Preemption does not happen when a job is qrun

Description

When a queued job is issued qrun. The running job is not preempted, as a result job remains queued. See below:

[root@centos1 ~]# qsub -lselect=1:ncpus=2 – /bin/sleep 999
0.centos1
[root@centos1 ~]# qsub -lselect=1:ncpus=2 – /bin/sleep 999
1.centos1
[root@centos1 ~]# qstat -s

centos1:
Req'd Req'd Elap
Job ID Username Queue Jobname SessID NDS TSK Memory Time S Time
--------------- -------- -------- ---------- ------ — — ------ ----- - -----
0.centos1 root workq STDIN 127034 1 2 – – R 00:00
Job run at Mon Oct 30 at 13:42 on (centos1:ncpus=2)
1.centos1 root workq STDIN – 1 2 – – Q –
Not Running: Not enough free nodes available
[root@centos1 ~]# qrun 1
qrun: Not enough free nodes available 1.centos1
[root@centos1 ~]# qstat -s

centos1:
Req'd Req'd Elap
Job ID Username Queue Jobname SessID NDS TSK Memory Time S Time
--------------- -------- -------- ---------- ------ — — ------ ----- - -----
0.centos1 root workq STDIN 127034 1 2 – – R 00:00
Job run at Mon Oct 30 at 13:42 on (centos1:ncpus=2)
1.centos1 root workq STDIN – 1 2 – – Q –
Not Running: Not enough free nodes available
[root@centos1 ~]# tracejob 1

Job: 1.centos1

10/30/2017 13:42:05 L Considering job to run
10/30/2017 13:42:05 S Job Queued at request of root@centos1.altair.com, owner = root@centos1.altair.com, job name = STDIN, queue = workq
10/30/2017 13:42:05 S Job Modified at request of Scheduler@centos1.altair.com
10/30/2017 13:42:05 L Not enough free nodes available
10/30/2017 13:42:05 S enqueuing into workq, state 1 hop 1
10/30/2017 13:42:05 A queue=workq
10/30/2017 13:42:11 L Received qrun request
10/30/2017 13:42:11 L Considering job to run
10/30/2017 13:42:11 L Not enough free nodes available
[root@centos1 ~]# tracejob 0

Job: 0.centos1

10/30/2017 13:42:04 L Considering job to run
10/30/2017 13:42:04 S Job Queued at request of root@centos1.altair.com, owner = root@centos1.altair.com, job name = STDIN, queue = workq
10/30/2017 13:42:04 S Job Run at request of Scheduler@centos1.altair.com on exec_vnode (centos1:ncpus=2)
10/30/2017 13:42:04 L Job run
10/30/2017 13:42:04 M Started, pid = 127034
10/30/2017 13:42:04 S enqueuing into workq, state 1 hop 1
10/30/2017 13:42:04 A queue=workq
10/30/2017 13:42:04 A user=root group=root project=_pbs_project_default jobname=STDIN queue=workq ctime=1509351124 qtime=1509351124 etime=1509351124 start=1509351124 exec_host=centos1/0*2 exec_vnode=(centos1:ncpus=2)
Resource_List.ncpus=2 Resource_List.nodect=1 Resource_List.place=free Resource_List.select=1:ncpus=2 resource_assigned.ncpus=2
[root@centos1 ~]#

Acceptance Criteria

None

Activity

Show:
Varun Sonkar
October 30, 2017, 8:18 AM

Added Watchers.

Assignee

Bhroam Mann

Reporter

Varun Sonkar

Severity

3-High

OS

None

Start Date

None

Pull Request URL

None

Story Points

1

Components

Fix versions

Affects versions

Priority

High
Configure