[torqueusers] Host key verification failed.

[Peter Praxmarer]

I doubt you can ssh from the client node to 'host3.cern.ch' (FQDN is 
important). This might be due to the fact that your 'ssh_known_hosts' 
file does not contain an entry for the FQDN.

You probably also want to examine the use of the '$usecp' flag in 
pbs_mom's config file in case your homedirectories are mounted with NFS 
throughout the whole cluster.

http://www.clusterresources.com/products/torque/docs/a.cmomconfig.shtml

HTH,
Peter

Haavard Bjerke wrote:

> Jobs aren't exiting properly because the client node isn't able to copy 
> the result back to the server node with scp.
> 
> I have two nodes. The server node is host3 and the client node is 
> host13-3. Manually, I can ssh without password authentication these ways:
> 
> root at host3 -> root at host13-3
> root at host3 -> user at host13-3
> user at host3 -> user at host13-3
> root at host13-3 -> user at host3
> root at host13-3 -> root at host3
> user at host13-3 -> user at host3
> 
> Still, I get the following error:
> 
> Unable to copy file /usr/spool/PBS/spool/110.host.OU to
> +user at host3.cern.ch:/home/user/STDIN.o110
>  >>> error from copy
> Host key verification failed.
> lost connection
>  >>> end error output
> 
> This is debug output from pbs_mom on host13-3:
> 
> pbs_mom: sys_copy, command '/usr/bin/scp -Brp 
> /usr/spool/PBS/spool/111.host.OU 
> user at host3.cern.ch:/home/user/STDIN.o111' failed with status=1, giving 
> up after 4 attempts
> 
> Any help would be appreciated
> 
> -- 
> Havard
> CERN Openlab
> _______________________________________________
> torqueusers mailing list
> torqueusers at supercluster.org
> http://www.supercluster.org/mailman/listinfo/torqueusers