1sprio(1) Slurm Commands sprio(1)
2
3
4
6 sprio - view the factors that comprise a job's scheduling priority
7
8
10 sprio [OPTIONS...]
11
12
14 sprio is used to view the components of a job's scheduling priority
15 when the multi-factor priority plugin is installed. sprio is a read-
16 only utility that extracts information from the multi-factor priority
17 plugin. By default, sprio returns information for all pending jobs.
18 Options exist to display specific jobs by job ID and user name.
19
20
22 --federation
23 Show jobs in federation if a member of one.
24
25
26 -h, --noheader
27 Do not print a header on the output.
28
29
30 --help Print a help message describing all options sprio.
31
32
33 -j <job_id_list>, --jobs=<job_id_list>
34 Requests a comma separated list of job ids to display. Defaults
35 to all jobs. Since this option's argument is optional, for
36 proper parsing the single letter option must be followed immedi‐
37 ately with the value and not include a space between them. For
38 example "-j1008,1009" and not "-j 1008,1009".
39
40
41 --local
42 Show only jobs local to this cluster. Ignore other clusters in
43 this federation (if any). Overrides --federation.
44
45
46 -l, --long
47 Report more of the available information for the selected jobs.
48
49
50 -M, --clusters=<string>
51 The cluster to issue commands to. Only one cluster name may be
52 specified. Note that the SlurmDBD must be up for this option to
53 work properly. This option implicitly sets the --local option.
54
55
56 -n, --norm
57 Display the normalized priority factors for the selected jobs.
58
59
60 -o <output_format>, --format=<output_format>
61 Specify the information to be displayed, its size and position
62 (right or left justified). The default formats when all factors
63 have been assigned non-zero weights are
64
65
66 default "%.15i %9r %.10Y %.10S %.10A %.10B %.10F %.10J
67 %.10P %.10Q %20T"
68
69 -l, --long "%.15i %9r %.8u %.10Y %.10S %.10A %.10B %.10F
70 %.10J %.10P %.10Q %.11N %.20T";
71
72
73 The format of each field is "%[.][size]type".
74
75 size is the minimum field size. If no size is specified,
76 whatever is needed to print the information will be
77 used.
78
79 . indicates the output should be left justified. By
80 default, output is right justified.
81
82
83 Valid type specifications include:
84
85
86 %a Normalized age priority
87
88 %A Weighted age priority
89
90 %b Normalized association priority
91
92 %B Weighted association priority
93
94 %c Cluster name. Only applicable for federated clusters
95
96 %f Normalized fair-share priority
97
98 %F Weighted fair-share priority
99
100 %i Job ID
101
102 %j Normalized job size priority
103
104 %J Weighted job size priority
105
106 %N Nice adjustment
107
108 %p Normalized partition priority
109
110 %P Weighted partition priority
111
112 %q Normalized quality of service priority
113
114 %Q Weighted quality of service priority
115
116 %r Partition name
117
118 %S Weighted admin priority.
119
120 %t Normalized TRES priorities
121
122 %T Weighted TRES priorities
123
124 %u User name for a job
125
126 %Y Job priority
127
128 %y Normalized job priority
129
130
131 -S <sort_list>, --sort=<sort_list>
132 Specification of the order in which jobs should be reported.
133 This uses the same field specification as <output_format>. Mul‐
134 tiple sorts may be performed by listing multiple sort fields
135 separated by commas. The field specifications may be preceded by
136 "+" or "-" for ascending (default) or descending respectively.
137 For example, a <sort_list> of "u,r,-y" will sort the job prior‐
138 ity reports by username, partition name, and descending job pri‐
139 ority, in that order. The default <sort_list> is "i" (ascending
140 job id).
141
142
143 -p <partition_list>, --partition=<partition_list>
144 Requests a comma separated list of partitions to display.
145 Defaults to all partitions.
146
147
148 --sibling
149 Show all sibling jobs on a federated cluster. Without this
150 option in a federated cluster, each job in each partition will
151 have its priority and its components reported for only one clus‐
152 ter. Each sibling job on the various clusters in the federation
153 may have different priority, which will not be reported without
154 using this option. Implicitly adds "%c" (cluster name) to the
155 output format.
156
157
158 -u <user_list>, --user=<user_list>
159 Request jobs from a comma separated list of users. The list can
160 consist of user names or user id numbers.
161
162
163 --usage
164 Print a brief help message listing the sprio options.
165
166
167 -v, --verbose
168 Report details of sprios actions.
169
170
171 -V , --version
172 Print version information and exit.
173
174
175 -w , --weights Display the configured weights for each
176 factor. This is for information purposes only. Actual job data
177 is suppressed.
178
179
181 Executing sprio sends a remote procedure call to slurmctld. If enough
182 calls from sprio or other Slurm client commands that send remote proce‐
183 dure calls to the slurmctld daemon come in at once, it can result in a
184 degradation of performance of the slurmctld daemon, possibly resulting
185 in a denial of service.
186
187 Do not run sprio or other Slurm client commands that send remote proce‐
188 dure calls to slurmctld from loops in shell scripts or other programs.
189 Ensure that programs limit calls to sprio to the minimum necessary for
190 the information you are trying to gather.
191
192
194 If no corresponding commandline option is specified, sprio will use the
195 value of the following environment variables.
196
197 SLURM_CLUSTERS Same as --clusters
198
199
200 SLURM_CONF The location of the Slurm configuration file.
201
202
203 SPRIO_FEDERATION Same as --federation
204
205
206 SPRIO_FORMAT Same as -o <output_format>, --format=<output_for‐
207 mat>
208
209
210 SPRIO_LOCAL Same as --local
211
212
213 SPRIO_SIBLING Same as --sibling
214
215
217 Print the list of all pending jobs with their weighted priorities
218 > sprio
219 JOBID PRIORITY AGE FAIRSHARE JOBSIZE PARTITION
220 QOS
221 65539 62664 0 51664 1000 10000
222 0
223 65540 62663 0 51663 1000 10000
224 0
225 65541 62662 0 51662 1000 10000
226 0
227
228 Print the list of all pending jobs with their normalized priorities
229 > sprio -n
230 JOBID PRIORITY AGE FAIRSHARE JOBSIZE PARTITION QOS
231 65539 0.00001459 0.0007180 0.5166470 1.0000000 1.0000000
232 0.0000000
233 65540 0.00001459 0.0007180 0.5166370 1.0000000 1.0000000
234 0.0000000
235 65541 0.00001458 0.0007180 0.5166270 1.0000000 1.0000000
236 0.0000000
237
238 Print the job priorities for specific jobs
239 > sprio --jobs=65548,65547
240 JOBID PRIORITY AGE FAIRSHARE JOBSIZE PARTITION
241 QOS
242 65547 62078 0 51078 1000 10000
243 0
244 65548 62077 0 51077 1000 10000
245 0
246
247 Print the job priorities for jobs of specific users
248 > sprio --users=fred,sally
249 JOBID USER PRIORITY AGE FAIRSHARE JOBSIZE PARTITION
250 QOS
251 65548 fred 62079 1 51077 1000 10000
252 0
253 65549 sally 62080 1 51078 1000 10000
254 0
255
256 Print the configured weights for each priority component
257 > sprio -w
258 JOBID PRIORITY AGE FAIRSHARE JOBSIZE PARTITION
259 QOS
260 Weights 1000 100000 1000 10000
261 1
262
263
265 Copyright (C) 2009 Lawrence Livermore National Security. Produced at
266 Lawrence Livermore National Laboratory (cf, DISCLAIMER).
267 Copyright (C) 2010-2017 SchedMD LLC.
268
269 This file is part of Slurm, a resource management program. For
270 details, see <https://slurm.schedmd.com/>.
271
272 Slurm is free software; you can redistribute it and/or modify it under
273 the terms of the GNU General Public License as published by the Free
274 Software Foundation; either version 2 of the License, or (at your
275 option) any later version.
276
277 Slurm is distributed in the hope that it will be useful, but WITHOUT
278 ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
279 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
280 for more details.
281
283 squeue(1), sshare(1)
284
285
286
287September 2018 Slurm Commands sprio(1)