Bug 471902

This has been fixed for Condor 7.2.0.  --Dan

Thanks Dan.

The fix will appear in 7.2.0-0.3

[root@north-13 ~]# ll /var/lib/condor/condor_master.pid 
ls: /var/lib/condor/condor_master.pid: No such file or directory
[root@north-13 ~]# service condor start
Starting Condor daemons:                                   [  OK  ]
[root@north-13 ~]# ll /var/lib/condor/condor_master.pid 
-rw-r--r-- 1 condor condor 6 Dec 11 13:10 /var/lib/condor/condor_master.pid
[root@north-13 ~]# condor_restart -master
Sent "Restart" command to local master
[root@north-13 ~]# ll /var/lib/condor/condor_master.pid 
-rw-r--r-- 1 condor condor 6 Dec 11 13:10 /var/lib/condor/condor_master.pid
[root@north-13 ~]# ps auxwww | grep condor_master
condor   13555  0.0  0.0 113596  7108 ?        Ssl  13:10   0:00 condor_master -pidfile /var/lib/condor/condor_master.pid
root     13806  0.0  0.0  61172   728 pts/1    S+   13:11   0:00 grep condor_master
[root@north-13 ~]# service condor stop
Stopping Condor daemons:                                   [  OK  ]
[root@north-13 ~]# ps auxwww | grep condor_master
root     13824  0.0  0.0  61172   728 pts/1    S+   13:12   0:00 grep condor_master
[root@north-13 ~]# ll /var/lib/condor/condor_master.pid 
ls: /var/lib/condor/condor_master.pid: No such file or directory
[root@north-13 ~]#

the pidfile abandoning issue comes about when the master restarts itself

the master monitors the executables, on disk, from DAEMON_LIST. when they are new it will re'exec the daemon. it even does this for itself. in the past it wouldn't keep its argv when re'execing, so it would not clean up. so, still more testing to do on 471902

Everything seems OK. PID file gets cleaned up after a self-restart followed by a shutdown sometime later. However the PID never changes since the master is restarted with execv.

"Like all of the exec functions, execv replaces the calling process image with a new process image. This has the effect of running a new progam with the process ID of the calling process. Note that a new process is not started; the new process image simply overlays the original process image. The execv function is most commonly used to overlay a process image that has been created by a call to the fork  function."

[root@north-13 condor]# cat condor_master.pid
17079
[root@north-13 condor]# !ps
ps -ef | grep condor_master
condor   17079     1  0 16:25 ?        00:00:00 condor_master -pidfile /var/lib/condor/condor_master.pid
[root@north-13 condor]# grep modified log/MasterLog 
12/11 16:31:00 /usr/sbin/condor_master was modified, restarting /usr/sbin/condor_master.
[root@north-13 condor]# !ps
ps -ef | grep condor_master
condor   17079     1  0 16:25 ?        00:00:00 condor_master -pidfile /var/lib/condor/condor_master.pid
[root@north-13 condor]# !cat
cat condor_master.pid
17079
[root@north-13 condor]# service condor stop
Stopping Condor daemons:                                   [  OK  ]
[root@north-13 condor]# ll
total 56
-rw-r--r-- 1 root   root    400 Nov 14 11:43 condor_config.local
-rw-r--r-- 1 root   root    106 Nov 24 18:07 condor_config.overrides
-rw-r--r-- 1 root   root     61 Nov 21 23:07 condor_config.overrides~
drwxr-xr-x 2 condor condor 4096 Dec 11 16:33 execute
drwxr-xr-x 2 root   root   4096 Dec  4 15:21 feature_configs
drwxr-xr-x 2 condor condor 4096 Dec 11 16:43 log
drwxr-xr-x 2 condor condor 4096 Dec 10 09:29 spool

An advisory has been issued which should help the problem
described in this bug report. This report is therefore being
closed with a resolution of ERRATA. For more information
on therefore solution and/or where to find the updated files,
please follow the link below. You may reopen this bug report
if the solution does not work for you.

http://rhn.redhat.com/errata/RHBA-2009-0036.html