Make New Entry,
Make Followup Entry
User name Ole Hansen
Log entry time 23:58:17 on October15,2009
Entry number 295144
keyword=Analysis crashes due to full disk
We noticed that the panguin analysis crashed every 10 minutes or so. We found /adaql1/work1 to be 100% full due to analysis results being written to /adaql1/work1/panguin_rootfiles. We cleaned up this directory, assuming the files can easily be reproduced. Now, that disk is really small (21 GB) - not the best place to put large analysis output. Someone should reconfigure the analysis to write to a bigger disk.
On this occasion, we noticed that /adaql3/work1 is also nearly 100% full (275GB worth of stuff), and /adaql10/work1 is approaching capacity (2.5 TB worth of stuff!!). Someone should CLEAN UP.