[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Condor-users] jobs submitted from windows don't execute

On 5/13/07, mohammed shambakey <shambakey1@xxxxxxxxx> wrote:
i checked the shadow log and it has the followng error:-

5/10 10:00:37 Using config source: C:\condor\condor_config
5/10 10:00:37 Using local config sources:
5/10 10:00:37    C:\condor/condor_config.local
5/10 10:00:37 DaemonCore: Command Socket at
5/10 10:00:38 Initializing a JAVA shadow for job 7.0
5/10 10:00:38 (7.0) (3316): Request to run on <> was
5/10 10:00:38 (7.0) (3316): ERROR "select, error # = 10038" at line 2417 in
file ..\src\condor_daemon_core.V6\daemon_core.C

and the schedd.log gives :-

5/13 10:41:11 (pid:3836) Using config source: C:\condor\condor_config
5/13 10:41:11 (pid:3836) Using local config sources:
5/13 10:41:11 (pid:3836)    C:\condor/condor_config.local
5/13 10:41:12 (pid:3836) DaemonCore: Command Socket at
5/13 10:41:12 (pid:3836) History file rotation is enabled.
5/13 10:41:12 (pid:3836)   Maximum history file size is: 20971520 bytes
5/13 10:41:12 (pid:3836)   Number of rotated history files is: 2
5/13 10:41:12 (pid:3836) my_popen: CreateProcess failed
5/13 10:41:12 (pid:3836) Failed to execute
C:\condor/bin/condor_shadow.pvm.exe, ignoring
5/13 10:41:12 (pid:3836) my_popen: CreateProcess failed
5/13 10:41:12 (pid:3836) Failed to execute
C:\condor/bin/condor_shadow.std.exe, ignoring
5/13 10:41:12 (pid:3836) Sent ad to central manager for
5/13 10:41:12 (pid:3836) Sent ad to 1 collectors for shambakey@shambakeyserv
5/13 10:41:12 (pid:3836) ERROR "select, error # = 10038" at line 2417 in
file ..\src\condor_daemon_core.V6\daemon_core.C

i looked for this daemon_core.c but didn't find it.
one thing, i have to start schedd.exe manually as it's not included with
services running in windows, i don't know if this has something to do with
the problem

I'm afraid you have a serious misunderstanding of how condor works
regarding the various daemons it needs.
You should not start the daemons directly (unless you are an expert
user doing something complicated at least)

Instead you should alter the condor_config file in you condor
installation and change the following line:

DAEMON_LIST = MASTER,   ... etc ...

to include SCHEDD in the comma separated list. On windows only one
service is required for condor which fires up the master (which then
creates the relevant schedd/startd/ whatever)

I suggest taking another read through the documentation in light of
this and see if this makes more sense and works.

Hope this helps