Re: [Condor-users] DAGman duplicating jobs on schedd restart

["R. Kent Wenger" <wenger@xxxxxxxxxxx>]

On Thu, 3 Nov 2011, Christopher Martin wrote:

> This is great Kent, thank you. We will probably be moving our logging off
> of the Windows share instead logging on the execute node and using the
> should_transfer_files option to transfer the logs back to the
> submit/scheduler machine. We've been having fileLock issues which is most
> likely related to this. Hopefully this will sort itself out with the
> logging changes.

Well, the logging is done by the schedd you submit to, so it's not 
affected by the file transfer options.  But if you can move the log files 
off of a shared file system and onto the local disk on the submit machine, 
that should help.

Kent Wenger
Condor Team