starting "pbs_mom -p" with running job results in SIGSEGV

Description

steps to reproduce:
1) start pbs_mom
2) start a job on the mom
3) stop pbs_mom
4) start "pbs_mom -p" and wait a few secs
5) following coredump is produced:

Core was generated by `pbs_mom -p -N'.
Program terminated with signal SIGSEGV, Segmentation fault.
#0 0x000000000044f783 in mom_over_limit (pjob=0x110a340) at mom_main.c:6986
6986 walltime_sum = preswalltime->rs_value.at_val.at_long;
(gdb) info stack
#0 0x000000000044f783 in mom_over_limit (pjob=0x110a340) at mom_main.c:6986
#1 0x000000000044ff6b in job_over_limit (pjob=0x110a340) at mom_main.c:7168
#2 0x00000000004543f7 in main (argc=3, argv=0x7fff37b20c68) at mom_main.c:9865
(gdb) print preswalltime
$1 = (resource *) 0x0
(gdb)

Acceptance Criteria

None

Activity

Show:
Václav Chlumský
November 20, 2017, 1:57 PM

Manual test:
-----------------------------------------------------------
TERMINAL1 (start pbs_mom):
(JESSIE)root@torque4:~# pbs_mom -p -N

TERINAL2 (start a job):
(JESSIE)vchlum@torque4:~$ qsub -I
qsub: waiting for job 849.torque4.ics.muni.cz to start
qsub: job 849.torque4.ics.muni.cz ready

(JESSIE)vchlum@torque4:~$

TERMINAL1 (Ctrl+C, mom is terminated):
^C(JESSIE)root@torque4:~#

TERMINAL1 (start pbs_mom again):
(JESSIE)root@torque4:~# pbs_mom -p -N
-----------------------------------------------------------
Both pbs_mom and job 849 remain running and used walltime is growing:
(JESSIE)vchlum@torque4:~$ qstat 849 -f | grep walltime
resources_used.walltime = 00:02:30
(JESSIE)vchlum@torque4:~$

Assignee

Unassigned

Reporter

Václav Chlumský

Severity

None

OS

None

Start Date

None

Pull Request URL

None

Story Points

1

Fix versions

Affects versions

Priority

Critical
Configure