PBS launched processes inherit extra fd (pipe) which makes dmtcp fail

Description

None

Acceptance Criteria

DMTCP is a open source distributed checkpointing framework. When used with PBS, the restart functionality does not work since processes launched under pbs inherit an extra fd (to a pipe) which makes dmtcp fail.

Possibly the pbs launcher forgot to set close on exec flag on the pipe.

[root@centosvm subhas]# lsof | grep 31108
lsof: WARNING: can't stat() fuse.gvfsd-fuse file system /run/user/1000/gvfs
Output information may be incomplete.
tst 31108 root cwd DIR 253,0 4096 201326721 /root
tst 31108 root rtd DIR 253,0 4096 128 /
tst 31108 root txt REG 253,0 8503 202458348 /root/tst
tst 31108 root mem REG 253,0 2112384 201328784 /usr/lib64/libc-2.17.so
tst 31108 root mem REG 253,0 164440 208542119 /usr/lib64/ld-2.17.so
tst 31108 root 0u CHR 136,1 0t0 4 /dev/pts/1
tst 31108 root 1u CHR 136,1 0t0 4 /dev/pts/1
tst 31108 root 2u CHR 136,1 0t0 4 /dev/pts/1

[root@centosvm subhas]# lsof | grep 31323
lsof: WARNING: can't stat() fuse.gvfsd-fuse file system /run/user/1000/gvfs
Output information may be incomplete.
tst 31323 root cwd DIR 253,0 4096 201326721 /root
tst 31323 root rtd DIR 253,0 4096 128 /
tst 31323 root txt REG 253,0 8503 202458348 /root/tst
tst 31323 root mem REG 253,0 2112384 201328784 /usr/lib64/libc-2.17.so
tst 31323 root mem REG 253,0 164440 208542119 /usr/lib64/ld-2.17.so
tst 31323 root 0r CHR 1,3 0t0 1028 /dev/null
tst 31323 root 1u REG 253,0 0 202897681 /var/spool/PBS/spool/22.centosvm.OU
tst 31323 root 2u REG 253,0 0 202898713 /var/spool/PBS/spool/22.centosvm.ER
tst 31323 root 12r FIFO 0,8 0t0 175941 pipe

The second lsof has the extra last line - the pipe

Status

Assignee

Unassigned

Reporter

Subhasis Bhattacharya

Severity

None

OS

None

Start Date

None

Pull Request URL

None

Story Points

1

Components

Fix versions

Affects versions

14.0.0

Priority

Low
Configure