Hi Stephane and Thomas I give you an update of the status of our robinhood server.
Increasing the kernel.sched_migration_cost to 10000000 didn't improve the situation. I tried 50000000 too but didn't change much. The load of the machine is always between 30 and 40 with the migration processes between 20 to 50 %CPU. So we decided to restore the kernel.sched_migration_cost to 500000 and try only numactl. Our server has two sockets with 8 cores each on hyper-threading. The result of numactl --hardware is: available: 2 nodes (0-1) node 0 cpus: 0 1 2 3 4 5 6 7 16 17 18 19 20 21 22 23 node 0 size: 65381 MB node 0 free: 974 MB node 1 cpus: 8 9 10 11 12 13 14 15 24 25 26 27 28 29 30 31 node 1 size: 65536 MB node 1 free: 1164 MB node distances: node 0 1 0: 10 11 1: 11 10 So reduced nb_threads = 8 and nb_threads_purge = 8 and I restarted robinhood with only --read-log --purge as following: numactl --cpunodebind=1 --membind=1 /usr/sbin/robinhood -d -f /etc/robinhood.d/tmpfs/cscs/rbh_daily.conf -p /var/run/robinhood.pid I will wait until tomorrow to see if the situation will go better but I can immediately noticed that the cpu usage of robinhood now is between 40% and 100%. The load of mysql didn't change: 2847 root 20 0 3867m 1.7g 1528 S 63.9 1.4 47:41.73 robinhood 3217 mysql 20 0 37.6g 1.3g 4500 S 0.0 1.0 1603:03 mysqld The next test I would like to try is deactivating the hyper-threading. Do you thing could help? regards Carmelo ------------------------------------------------------------------------------ BPM Camp - Free Virtual Workshop May 6th at 10am PDT/1PM EDT Develop your own process in accordance with the BPMN 2 standard Learn Process modeling best practices with Bonita BPM through live exercises http://www.bonitasoft.com/be-part-of-it/events/bpm-camp-virtual- event?utm_ source=Sourceforge_BPM_Camp_5_6_15&utm_medium=email&utm_campaign=VA_SF _______________________________________________ robinhood-support mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/robinhood-support
