10
10
\fBscontrol\fR is used to view or modify Slurm configuration including: job,
11
job step, node, partition, and overall system configuration. Most of the
12
commands can only be executed by user root. If an attempt to view or modify
11
job step, node, partition, reservation, and overall system configuration. Most
12
of the commands can only be executed by user root. If an attempt to view or modify
13
13
configuration information is made by an unauthorized user, an error message
14
14
will be printed and the requested action will not occur. If no command is
15
15
entered on the execute line, \fBscontrol\fR will operate in an interactive
16
16
mode and prompt for input. It will continue prompting for input and executing
17
17
commands until explicitly terminated. If a command is entered on the execute
18
18
line, \fBscontrol\fR will execute that command and terminate. All commands
19
and options are case\-insensitive, although node names and partition names
20
are case\-sensitive (node names "LX" and "lx" are distinct). Commands can
21
be abbreviated to the extent that the specification is unique.
19
and options are case\-insensitive, although node names, partition names, and
20
reservation names are case\-sensitive (node names "LX" and "lx" are distinct).
21
All commands and options can be abbreviated to the extent that the
22
specification is unique.
84
85
\fIerror\fP (report the result for the last checkpoint request, error code and message)
86
87
\fIrestart\fP (restart execution of the previously checkpointed job steps)
89
Acceptable values for \fICKPT_OP\fP include:
91
\fIMaxWait=<seconds>\fP maximum time for checkpoint to be written.
92
Default value is 10 seconds.
93
Valid with \fIcreate\fP and \fIvacate\fP options only.
95
\fIImageDir=<directory_name>\fP Location of checkpoint file.
96
Valid with \fIcreate\fP, \fIvacate\fP and \fIrestart\fP options only.
97
This value takes precedent over any \-\-checkpoint\-dir value specified
98
at job submission time.
100
\fIStickToNodes\fP If set, resume job on the same nodes are previously used.
101
Valid with the \fIrestart\fP option only.
105
\fBcreate\fP \fISPECIFICATION\fP
106
Create a new partition or reservation. See the full list of parameters
107
below. Include the tag "res" to create a reservation without specifying
91
112
Display all jobs in a COMPLETING state along with associated nodes in either a
92
113
COMPLETING or DOWN state.
196
217
\fBshow\fP \fIENTITY\fP \fIID\fP
197
218
Display the state of the specified entity with the specified identification.
198
219
\fIENTITY\fP may be \fIconfig\fP, \fIdaemons\fP, \fIjob\fP, \fInode\fP,
199
\fIpartition\fP, \fIslurmd\fP, \fIstep\fP, \fIhostlist\fP or \fIhostnames\fP
220
\fIpartition\fP, \fIreservation\fP, \fIslurmd\fP, \fIstep\fP, \fItopology\fP,
221
\fIhostlist\fP or \fIhostnames\fP
200
222
(also \fIblock\fP or \fIsubbp\fP on BlueGene systems).
201
223
\fIID\fP can be used to identify a specific element of the identified
202
224
entity: the configuration parameter name, job ID, node name, partition name,
203
or job step ID \fIconfig\fP, \fIjob\fP, \fInode\fP, \fIpartition\fP,
204
or \fIstep\fP respectively.
225
reservation name, or job step ID for \fIconfig\fP, \fIjob\fP, \fInode\fP,
226
\fIpartition\fP, or \fIstep\fP respectively.
227
For an \fIENTITY\fP of \fItopology\fP, the \fIID\fP may be a node or switch name.
228
If one node name is specified, all switches connected to that node (and
229
their parent switches) will be shown.
230
If more than one node name is specified, only switches that connect to all
231
named nodes will be shown.
205
232
\fIhostnames\fP takes an optional hostlist expression as input and
206
233
writes a list of individual host names to standard output (one per
207
234
line). If no hostlist expression is supplied, the contents of the
208
SLURM_NODELIST environment variable is used. For example "tux[1-3]"
235
SLURM_NODELIST environment variable is used. For example "tux[1\-3]"
209
236
is mapped to "tux1","tux2" and "tux3" (one hostname per line).
210
237
\fIhostlist\fP takes a list of host names and prints the hostlist
211
238
expression for them (the inverse of \fIhostnames\fP).
237
264
Not all architectures and configurations support job suspension.
268
Instruct SLURM's backup controller (slurmctld) to take over system control.
269
SLURM's backup controller requests control from the primary and waits for
270
its termination. After that, it switches from backup mode to controller
271
mode. If primary controller can not be contacted, it directly switches to
272
controller mode. This can be used to speed up the SLURM controller
273
fail\-over mechanism when the primary node is down.
274
This can be used to minimize disruption if the computer executing the
275
primary SLURM controller is scheduled down.
276
(Note: SLURM's primary controller will take the control back at startup.)
240
279
\fBupdate\fP \fISPECIFICATION\fP
241
Update job, node or partition configuration per the supplied specification.
242
\fISPECIFICATION\fP is in the same format as the Slurm configuration file
243
and the output of the \fIshow\fP command described above. It may be desirable
244
to execute the \fIshow\fP command (described above) on the specific entity
245
you which to update, then use cut\-and\-paste tools to enter updated configuration
246
values to the \fIupdate\fP. Note that while most configuration values can be
247
changed using this command, not all can be changed using this mechanism. In
248
particular, the hardware configuration of a node or the physical addition or
249
removal of nodes from the cluster may only be accomplished through editing
250
the Slurm configuration file and executing the \fIreconfigure\fP command
280
Update job, node, partition, or reservation configuration per the supplied
281
specification. \fISPECIFICATION\fP is in the same format as the Slurm
282
configuration file and the output of the \fIshow\fP command described above. It
283
may be desirable to execute the \fIshow\fP command (described above) on the
284
specific entity you which to update, then use cut\-and\-paste tools to enter
285
updated configuration values to the \fIupdate\fP. Note that while most
286
configuration values can be changed using this command, not all can be changed
287
using this mechanism. In particular, the hardware configuration of a node or
288
the physical addition or removal of nodes from the cluster may only be
289
accomplished through editing the Slurm configuration file and executing the \fIreconfigure\fP command (described above).
264
302
Repeat the last command executed.
267
\fBSPECIFICATIONS FOR SHOW AND UPDATE COMMANDS, JOBS\fR
305
\fBSPECIFICATIONS FOR UPDATE COMMAND, JOBS\fR
269
307
\fIAccount\fP=<account>
270
308
Account name to be changed for this job's resource use.
271
309
Value may be cleared with blank data value, "Account=".
311
\fIConnection\fP=<type>
312
Reset the node connection type.
313
Possible values on Blue Gene are "MESH", "TORUS" and "NAV"
273
316
\fIContiguous\fP=<yes|no>
274
317
Set the job's requirement for contiguous (consecutive) nodes to be allocated.
275
Possible values are"YES" and "NO".
318
Possible values are "YES" and "NO".
277
320
\fIDependency\fP=<job_id>
278
321
Defer job's initiation until specified job_id completes.
279
Cancel dependency with job_id value of "0", "Depedency=0".
322
Cancel dependency with job_id value of "0", "Dependency=0".
324
\fIEligibleTime\fP=<time_spec>
325
Equivalent to \fIStartTime\fP.
281
327
\fIExcNodeList\fP=<nodes>
282
328
Set the job's list of excluded node. Multiple node names may be
290
336
it was terminated by a signal.
292
338
\fIFeatures\fP=<features>
293
Set the job's required features on nodes specified value. Multiple values
339
Set the job's required node features. Multiple values
294
340
may be comma separated if all features are required (AND operation) or
295
341
separated by "|" if any of the specified features are required (OR operation).
296
342
Value may be cleared with blank data value, "Features=".
345
Reset the required job geometry.
346
On Blue Gene the value should be three digits separated by
347
"x" or ",". The digits represent the allocation size in
348
X, Y and Z dimentions (e.g. "2x3x4").
350
\fIRotate\fP=<yes|no>
351
Permit the job's geometry to be rotated.
352
Possible values are "YES" and "NO".
299
355
Identify the job to be updated. This specification is required.
301
357
\fIMinCores\fP=<count>
302
358
Set the job's minimum number of cores per socket to the specified value.
304
\fIMinMemory\fP=<megabytes>
305
Set the job's minimum real memory required per nodes to the specified value.
360
\fIMinMemoryCPU\fP=<megabytes>
361
Set the job's minimum real memory required per allocated CPU to the specified
363
Either \fIMinMemoryCPU\fP or \fIMinMemoryNode\fP may be set, but not both.
365
\fIMinMemoryNode\fP=<megabytes>
366
Set the job's minimum real memory required per node to the specified value.
367
Either \fIMinMemoryCPU\fP or \fIMinMemoryNode\fP may be set, but not both.
307
369
\fIMinProcs\fP=<count>
308
Set the job's minimum number of processors per nodes to the specified value.
370
Set the job's minimum number of processors per node to the specified value.
310
372
\fIMinSockets\fP=<count>
311
373
Set the job's minimum number of sockets per node to the specified value.
314
376
Set the job's minimum number of threads per core to the specified value.
316
378
\fIMinTmpDisk\fP=<megabytes>
317
Set the job's minimum temporary disk space required per nodes to the specified value.
379
Set the job's minimum temporary disk space required per node to the specified value.
319
381
\fIName\fP=<name>
320
382
Set the job's name to the specified value.
385
Adjust job's priority by the specified value. Default value is 100.
386
The adjustment range is from \-10000 (highest priority)
387
to 10000 (lowest priority).
388
Nice value changes are not additive, but overwrite any prior nice
389
value and are applied to the job's base priority.
390
Only privileged users can specify a negative adjustment.
322
392
\fIPartition\fP=<name>
323
393
Set the job's partition to the specified value.
327
397
Note that a job priority of zero prevents the job from ever being scheduled.
328
398
By setting a job's priority to zero it is held.
329
399
Set the priority to a non\-zero value to permit it to run.
332
Adjust job's priority by the specified value. Default value is 100.
334
\fIReqProcs\fP=<count>
335
Set the job's count of required processes to the specified value.
400
Explicitly setting a job's priority clears any previously set nice value.
402
\fIReqCores\fP=<count>
403
Set the job's count of required cores to the specified value.
337
405
\fIReqNodeList\fP=<nodes>
338
406
Set the job's list of required node. Multiple node names may be specified using
339
407
simple node range expressions (e.g. "lx[10\-20]").
340
408
Value may be cleared with blank data value, "ReqNodeList=".
342
\fIReqNodes\fP=<min_count>[-<max_count>]
410
\fIReqNodes\fP=<min_count>[\-<max_count>]
343
411
Set the job's minimum and optionally maximum count of nodes to be allocated.
413
\fIReqProcs\fP=<count>
414
Set the job's count of required processes to the specified value.
345
416
\fIReqSockets\fP=<count>
346
417
Set the job's count of required sockets to the specified value.
348
\fIReqCores\fP=<count>
349
Set the job's count of required cores to the specified value.
351
419
\fIReqThreads\fP=<count>
352
420
Set the job's count of required threads to the specified value.
364
432
\fIteatime\fR (4pm) and you can have a time\-of\-day suffixed
365
433
with \fIAM\fR or \fIPM\fR for running in the morning or the evening.
366
434
You can also say what day the job will be run, by specifying
367
a date of the form \fIMMDDYY\fR or \fIMM/DD/YY\fR
368
or \fIMM.DD.YY\fR. You can also
435
a date of the form \fIMMDDYY\fR or \fIMM/DD/YY\fR or \fIMM.DD.YY\fR,
436
or a date and time as \fIYYYY\-MM\-DD[THH[:MM[:SS]]]\fR. You can also
369
437
give times like \fInow + count time\-units\fR, where the time\-units
370
438
can be \fIminutes\fR, \fIhours\fR, \fIdays\fR, or \fIweeks\fR
371
439
and you can tell SLURM to run the job today with the keyword
372
440
\fItoday\fR and to run the job tomorrow with the keyword
376
443
\fITimeLimit\fP=<time>
377
444
The job's time limit.
423
474
\fIState\fP=<state>
424
475
Identify the state to be assigned to the node. Possible values are "NoResp",
425
"ALLOC", "ALLOCATED", "DOWN", "DRAIN", "FAIL", "FAILING", "IDLE" or "RESUME".
426
"RESUME is not an actual node state, but will return a DRAINED, DRAINING,
427
or DOWN node to service, either IDLE or ALLOCATED state as appropriate.
428
Setting a node "DOWN" will cause all running and suspended jobs on that
429
node to be terminated.
476
"ALLOC", "ALLOCATED", "DOWN", "DRAIN", "FAIL", "FAILING", "IDLE", "MAINT",
477
"POWER_DOWN", "POWER_UP", or "RESUME".
430
478
If you want to remove a node from service, you typically want to set
431
479
it's state to "DRAIN".
432
480
"FAILING" is similar to "DRAIN" except that some applications will
433
481
seek to relinquish those nodes before the job completes.
482
"RESUME" is not an actual node state, but will return a "DRAINED", "DRAINING",
483
or "DOWN" node to service, either "IDLE" or "ALLOCATED" state as appropriate.
484
Setting a node "DOWN" will cause all running and suspended jobs on that
485
node to be terminated.
486
"POWER_DOWN" and "POWER_UP" will use the configured \fISuspendProg\fR and
487
\fIResumeProg\fR programs to explicitly place a node in or out of a power
434
489
The "NoResp" state will only set the "NoResp" flag for a node without
435
490
changing its underlying state.
436
491
While all of the above states are valid, some of them are not valid new
438
493
Generally only "DRAIN", "FAIL" and "RESUME" should be used.
441
\fBSPECIFICATIONS FOR UPDATE AND DELETE COMMANDS, PARTITIONS\fR
496
\fIWeight\fP=<weight>
497
Identify weight to be associated with specified nodes. This allows
498
dynamic changes to weight associated with nodes, which will be used
499
for the subsequent node allocation decisions.
500
Any previously identified weight will be overwritten with the new value.\fBNOTE:\fR The \fIWeight\fP associated with nodes will be reset to
501
the values specified in slurm.conf (if any) upon slurmctld restart
503
Update slurm.conf with any changes meant to be persistent.
506
\fBSPECIFICATIONS FOR CREATE, UPDATE, AND DELETE COMMANDS, PARTITIONS\fR
443
508
\fIAllowGroups\fP=<name>
444
509
Identify the user groups which may use this partition.
513
578
must be explicitly canceled to force their termination.
581
\fBSPECIFICATIONS FOR CREATE, UPDATE, AND DELETE COMMANDS, RESERVATIONS\fR
585
\fIReservation\fP=<name>
586
Identify the name of the reservation to be created, updated, or deleted.
587
This parameter is required for update and is the only parameter for delete.
588
For create, if you do not want to give a reservation name, use
589
"scontrol create res ..." and a name will be created automatically.
593
Identify number of nodes to be reserved. A new reservation must specify either
598
Identify the node(s) to be reserved. Multiple node names
599
may be specified using simple node range expressions (e.g. "Nodes=lx[10\-20]").
600
Specify a blank data value to remove all nodes from a reservation: "Nodes=".
601
A new reservation must specify either NodeCnt or Nodes.
604
\fIStartTime\fP=<time_spec>
605
The start time for the reservation. A new reservation must specify a start
606
time. It accepts times of the form \fIHH:MM:SS\fR for
607
a specific time of day (seconds are optional).
608
(If that time is already past, the next day is assumed.)
609
You may also specify \fImidnight\fR, \fInoon\fR, or
610
\fIteatime\fR (4pm) and you can have a time\-of\-day suffixed
611
with \fIAM\fR or \fIPM\fR for running in the morning or the evening.
612
You can also say what day the job will be run, by specifying
613
a date of the form \fIMMDDYY\fR or \fIMM/DD/YY\fR or \fIMM.DD.YY\fR,
614
or a date and time as \fIYYYY\-MM\-DD[THH[:MM[:SS]]]\fR. You can also
615
give times like \fInow + count time\-units\fR, where the time\-units
616
can be \fIminutes\fR, \fIhours\fR, \fIdays\fR, or \fIweeks\fR
617
and you can tell SLURM to run the job today with the keyword
618
\fItoday\fR and to run the job tomorrow with the keyword
622
\fIEndTime\fP=<time_spec>
623
The end time for the reservation. A new reservation must specify an end
624
time or a duration. Valid formats are the same as for StartTime.
627
\fIDuration\fP=<time>
628
The length of a reservation. A new reservation must specify an end
629
time or a duration. Valid formats are minutes, minutes:seconds,
630
hours:minutes:seconds, days\-hours, days\-hours:minutes,
631
days\-hours:minutes:seconds, or UNLIMITED. Time resolution is one minute and
632
second values are rounded up to the next minute.
635
\fIPartitionName\fP=<name>
636
Identify the partition to be reserved.
640
Flags associated with the reservation.
641
In order to remove a flag with the update option, preceed the name with
642
a minus sign. For example: Flags=\-MAINT.
643
Currently supported flags include:
647
Maintenance mode, receives special accounting treatment
650
Ignore currently running jobs when creating the reservation.
651
This can be especially useful when reserving all nodes in the system for maintenance
654
Repeat the reservation at the same time every day
657
Repeat the reservation at the same time every week
660
Reservation is for specific nodes (output only)
664
\fIFeatures\fP=<features>
665
Set the reservation's required node features. Multiple values
666
may be comma separated if all features are required (AND operation) or
667
separated by "|" if any of the specified features are required (OR operation).
668
Value may be cleared with blank data value, "Features=".
671
\fIUsers\fP=<user list>
672
List of users permitted to use the reserved nodes.
673
E.g. Users=jones1,smith2.
674
A new reservation must specify Users and/or Accounts.
677
\fIAccounts\fP=<account list>
678
List of accounts permitted to use the reserved nodes.
679
E.g. Accounts=physcode1,physcode2. A user in any of the accounts
680
may use the reserved nodes.
681
A new reservation must specify Users and/or Accounts.
516
684
\fBSPECIFICATIONS FOR UPDATE, BLOCK \fR
518
686
Bluegene systems only!
612
788
\fBscancel\fR(1), \fBsinfo\fR(1), \fBsqueue\fR(1),
613
789
\fBslurm_checkpoint\fR(3),
790
\fBslurm_create_partition\fR(3),
614
791
\fBslurm_delete_partition\fR(3),
615
792
\fBslurm_load_ctl_conf\fR(3),
616
793
\fBslurm_load_jobs\fR(3), \fBslurm_load_node\fR(3),
617
794
\fBslurm_load_partitions\fR(3),
618
795
\fBslurm_reconfigure\fR(3), \fBslurm_requeue\fR(3), \fBslurm_resume\fR(3),
619
796
\fBslurm_shutdown\fR(3), \fBslurm_suspend\fR(3),
797
\fBslurm_takeover\fR(3),
620
798
\fBslurm_update_job\fR(3), \fBslurm_update_node\fR(3),
621
799
\fBslurm_update_partition\fR(3),
622
800
\fBslurm.conf\fR(5)