From: Greenseid, J. M. <Jos...@ng...> - 2008-04-29 13:43:30
|
Did you mark the node offline in the middle of the job? If the node is marked as offline, no jobs should be sent to it, so the scheduler/queueing system will just consider it unavailable and use only the other available resources. --Joe ________________________________ From: osc...@li... on behalf of Zhimin Xiong Sent: Tue 4/29/2008 8:16 AM To: osc...@li... Subject: [Oscar-users] Help on qsub problem Dear OSCAR users, I had a question on qsub. I found the job exited with -2 from the file /var/spool/server_priv/accounting/20080429 when I submit a job via a script, and no other output file. It may be ascribed that I marked one node offline due to some problem, that it reported " EDAC MC0: CE - no information available: e7xxx CE log register overflow" on that node when booting over. So the questions are: 1) what is the reason for the above error on that node? 2) if I don't use that node, for example, I marked it offline, then how to make the system work well as usual? Thanks! -- Zhimin Xiong State Key Lab. of Polymer Physics & Chemistry Institute of Chemistry, Chinese Academy of Sciences(ICCAS) |