[Condor-users] Having probs upgrading 7.4.3 to 7.5.4 Windows XP submit node

[<Greg.Hitchen@xxxxxxxx>]

Hi All

 

We currently have a number of pools of windows execute nodes (7.2.4), windows submit nodes (7.4.3), and

linux central managers (7.2.3, 7.4.3). Just thought I'd try one windows submit node with 7.5.4 given the

nice sounding improvements in scalability for running jobs. Unfortunately the schedd keeps crashing.

See the excerpts from the log files below. Interestingly we have NEVER had any procd specific stuff in the

condor_config files previously but I cut/pasted/modified stuff from the example condor_config.generic as

I initially had an error like:

 

SchedLog

 

11/24/10 15:25:18 Create_Process(): Failed to extract the extension from file C:\PROGRA~1\condor/bin/condor_procd.
11/24/10 15:25:18 SharedPortEndpoint: Destructor: Problem in thread shutdown notification: 2

 

until I explicitly set PROCD = $(SBIN)/condor_procd.exe That stopped the above error but I am still stuck at the moment

with the errors listed below.

Thanks for any help.

 

Cheers

 

Greg

 

 

SchedLog

 

11/24/10 16:30:14 DaemonCore: pid 15660 exited with status 1, invoking reaper 1 <condor_procd reaper>
11/24/10 16:30:14 procd (pid = 15660) exited unexpectedly with status 1
11/24/10 16:30:14 ERROR "ProcD has failed" at line 608 in file ..\src\condor_c++_util\proc_family_proxy.cpp

ProcLog.SCHEDD

 

11/24/10 16:30:14 : ...snapshot complete
11/24/10 16:30:14 : CreateNamedPipe(C:\PROGRA~1\condor/log/procd_pipe.SCHEDD) error: 123
11/24/10 16:30:14 : ERROR: ProcFamilyServer: could not initialize LocalServer

condor_config

 

PROCD = $(SBIN)/condor_procd.exe
PROCD_ADDRESS = $(LOCK)/procd_pipe
PROCD_LOG = $(LOG)/ProcLog
PROCD_MAX_SNAPSHOT_INTERVAL = 60
WINDOWS_SOFTKILL = $(SBIN)/condor_softkill.exe