Re: [Condor-users] ERROR "unable to spawn the ProcD"

[Matthew Farrellee <matt@xxxxxxxxxx>]

Mark Tigges wrote:
> Hey there,
> 
> I'm trying to set condor 7.0.5 up remotely at a remote studio.  I've
> almost got it working.  Unfortunately, when I submit a job, the
> machine that it's scheduled to fails.
> 
> Here's the relevant bit of the start log on that machine:
> 
> 9/12 01:13:06 slot1: Total execute space: 23843728
> 9/12 01:13:06 slot1: Remote job ID is 8.0
> 9/12 01:13:06 slot1: Remote global job ID is
> condor@xxxxxxxxxxxxxxxxxxxxxxxxxx#1252684660#8.0
> 9/12 01:13:06 slot1: JobLeaseDuration defined in job ClassAd: 1200
> 9/12 01:13:06 slot1: Resetting ClaimLease timer (17) with new duration
> 9/12 01:13:06 slot1: About to Create_Process "condor_starter -f -a
> slot1 apshadesk431.aaa.ad.ea.com"
> 9/12 01:13:06 GetBinaryType() returned 0
> 9/12 01:13:06 return value from my_system for softkill: 2
> 9/12 01:13:06 start_procd: error received from procd: error:
> getProcInfo failed on own PID
> 9/12 01:13:06 ERROR "unable to spawn the ProcD" at line 110 in file
> ..\src\condor_c++_util\proc_family_proxy.C
> 9/12 01:13:06 CronMgr: 0 jobs alive
> 9/12 01:13:06 slot1: Canceled ClaimLease timer (17)
> 9/12 01:13:06 slot1: Changing state and activity: Claimed/Idle ->
> Preempting/Killing
> 9/12 01:13:06 Entered vacate_client <10.88.70.26:1027>
> apshadesk431.aaa.ad.ea.com...
> 9/12 01:13:06 slot1: State change: No preempting claim, returning to owner
> 9/12 01:13:06 slot1: Changing state and activity: Preempting/Killing
> -> Owner/Idle
> 9/12 01:13:06 slot1: State change: IS_OWNER is false
> 9/12 01:13:06 slot1: Changing state: Owner -> Unclaimed
> 9/12 01:13:06 startd exiting because of fatal exception.
> 
> So, when I do some googling, I find this link:
> http://www.cs.wisc.edu/condor/manual/v7.2/7_4Condor_on.html#SECTION008415000000000000000
> 
> Hopeful, I went in to the registry, unfortunatly there are no "Disable
> Performance Counter" name value pairs anywhere in the entire registry.
> 
> Is there any other cause for this kind of problem?  The executable in
> the submission is just a simple batch file.  There doesn't seem to be
> any permission problem.  I'm really stymied.  Any insight from the
> experts would be helpful.
> 
> Mark.

FYI, the STARTING UP section of any log is always helpful.

If you get really stuck try USE_PROCD=FALSE

http://www.cs.wisc.edu/condor/manual/v7.2/3_3Configuration.html#19493

Best,


matt