Re: [Condor-users] Why do remote job submissions on Linux require 'should_transfer_files'

[Tim St Clair <tstclair@xxxxxxxxxx>]

inline

----- Original Message -----
> From: "Thomas Luff" <Thomas.Luff@xxxxxxx>
> To: condor-users@xxxxxxxxxxx
> Sent: Friday, March 9, 2012 8:27:45 AM
> Subject: [Condor-users] Why do remote job submissions on Linux require 'should_transfer_files'
> 
> I'm currently toying with remote job submission for our validation
> cluster and have come across a bit of an issue.
> 
> Jobs that are remotely submit to be run on Linux without specifying
> 'should_transfer_files = true' get stuck in the hold state with the
> following error:
> 
>     023.000:  Request is held.
>     Hold reason: Error from slot1@xxxxxxxxxx: Failed to execute
>     '/bin/sleep' with arguments 1m: Cannot access specified iwd
>     "/work/condor/local/spool/23/0/cluster23.proc0.subproc0"
>     (errno=2: 'No such file or directory')
> 
> The Linux job only gets stuck when remotely submitted (condor_submit
> -remote). It executes successfully when submitted locally.
> 
> A similar job can be run fine on a windows node with
> 'should_transfer_files = IF_NEEDED'.
> 
> My Jobs are as follows:
>     Linux:
> <pre>
>         Executable = /bin/sleep
>         Arguments = "1m"
> 
>         Requirements = (ARCH=="INTEL" || ARCH=="X86_64") &&
>         OpSys=="LINUX"
> 
>         transfer_executable = false
>         should_transfer_files = TRUE
>         WhenToTransferOutput = ON_EXIT
>         Run_As_Owner = True
>         Queue
> </pre>
> 
>     Windows:
> <pre>
>          Executable = \\network\path\to\bin\perl
>          Arguments = "-e '$time = time+60; sleep 1 until time >
>          $time'" #Sleep 60 seconds
> 
>          Requirements = ((ARCH=="INTEL" || ARCH=="X86_64") &&
>          OpSys=="WINNT51")
> 
>          transfer_executable = false
>          should_transfer_files = IF_NEEDED
>          WhenToTransferOutput = ON_EXIT
> 
>          Run_As_Owner = True
> 
>          Queue
> </pre>
> 
> My question is, Why is this needed and why is it different between
> Windows and Linux?

condor_submit fills in a series of defaults when working locally, a simple way to check the full list to compare the differences is to run `condor_q -long cluster.proc` and compare the differences.

In general: The path of least resistance for remote submission, in my experience, has been to set:

should_transfer_files = TRUE
# something sane on windows, but not needed for windows.
iwd = /tmp  

> 
> Thanks
> 
> P.S Has anyone successfully built Condor for ARM Linux?

Not yet.  Once ARM makes it's way to the data center it might become more of a priority, but right now it would just melt a cell phone.

> 
> 
> 
> 
> -- IMPORTANT NOTICE: The contents of this email and any attachments
> are confidential and may also be privileged. If you are not the
> intended recipient, please notify the sender immediately and do not
> disclose the contents to any other person, use it for any purpose,
> or store or copy the information in any medium.  Thank you.
> 
> _______________________________________________
> 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/
>