jobsub_fetchlog timeout problem
When a large number of files are in a sandbox the server occasionally times out creating a tarfile, leaving an orphan .tgz in the users sandbox directory and a frustrated user.
We need 2 things to correct this situation
- lengthen the timeout for a fetchlog
- catch a timeout, clean up the intermediate tgz file, and inform the user as to what happened. If the timeout value is in a config file they can then open a ticket to have the timeout value increased