<pre>>Yes, people have their own cronjobs handle this stuff.<br><br>>My cluster has a daily cronjob that ssh's to each node, moving all log files<br>>back to a head node.<br></pre><br>Thanks! I wrote my own cron scripts now. The logs are handled correctly.<br>
<br>The /var/spool/torque/spool is still a sticking point though. There are some huge files here eg. "<a href="http://233139.supe.ER">233139.supe.ER</a>". <br><br>The corresponding job has ended qstat reveals. Isn't pbs (torque) supposed to remove these. What might I be doing wrong? <br>
<br>The only other option for me seems again a cron every hour that looks to see if a jobid has ended and then manually deletes files. But I think that shouldn't be needed. Or should it?<br><br>-- <br>Rahul<br>