[torqueusers] Post job file processing error / permission denied

glen.beane at gmail.com glen.beane at gmail.com
Wed Aug 14 20:57:53 MDT 2013


You're right, it should be a viable option. If passwordless ssh works from the compute nodes to the submit node the the scp should succeed.   I think in most cases Torque users have NFS home directories, which is why I suggested $usecp. 


On Aug 14, 2013, at 9:08 PM, "Mark Christiansen" <mchristi at uw.edu> wrote:

> Glen,
>  
> Thank you for your suggestion.  By moving my files to an NFS share and configuring the $usecp parameter, I was able to get things to work without an error.  However, I still don’t understand why using ssh keys didn’t work.  Perhaps it is because of the use of the Scheduler@ user that I remember seeing in one of the logs.  It seems like using password-less ssh keys should be a viable solution to solving this problem.
>  
> Best regards,
>  
> Mark
>  
> From: torqueusers-bounces at supercluster.org [mailto:torqueusers-bounces at supercluster.org] On Behalf Of glen.beane at gmail.com
> Sent: Wednesday, August 14, 2013 5:05 PM
> To: Torque Users Mailing List
> Subject: Re: [torqueusers] Post job file processing error / permission denied
>  
> You don't need to use ssh keys if this is an NFS filesystem. Check the $usecp pbs_mom config file parameter
> 
> Sent from my iPhone
> 
> On Aug 14, 2013, at 2:52 PM, "Mark Christiansen" <mchristi at uw.edu> wrote:
> 
> Hi Everyone,
>  
> I am trying to set up a torque cluster and I get the following error:
>  
> An error has occurred processing your job, see below.
> Post job file processing error; job 4
>  
> Unable to copy file /var/lib/torque/spool/4.localhost.localdomain.OU to user at host.local:/home/user/job.out
> *** error from copy
> Permission denied (publickey,gssapi-keyex,gssapi-with-mic,password,hostbased).
> lost connection
> *** end error output
> Output retained on that host in: /var/lib/torque/undelivered/4.localhost.localdomain.OU
>  
> I have looked through the archives, and there are some threads around using ssh keys to fix this problem.  However, I did configure password-less ssh logins and it works well.  If I log into the machine giving me the above error, I can perform a copy using “scp” without any problem.  I can also ssh from any machine to any machine in the cluster without getting prompted for a password.  
>  
> My cluster is configured to use LDAP for authenticating users. 
>  
> Has anyone run into this problem after configuring their ssh keys? 
>  
> Thank you in advance,
>  
> Mark
>  
> _______________________________________________
> torqueusers mailing list
> torqueusers at supercluster.org
> http://www.supercluster.org/mailman/listinfo/torqueusers
> _______________________________________________
> torqueusers mailing list
> torqueusers at supercluster.org
> http://www.supercluster.org/mailman/listinfo/torqueusers
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.supercluster.org/pipermail/torqueusers/attachments/20130814/3d40843c/attachment.html 


More information about the torqueusers mailing list