From: Flemming Videbaek (videbaek@sgs1.hirg.bnl.gov)
Date: Wed Apr 16 2003 - 12:22:47 EDT
Some comments as ws well actions taken. A long time ago we agreed to participating in a general sharing of some LSF resources at RCF. This is the origin of the gen_hepnp_rhic LSF queue. It was meant not to interfere with priority usage of the cas machines. As noticed from the e-mail yesterday this was not the case. I have communicated the issues with o.rind and he changed a couple of things for the queues such that. citation: "I raised the nice level on your queues and lowered the gen queue nice level to 20 2) I raised the r1m scheduling load limit to 1.5. This last change should enable Brahms jobs to preempt any gen jobs on your nodes. I'll keep an eye on things for a little while longer, but these changes already seem to be doing the trick and I see the jobs in your queues moving along. " It is also our privelige to with-draw from this arrangement, but in general I find this not so justefiable, since the average use of the CAS machines from our side is probably 20% or less. If you doubt this go to the RCF metric page and look at historical information. The agreement also implies that we can use the general queues. The machine that are shared from our end should only be rcas0001-0009 (5 excluded). /fv ------------------------------------------------------ Flemming Videbaek Physics Department Brookhaven National Laboratory tlf: 631-344-4106 fax 631-344-1334 e-mail: videbaek@bnl.gov
This archive was generated by hypermail 2.1.5 : Wed Apr 16 2003 - 12:19:52 EDT