BIND 10 #2410: race condition(?) in dhcp[46]_test.py

[BIND 10 Development]

#2410: race condition(?) in dhcp[46]_test.py
-------------------------------------+-------------------------------------
                   Reporter:  jelte  |                 Owner:  stephen
                       Type:         |                Status:  reviewing
  defect                             |             Milestone:  Sprint-
                   Priority:         |  DHCP-20121129
  medium                             |            Resolution:
                  Component:  dhcp   |             Sensitive:  0
                   Keywords:         |           Sub-Project:  DHCP
            Defect Severity:  N/A    |  Estimated Difficulty:  4
Feature Depending on Ticket:         |           Total Hours:  0
        Add Hours to Ticket:  0      |
                  Internal?:  0      |
-------------------------------------+-------------------------------------
Changes (by marcin):

 * owner:  marcin => stephen


Comment:

 Reviewed commit aac0dd304b86ee38049480c8155f5eb97142c435

 I verified the fix on the VM in the following way...
 - I was running the Debian6 system on VirtualBox: Linux marcin-debian
 2.6.32-5-amd64 #1 SMP Sun Sep 23 10:07:46 UTC 2012 x86_64 GNU/Linux
 - I checked out two versions of bind10 to different folders: master branch
 to one folder and trac2410 with a fix to the other folder
 - I built both versions of bind10 and for each of them I launched the
 DHCPv6 test in the loop with this script in src/bin/dhcp6 folder:
 {{{
 #!/bin/bash
 for i in {1..1000}
 do
     make check
     if [[ $? -ne 0 ]]
     then
         exit 1;
     fi

 done
 }}}
 While script was running I was doing expensive folder copy operation of
 '''Downloads''' folder to '''Downloads2''' in my home folder.

 I hit the issue described in this ticket during the first iteration of the
 test when running it on master branch.

 I made around 30 iterations of this test on trac2410 and did not hit this
 issue. This indicates that the fix works for the certain machine load.

 '''Important:''' this change fixes the test fatal error but it does not
 take away the resource leak error also described in this ticket:
 {{{
 Running command: ../b10-dhcp6 -v
 /home/marcin/devel/bbind10/src/bin/dhcp6/tests/dhcp6_test.py:151:
 ResourceWarning: unclosed file <_io.FileIO name=10 mode='wb'>
 }}}

 However, in the code there is a TODO statement that explains that resource
 warning is actually a false alarm. Is the following line neccessary:
 {{{
 pi = None
 }}}
 ?
 Doesn't this object die when the function ends anyway?

-- 
Ticket URL: <http://bind10.isc.org/ticket/2410#comment:11>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development