,
555 resv, and unk respectively. Note that the suffix "*" identifies
556 nodes that are presently not responding.
557
558 TMP_DISK
559 Size of temporary disk space in megabytes on these nodes.
560
561
563 Node state codes are shortened as required for the field size. These
564 node states may be followed by a special character to identify state
565 flags associated with the node. The following node sufficies and
566 states are used:
567
568 * The node is presently not responding and will not be allocated any
569 new work. If the node remains non-responsive, it will be placed in
570 the DOWN state (except in the case of COMPLETING, DRAINED, DRAIN‐
571 ING, FAIL, FAILING nodes).
572
573 ~ The node is presently in a power saving mode (typically running at
574 reduced frequency).
575
576 # The node is presently being powered up or configured.
577
578 $ The node is currently in a reservation with a flag value of "main‐
579 tenance".
580
581 @ The node is pending reboot.
582
583 ALLOCATED The node has been allocated to one or more jobs.
584
585 ALLOCATED+ The node is allocated to one or more active jobs plus one
586 or more jobs are in the process of COMPLETING.
587
588 COMPLETING All jobs associated with this node are in the process of
589 COMPLETING. This node state will be removed when all of
590 the job's processes have terminated and the Slurm epilog
591 program (if any) has terminated. See the Epilog parameter
592 description in the slurm.conf man page for more informa‐
593 tion.
594
595 DOWN The node is unavailable for use. Slurm can automatically
596 place nodes in this state if some failure occurs. System
597 administrators may also explicitly place nodes in this
598 state. If a node resumes normal operation, Slurm can auto‐
599 matically return it to service. See the ReturnToService and
600 SlurmdTimeout parameter descriptions in the slurm.conf(5)
601 man page for more information.
602
603 DRAINED The node is unavailable for use per system administrator
604 request. See the update node command in the scontrol(1)
605 man page or the slurm.conf(5) man page for more informa‐
606 tion.
607
608 DRAINING The node is currently executing a job, but will not be
609 allocated to additional jobs. The node state will be
610 changed to state DRAINED when the last job on it completes.
611 Nodes enter this state per system administrator request.
612 See the update node command in the scontrol(1) man page or
613 the slurm.conf(5) man page for more information.
614
615 FAIL The node is expected to fail soon and is unavailable for
616 use per system administrator request. See the update node
617 command in the scontrol(1) man page or the slurm.conf(5)
618 man page for more information.
619
620 FAILING The node is currently executing a job, but is expected to
621 fail soon and is unavailable for use per system administra‐
622 tor request. See the update node command in the scon‐
623 trol(1) man page or the slurm.conf(5) man page for more
624 information.
625
626 FUTURE The node is currently not fully configured, but expected to
627 be available at some point in the indefinite future for
628 use.
629
630 IDLE The node is not allocated to any jobs and is available for
631 use.
632
633 MAINT The node is currently in a reservation with a flag value of
634 "maintainence".
635
636 REBOOT The node is currently scheduled to be rebooted.
637
638 MIXED The node has some of its CPUs ALLOCATED while others are
639 IDLE.
640
641 PERFCTRS (NPC)
642 Network Performance Counters associated with this node are
643 in use, rendering this node as not usable for any other
644 jobs
645
646 POWER_DOWN The node is currently powered down and not capable of run‐
647 ning any jobs.
648
649 POWER_UP The node is currently in the process of being powered up.
650
651 RESERVED The node is in an advanced reservation and not generally
652 available.
653
654 UNKNOWN The Slurm controller has just started and the node's state
655 has not yet been determined.
656
657
659 Some sinfo options may be set via environment variables. These environ‐
660 ment variables, along with their corresponding options, are listed
661 below. (Note: Commandline options will always override these settings.)
662
663 SINFO_ALL -a, --all
664
665 SINFO_FEDERATION Same as --federation
666
667 SINFO_FORMAT -o <output_format>, --format=<output_format>
668
669 SINFO_LOCAL Same as --local
670
671 SINFO_PARTITION -p <partition>, --partition=<partition>
672
673 SINFO_SORT -S <sort>, --sort=<sort>
674
675 SLURM_CLUSTERS Same as --clusters
676
677 SLURM_CONF The location of the Slurm configuration file.
678
679 SLURM_TIME_FORMAT Specify the format used to report time stamps. A
680 value of standard, the default value, generates
681 output in the form
682 "year-month-dateThour:minute:second". A value of
683 relative returns only "hour:minute:second" if the
684 current day. For other dates in the current year
685 it prints the "hour:minute" preceded by "Tomorr"
686 (tomorrow), "Ystday" (yesterday), the name of the
687 day for the coming week (e.g. "Mon", "Tue", etc.),
688 otherwise the date (e.g. "25 Apr"). For other
689 years it returns a date month and year without a
690 time (e.g. "6 Jun 2012"). All of the time stamps
691 use a 24 hour format.
692
693 A valid strftime() format can also be specified.
694 For example, a value of "%a %T" will report the day
695 of the week and a time stamp (e.g. "Mon 12:34:56").
696
697
699 Report basic node and partition configurations:
700
701
702 > sinfo
703 PARTITION AVAIL TIMELIMIT NODES STATE NODELIST
704 batch up infinite 2 alloc adev[8-9]
705 batch up infinite 6 idle adev[10-15]
706 debug* up 30:00 8 idle adev[0-7]
707
708
709 Report partition summary information:
710
711 > sinfo -s
712 PARTITION AVAIL TIMELIMIT NODES(A/I/O/T) NODELIST
713 batch up infinite 2/6/0/8 adev[8-15]
714 debug* up 30:00 0/8/0/8 adev[0-7]
715
716
717 Report more complete information about the partition debug:
718
719 > sinfo --long --partition=debug
720 PARTITION AVAIL TIMELIMIT JOB_SIZE ROOT OVERSUBS GROUPS NODES STATE NODELIST
721 debug* up 30:00 8 no no all 8 idle dev[0-7]
722
723 Report only those nodes that are in state DRAINED:
724
725 > sinfo --states=drained
726 PARTITION AVAIL NODES TIMELIMIT STATE NODELIST
727 debug* up 2 30:00 drain adev[6-7]
728
729
730 Report node-oriented information with details and exact matches:
731
732 > sinfo -Nel
733 NODELIST NODES PARTITION STATE CPUS MEMORY TMP_DISK WEIGHT FEATURES REASON
734 adev[0-1] 2 debug* idle 2 3448 38536 16 (null) (null)
735 adev[2,4-7] 5 debug* idle 2 3384 38536 16 (null) (null)
736 adev3 1 debug* idle 2 3394 38536 16 (null) (null)
737 adev[8-9] 2 batch allocated 2 246 82306 16 (null) (null)
738 adev[10-15] 6 batch idle 2 246 82306 16 (null) (null)
739
740
741 Report only down, drained and draining nodes and their reason field:
742
743 > sinfo -R
744 REASON NODELIST
745 Memory errors dev[0,5]
746 Not Responding dev8
747
748
749
751 Copyright (C) 2002-2007 The Regents of the University of California.
752 Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
753 Copyright (C) 2008-2009 Lawrence Livermore National Security.
754 Copyright (C) 2010-2017 SchedMD LLC.
755
756 This file is part of Slurm, a resource management program. For
757 details, see <https://slurm.schedmd.com/>.
758
759 Slurm is free software; you can redistribute it and/or modify it under
760 the terms of the GNU General Public License as published by the Free
761 Software Foundation; either version 2 of the License, or (at your
762 option) any later version.
763
764 Slurm is distributed in the hope that it will be useful, but WITHOUT
765 ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
766 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
767 for more details.
768
769
771 scontrol(1), smap(1), squeue(1), slurm_load_ctl_conf [22m(3),
772 slurm_load_jobs [22m(3), slurm_load_node [22m(3), slurm_load_partitions (3),
773 slurm_reconfigure [22m(3), slurm_shutdown [22m(3), slurm_update_job [22m(3),
774 slurm_update_node (3), slurm_update_partition (3), slurm.conf(5)
775
776
777
778November 2016 Slurm Commands sinfo(1)