Hiro, 2 hours is better. Some of my jobs take 1400seconds to finish. If I turn on the mixed event part of the analysis I believe it takes about 2800seconds. Both are less than 2 hours, but a good buffer will work best in order to not stop any jobs prematurely. Erik -----Original Message----- From: brahms-dev-l-bounces+ebj=ku.edu@lists.bnl.gov on behalf of Hironori Ito Sent: Thu 9/15/2005 2:34 PM To: flemming videbaek Cc: devlist Subject: [Brahms-dev-l] Re: proofserver Hello. When root session with proof dies, sometimes it does not kill your proof service. This is quite annoying since it requires a user to logon to every slave machines and to kill all proof jobs manually. However, any console jobs that is running longtime (I don't remember the duration) will be automatically reniced after some time for jobs by any other users. I am thinking to make a program to kill just proof jobs with more than a few hours of cpu time. (I am assuming nobody needs proof for more than a hour or two. For example, I can loop through run04 AuAu one angle (with many magnet settings) in about 30 minutes.) Hiro flemming videbaek wrote: > Hi Eun-Joo > you have a proffserver running on rcas0005 that has accumulated 70555 > minutes of cpu; > I have seen something like this happen to me but was discovered by > erik after 24 hours > Could you kill the process > > Flemming > > ---------------------------------------------------------------- > Flemming Videbaek > Physics Department > Brookhaven National Laboratory > > e-mail: videbaek@bnl.gov <mailto:videbaek@bnl.gov> > phone: 631-344-4106 _______________________________________________ Brahms-dev-l mailing list Brahms-dev-l@lists.bnl.gov http://lists.bnl.gov/mailman/listinfo/brahms-dev-l _______________________________________________ Brahms-dev-l mailing list Brahms-dev-l@lists.bnl.gov http://lists.bnl.gov/mailman/listinfo/brahms-dev-lReceived on Thu Sep 15 15:50:24 2005
This archive was generated by hypermail 2.1.8 : Thu Sep 15 2005 - 15:50:34 EDT