Article ID: 962 - Last Modified: December 4, 2010
Long jobs submitted to an SGE queue sometimes fail. What can I do?
Jobs can fail because SGE writes some of its job files to /tmp, and if this directory is automatically cleaned up via an automatic crontab task that uses tmpwatch(8), jobs that are longer than the cleanup time period can have job files removed. The workaround is to configure the cleanup of /tmp to an interval that is longer than the expected duration of any job.
Type the words or phrases on which you would like to search, or click here to view a list of all
Knowledge Base articles