Something strange is going on with my scheduling

Author	Message
KWSN Sir Clark Send message Joined: 24 Jul 05 Posts: 35 Credit: 19,384 RAC: 0	Message 9378 - Posted: 13 Aug 2005, 12:26:04 UTC Last modified: 13 Aug 2005, 12:28:24 UTC I'm running only two projects at the moment LHC & CPDN. CPDN is 11% through and has a deadline in 2006. LHC is 10% through and has a deadline of 21st. LHC has resource share of 200, CPDN's is 100 Why is my computer using Earliest Deadline First???? I'm set to connect every 4 days. From my messages: 13/08/2005 09:13:26\|\|Using earliest-deadline-first scheduling because computer is overcommitted. 13/08/2005 10:11:32\|LHC@home\|Result wjun2C_v6s4hvpac_mqx__15__64.295_59.305__4_6__6__50_1_sixvf_boinc19353_2 exited with zero status but no 'finished' file 13/08/2005 10:11:32\|LHC@home\|If this happens repeatedly you may need to reset the project. 13/08/2005 10:11:32\|\|request_reschedule_cpus: process exited 13/08/2005 10:11:32\|LHC@home\|Restarting result wjun2C_v6s4hvpac_mqx__15__64.295_59.305__4_6__6__50_1_sixvf_boinc19353_2 using sixtrack version 4.67 13/08/2005 12:27:07\|\|request_reschedule_cpus: process exited 13/08/2005 12:27:07\|LHC@home\|Computation for result wjun2C_v6s4hvpac_mqx__15__64.295_59.305__4_6__6__50_1_sixvf_boinc19353_2 finished 13/08/2005 12:27:07\|\|Resuming round-robin CPU scheduling. 13/08/2005 12:27:08\|LHC@home\|Sending scheduler request to http://lhcathome-sched1.cern.ch/scheduler/cgi 13/08/2005 12:27:08\|LHC@home\|Requesting 432000 seconds of work, returning 0 results 13/08/2005 12:27:08\|LHC@home\|Started upload of wjun2C_v6s4hvpac_mqx__15__64.295_59.305__4_6__6__50_1_sixvf_boinc19353_2_0 13/08/2005 12:27:09\|LHC@home\|Scheduler request to http://lhcathome-sched1.cern.ch/scheduler/cgi succeeded 13/08/2005 12:27:10\|\|Using earliest-deadline-first scheduling because computer is overcommitted. 13/08/2005 12:27:10\|LHC@home\|Started download of wjun2C_v6s4hvpac_mqx__12__64.293_59.303__6_8__6__10_1_sixvf_boinc15129.zip 13/08/2005 12:27:11\|LHC@home\|Finished upload of wjun2C_v6s4hvpac_mqx__15__64.295_59.305__4_6__6__50_1_sixvf_boinc19353_2_0 13/08/2005 12:27:11\|LHC@home\|Throughput 29342 bytes/sec 13/08/2005 12:27:11\|LHC@home\|Finished download of wjun2C_v6s4hvpac_mqx__12__64.293_59.303__6_8__6__10_1_sixvf_boinc15129.zip 13/08/2005 12:27:11\|LHC@home\|Throughput 50328 bytes/sec 13/08/2005 12:27:11\|\|request_reschedule_cpus: files downloaded 13/08/2005 12:27:11\|LHC@home\|Starting result wjun2C_v6s4hvpac_mqx__12__64.293_59.303__6_8__6__10_1_sixvf_boinc15129_3 using sixtrack version 4.67 www.chris-kent.co.uk aka Chief.com ID: 9378 · Reply Quote

KWSN Sir Clark Send message Joined: 24 Jul 05 Posts: 35 Credit: 19,384 RAC: 0	Message 9379 - Posted: 13 Aug 2005, 14:18:06 UTC Got it sorted. It was because it was still looking at the Resource share as if I only have 7% of time assigned to it. I detached from the projects I'm not running and hey presto it worked due to more time allocation being given to it. www.chris-kent.co.uk aka Chief.com ID: 9379 · Reply Quote

John McLeod VII Send message Joined: 2 Sep 04 Posts: 165 Credit: 146,925 RAC: 0	Message 9384 - Posted: 13 Aug 2005, 16:49:17 UTC EDF only looks at the work on hand. If the deadline is less than 2 * the connect every X days setting, then EDF will be entered. THis is so that modem users will get work done before the connection that is just earlier than the deadline so theat it can be freported on time. Unless you are using a modem (or otherwise have a computer that is disconnected from the Internet for some period of time) you should set your connect every X days to be fairly short and let the other project(s) pick up the slack if one project is down. BOINC WIKI ID: 9384 · Reply Quote

LHC@home