"nice" resource


#1

Hello All,
I have a question about a “nice” resource. How this resource should actually work? This is some kind of priority? I added this resource to the “resources” in sched_config and assigned different “nice” value to each node but it does nothing… I tried “qsub -I -l nice=” and always got the same node. Documentation mention this resource(AG-316, RG-323) but it is too brief.
Thank you


#2

Have you tried adding it to you select line. For example

-l select=1:ncpus=3:nice=2


#3

RG-323: “Can be requested only outside of a select statement.”
$ qsub -I -l select=2:ncpus=1:nice=2
qsub: Resource invalid in “select” specification: nice


#4

The special resource “nice” would not be expected to have an effect on the node selection for the job, it is expected to change the nice process value, though. Here is an example showing how it works (**emphasis in top output mine**):

[user1@centos7 tmp]$ echo "sleep 100000" | qsub
33.centos7.prog.altair.com

[user1@centos7 tmp]$ echo "sleep 100000" | qsub -lnice=10
34.centos7.prog.altair.com


[root@centos7 tmp]# qstat -a

centos7.prog.altair.com:
                                                            Req'd  Req'd   Elap
Job ID          Username Queue    Jobname    SessID NDS TSK Memory Time  S Time
--------------- -------- -------- ---------- ------ --- --- ------ ----- - -----
33.centos7.prog user1    workq    STDIN        4327   1   1    --    --  R 00:01
34.centos7.prog user1    workq    STDIN        4368   1   1    --    --  R 00:01


[root@centos7 tmp]# qstat -a

centos7.prog.altair.com:
                                                            Req'd  Req'd   Elap
Job ID          Username Queue    Jobname    SessID NDS TSK Memory Time  S Time
--------------- -------- -------- ---------- ------ --- --- ------ ----- - -----
33.centos7.prog user1    workq    STDIN        4327   1   1    --    --  R 00:01
34.centos7.prog user1    workq    STDIN        4368   1   1    --    --  R 00:01

Job 33, no nice request:

[root@centos7 tmp]# ps -efj | grep 4327
user1     4327  3265  4327  4327  0 12:50 ?        00:00:00 -bash
user1     4361  4327  4327  4327  0 12:50 ?        00:00:00 -bash
user1     4362  4361  4327  4327  0 12:50 ?        00:00:00 sleep 100000
root      4440  3079  4439  3079  0 12:52 pts/0    00:00:00 grep --color=auto 4327

[root@centos7 tmp]# top -b -p4362 -n1
top - 12:53:43 up 19:13,  3 users,  load average: 0.01, 0.06, 0.06
Tasks:   1 total,   0 running,   1 sleeping,   0 stopped,   0 zombie
%Cpu(s):  0.1 us,  0.1 sy,  0.0 ni, 99.8 id,  0.0 wa,  0.0 hi,  0.0 si,  0.0 st
KiB Mem :  8011144 total,  7038784 free,   391616 used,   580744 buff/cache
KiB Swap:  2183164 total,  2183164 free,        0 used.  7340456 avail Mem

  PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND
 4362 user1     20   **0**  107892    624    532 S   0.0  0.0   0:00.00 sleep

Job 34, nice=10 request:

[root@centos7 tmp]# ps -efj | grep 4368
user1     4368  3265  4368  4368  0 12:50 ?        00:00:00 -bash
user1     4402  4368  4368  4368  0 12:50 ?        00:00:00 -bash
user1     4403  4402  4368  4368  0 12:50 ?        00:00:00 sleep 100000
root      4456  3079  4455  3079  0 12:54 pts/0    00:00:00 grep --color=auto 4368

[root@centos7 tmp]# top -b -p4403 -n1
top - 12:54:13 up 19:14,  3 users,  load average: 0.01, 0.06, 0.06
Tasks:   1 total,   0 running,   1 sleeping,   0 stopped,   0 zombie
%Cpu(s):  0.1 us,  0.1 sy,  0.0 ni, 99.8 id,  0.0 wa,  0.0 hi,  0.0 si,  0.0 st
KiB Mem :  8011144 total,  7039156 free,   391228 used,   580760 buff/cache
KiB Swap:  2183164 total,  2183164 free,        0 used.  7340836 avail Mem

  PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND
 4403 user1     30  **10**  107892    624    532 S   0.0  0.0   0:00.00 sleep

#5

OK, thank you for the answer.


#6

I’ve filed a doc bug for this.