pbs_python-4.6.0 memory leak

[aroudgar@…]

Hi,

I am using pbs_python-4.6.0 with torque-5.1.1.2. We found a memory leak for this module. Simply run the following program and hopefully you can reproduce this issue. It gradually takes more and more memory during the run until it takes all available physical memory of our server and crashes.

We also used torque-5.1.0 but the result is the same. Any suggestion or advice would be gratefully appreciated.

import pbs
import time
server_name = pbs.pbs_default()
print "Sername name",server_name
c = pbs.pbs_connect(server_name)
attropl = pbs.new_attropl(1)
attropl[0].name  = pbs.ATTR_N
attropl[0].value = 'test'
job_id = pbs.pbs_submit(c, attropl, 'A1.tsk', '', 'NULL')
e, e_txt = pbs.error()
if e:
     print e,e_txt
print job_id
pbs.pbs_disconnect( c )
j=0
while 1:
     time.sleep( 1 )
     j=j+1
     if j >= 200:
       break
     c = pbs.pbs_connect(server_name)
     stat_attrl = pbs.new_attrl(3)
     stat_attrl[0].name = pbs.ATTR_state
     stat_attrl[1].name = pbs.ATTR_used
     stat_attrl[2].name = pbs.ATTR_exitstat
     jobs = pbs.pbs_statjob( c, None, stat_attrl, None )
     pbs.pbs_disconnect( c )

Cheers, Ata

[bas]

I will test it thanks for the report