Re: [Condor-users] Shadow Exception !!!

[John Wheez <john@xxxxxxxxxx>]

I have had this error before. You need to make sure that the user which 
executes "condor_submit" or "condor_submit_dag" has read/write 
proiveledges in it's current working directory.
If you are on linux this is probably the "condor" user account...some 
people try to run condor as root to overcome file permission problems 
but, condor will not submit a job as user root..so do not try this.
In my scripts i always change to the directory where my condor files are 
located to ensure that i am in a directory  which will allow read and 
write priveledges for the user.

In my case i was using PhP scripts to automatically submit and generate 
jobs.
So i had to tell PhP to switch to teh directory where my condor files 
were located and use that for it's current working directory..


Hope this helps out.

JW


Condor Grib wrote:

> I've installed the last version of condor in my PC,
> and it's running ok under linux (Redhat 9). The
> problem appears when I sent a process to run, after
> running for a few seconds a shadow exception appears,
> and the the process starts again until another shadow
> exception stops it. I dunno what's happening, if you
> run the process in one pc wothout using condors works
> perfect. 
> Looking at the logs first I saw in the log of my
> program this:
> 000 (003.000.000) 09/21 13:40:49 Job submitted from
> host: <193.147.240.233:36284
>  
>
> ...
> 001 (003.000.000) 09/21 13:40:52 Job executing on
> host: <193.147.240.233:36286>
> ...
> 006 (003.000.000) 09/21 13:41:00 Image size of job
> updated: 1332
> ...
> 007 (003.000.000) 09/21 13:41:42 Shadow exception!
>        Can no longer talk to condor_starter
> <193.147.240.233:36286>
>        0  -  Run Bytes Sent By Job
>        14829  -  Run Bytes Received By Job
> *****************************************************
>
> So i looked in the Starterlog and this is what I've
> got:
>
> Starterlog:
>
> 9/21 13:40:52
> ******************************************************
> 9/21 13:40:52 ** condor_starter (CONDOR_STARTER)
> STARTING UP
> 9/21 13:40:52 **
> /home/condor/condor-6.7.1/sbin/condor_starter
> 9/21 13:40:52 ** $CondorVersion: 6.7.1 Aug 10 2004 $
> 9/21 13:40:52 ** $CondorPlatform: I386-LINUX_RH9 $
> 9/21 13:40:52 ** PID = 15935
> 9/21 13:40:52
> ******************************************************
> 9/21 13:40:52 Using config file:
> /home/condor/condor-6.7.1/etc/condor_config
> 9/21 13:40:52 Using local config files:
> /home/condor/condor-6.7.1/local.golem/c
> o
> ndor_config.local
> 9/21 13:40:52 DaemonCore: Command Socket at
> <193.147.240.233:36308>
> 9/21 13:40:52 Done setting resource limits
> 9/21 13:40:52 Communicating with shadow
> <193.147.240.233:36306>
> 9/21 13:40:52 Submitting machine is "golem.imim.es"
> 9/21 13:40:52 File transfer completed successfully.
> 9/21 13:40:52 Starting a VANILLA universe job with ID:
> 3.0
> 9/21 13:40:52 IWD:
> /home/condor/condor-6.7.1/local.golem/execute/dir_15935
> 9/21 13:40:52 Output file:
> /home/condor/condor-6.7.1/local.golem/execute/dir_15
> 935/2program.out
> 9/21 13:40:52 Error file:
> /home/condor/condor-6.7.1/local.golem/execute/dir_159
> 35/2program.err
> 9/21 13:40:52 About to exec
> /home/condor/condor-6.7.1/local.golem/execute/dir_1
> 5935/condor_exec.exe
> 9/21 13:40:52 Create_Process succeeded, pid=15937
> 9/21 13:41:42 Process exited, pid=15937, status=0
> 9/21 13:41:42 ReliSock: put_file: Failed to open file
> /home/condor/condor-6.7.1
> /local.golem/execute/dir_15935/2program.log, errno =
> 2.
> 9/21 13:41:42 ERROR "DoUpload: Failed to send file
> /home/condor/condor-6.7.1/lo
> cal.golem/execute/dir_15935/2program.log, exiting at
> 1408
> " at line 1407 in file file_transfer.C
> 9/21 13:41:42 ShutdownFast all jobs.
> *****************************************************
>
> the Shadowlog :
> 9/21 13:40:52
> ******************************************************
> 9/21 13:40:52 ** condor_shadow (CONDOR_SHADOW)
> STARTING UP
> 9/21 13:40:52 **
> /home/condor/condor-6.7.1/sbin/condor_shadow
> 9/21 13:40:52 ** $CondorVersion: 6.7.1 Aug 10 2004 $
> 9/21 13:40:52 ** $CondorPlatform: I386-LINUX_RH9 $
> 9/21 13:40:52 ** PID = 15934
> 9/21 13:40:52
> ******************************************************
> 9/21 13:40:52 Using config file:
> /home/condor/condor-6.7.1/etc/condor_config
> 9/21 13:40:52 Using local config files:
> /home/condor/condor-6.7.1/local.golem/c
> ondor_config.local
> 9/21 13:40:52 DaemonCore: Command Socket at
> <193.147.240.233:36306>
> 9/21 13:40:52 Initializing a VANILLA shadow for job
> 3.0
> 9/21 13:40:52 (3.0) (15934): Request to run on
> <193.147.240.233:36286> was ACCE
> PTED
> 9/21 13:41:42 (3.0) (15934): ERROR "Can no longer talk
> to condor_starter <193.1
> 47.240.233:36286>" at line 93 in file NTreceivers.C
> *********************
>
> Anyone knows where the problem is? 
>
> BTW, I just have only one machine that everytime a
> process is send it starts running imediately.
> If you need more info let me know
>
>
>
>
> 		
> ______________________________________________
> Renovamos el Correo Yahoo!: ¡100 MB GRATIS!
> Nuevos servicios, más seguridad
> http://correo.yahoo.es
> _______________________________________________
> Condor-users mailing list
> Condor-users@xxxxxxxxxxx
> http://lists.cs.wisc.edu/mailman/listinfo/condor-users
>