Schedd restarts after condor_procd exit in MRG Grid 2.3

Issue

• Schedd restarts after condor_procd exits
• Using a multi-schedd configuration
• condor_procd log indicates:

05/21/13 12:47:15 : open for write-only of /var/run/condor/procd_pipe.SCHEDD.watchdog failed: No such file or directory (2)
05/21/13 12:47:15 : failed to initialize watchdog named pipe at /var/run/condor/procd_pipe.SCHEDD.watchdog
05/21/13 12:47:15 : ERROR: ProcFamilyServer: could not initialize LocalServer
...
05/21/13 12:48:25 : NamedPipeReader::consistent(): The named pipe at m_addr: '/var/run/condor/procd_pipe.SCHEDD' is inconsistent with the originally opened m_addr when the procd was started.
05/21/13 12:48:25 : ERROR: ProcFamilyServer: Namedpipe reader isn't consistent

• Condor_schedd log indicates:

05/21/13 12:47:15 (pid:7217) procd (pid = 7219) exited unexpectedly with status 256
05/21/13 12:47:15 (pid:7217) ERROR "ProcD has failed" at line 621 in file /home/jrthomas/rpmbuild/BUILD/condor-7.8.6/src/condor_utils/proc_family_proxy.cpp