Re: [Condor-users] condor job has exited normallywithstatus-1073741515

Mailing List Archives Authenticated access	UW Madison Computer Sciences Department Computer Systems Lab

In general the environment your job runs with via Condor is considerably reduced. Condor sets up the scantest environment possible for your job before executing it. Try running a job that does:

myjob.bat:

set

To dump the remote environment for the job. You might need to export pieces of your submitting environment to get a specific command to work.

On Windows, Condor also runs your job in a fairly restricted desktop. The account under which it runs the job, assuming you haven’t set up Condor’s credential daemon, is a temporary, local user account on the machine with limited rights. The systeminfo.exe command may require better user privilege than what the account under which the remote job is executing has.

Hope that helps debugging the problem.

- Ian

From: condor-users-bounces@xxxxxxxxxxx [mailto:condor-users-bounces@xxxxxxxxxxx] On Behalf Of Alas, Alex [FEDI]
Sent: Monday, November 17, 2008 5:05 PM
To: Condor-Users Mail List
Subject: Re: [Condor-users] condor job has exited normallywithstatus-1073741515

Dave,

I tried that and when you run the batch file or the systeminfo executable, it runs without any problems is only when you try to run it through condor that you see the error message. I will download the depens.exe tool to track the problem,

So the logs I sent you were not useful at all?

Thanks for your help.

Alex

From: condor-users-bounces@xxxxxxxxxxx [mailto:condor-users-bounces@xxxxxxxxxxx] On Behalf Of David Watrous
Sent: Monday, November 17, 2008 4:36 PM
To: Condor-Users Mail List
Subject: Re: [Condor-users] condor job has exited normally withstatus-1073741515

Hi Alex,

Sorry for any confusion. By the "execute node" I meant the machine that runs the condor_startd daemon, which actually runs the job. By "scheduler node" I meant the machine running the condor_schedd daemon. Most of the time this is where you run the "condor_submit" command.

I did a little research about the exit code, and a value of -1073741515 is actually Windows error C0000135. At a system level this means "Unable to Locate DLL" or "Unable To Locate Component." It's possible that the systeminfo.exe program has an external dependency on a DLL or a COM component that is not installed on the execute machine.

If possible, I would try going to that machine and running systeminfo.exe by hand. This may pop up an error message that might give you more information. If that is not possible, you can use a tool called Depends.exe:

http://www.dependencywalker.com/

This program examines executable files and lists any DLL dependencies it may have. Most of these DLLs should already be installed on the remote machine, but some of these DLL files may need to be transferred to the remote machine using Condor's "transfer_input_files" mechanism.

Good luck!

- Dave

On Nov 17, 2008, at 3:32 PM, Alas, Alex [FEDI] wrote:

Dave,

Thanks for your input. I am kind of a newbie here so I interpreted execute node as the computer where I ran the condor_submit job and the Scheduler node where it intends to run, am I right or wrong? Please correct me If I am.

This is what the Starterlog on the execute node says

11/17 13:30:47 ******************************************************

11/17 13:30:47 ** condor_starter (CONDOR_STARTER) STARTING UP

11/17 13:30:47 ** C:\condor\bin\condor_starter.exe

11/17 13:30:47 ** $CondorVersion: 6.8.7 Nov 29 2007 $

11/17 13:30:47 ** $CondorPlatform: INTEL-WINNT50 $

11/17 13:30:47 ** PID = 2504

11/17 13:30:47 ** Log last touched 11/17 13:22:58

11/17 13:30:47 ******************************************************

11/17 13:30:47 Using config source: C:\condor\condor_config

11/17 13:30:47 Using local config sources:

11/17 13:30:47 C:\condor/condor_config.local

11/17 13:30:47 DaemonCore: Command Socket at <1x.xx.xx.x9:3354>

11/17 13:30:47 Setting resource limits not implemented!

11/17 13:30:47 Communicating with shadow <1x.xx.xx.x4:2784>

11/17 13:30:47 Submitting machine is "theisman.domain.com"

11/17 13:30:47 File transfer completed successfully.

11/17 13:30:48 Starting a VANILLA universe job with ID: 23.0

11/17 13:30:48 IWD: C:\condor/execute\dir_2504

11/17 13:30:48 Output file: C:\condor/execute\dir_2504\Batch4testv3.out.0

11/17 13:30:48 Error file: C:\condor/execute\dir_2504\Batch4testv3.err.0

11/17 13:30:48 Renice expr "10" evaluated to 10

11/17 13:30:48 About to exec C:\WINDOWS\system32\cmd.exe /Q /C condor_exec.bat

11/17 13:30:48 Create_Process succeeded, pid=4080

11/17 13:30:48 Process exited, pid=4080, status=-1073741515

11/17 13:30:48 Got SIGQUIT. Performing fast shutdown.

11/17 13:30:48 ShutdownFast all jobs.

11/17 13:30:48 **** condor_starter (condor_STARTER) EXITING WITH STATUS 0

11/17 14:34:32 *************************************************

Scheduler Node’s shadowlog is the following:

11/17 13:30:38 Using config source: C:\Condor\condor_config

11/17 13:30:38 Using local config sources:

11/17 13:30:38 C:\Condor/condor_config.local

11/17 13:30:38 DaemonCore: Command Socket at <1x.xx.xx.x4:2784>

11/17 13:30:38 Initializing a VANILLA shadow for job 23.0

11/17 13:30:38 (23.0) (1012): Request to run on <1x.xx.xx.9:1104> was ACCEPTED

11/17 13:30:40 (23.0) (1012): ZKM: setting default map to (null)

11/17 13:30:40 (23.0) (1012): Job 23.0 terminated: exited with status -1073741515

11/17 13:30:40 (23.0) (1012): **** condor_shadow (condor_SHADOW) EXITING WITH STATUS 100

Thanks for your input and I hope this logs can help,

Alex

From: condor-users-bounces@xxxxxxxxxxx [mailto:condor-users-bounces@xxxxxxxxxxx] On Behalf Of David Watrous
Sent: Monday, November 17, 2008 2:55 PM
To: Condor-Users Mail List
Subject: Re: [Condor-users] condor job has exited normally with status-1073741515

Alex,

What does your StarterLog on the execute node and the ShadowLog on the scheduler say about these jobs as they run? Do you get an "ERROR: Provider load failure" message in the output/error file from systeminfo?

If you don't see anything interesting in those logs, I'm assuming that your scheduler and execute nodes all have the referenced directories, so what happens when your batch file just does an 'echo "Hello World"'?

I hope this helps!

Good luck,

- Dave

--
===================================
David Watrous
main: 888.292.5320

Cycle Computing, LLC

Leader in Condor Grid Solutions

Enterprise Condor Support and Management Tools

http://www.cyclecomputing.com

http://www.cyclecloud.com

On Nov 17, 2008, at 2:34 PM, Alas, Alex [FEDI] wrote:

I am trying to run a job but it fails exiting with the code 107374515. I know if the job succeeds it will exit out with code 0 and anything else non-zero code means it failed but I don’t know if this code means anything or if it is a generic error code.

My description file is the following:

#########################################################################################

# Description file for Batch File for TESTING purposes

# Prepared by Alex Alas

##########################################################################################

universe = vanilla

requirements = (Arch == "INTEL" && OpSys == "WINNT51")

notify_user=username@xxxxxxxxxx

initialdir = c:\condor\execute_bk

should_transfer_files = YES

when_to_transfer_output = ON_EXIT

transfer_input_files = c:\windows\system32\systeminfo.exe

run_as_owner = true

executable = Batch4testv2.bat

output = Batch4testv3.out.$(Process)

error = Batch4testv3.err.$(Process)

log = Batch4testv3.log

queue 1

The batch file I am running is as follow:

> systeminfo.exe

Any input is much appreciated,

Respectfully,

Alex Alas

Systems Administrator
Fugro EarthData Inc.

_______________________________________________
Condor-users mailing list
To unsubscribe, send a message to condor-users-request@xxxxxxxxxxx with a
subject: Unsubscribe
You can also unsubscribe by visiting
https://lists.cs.wisc.edu/mailman/listinfo/condor-users

The archives can be found at:
https://lists.cs.wisc.edu/archive/condor-users/

--
===================================

David Watrous

main: 888.292.5320

Cycle Computing, LLC

Leader in Condor Grid Solutions

Enterprise Condor Support and Management Tools

http://www.cyclecomputing.com
http://www.cyclecloud.com

Mailing List Archives

Authenticated access

Re: [Condor-users] condor job has exited normallywithstatus-1073741515