topic BL460c G6 "update_cpu_power" in BladeSystem - General

BL460c G6 "update_cpu_power"

Romualdo — Tue, 06 Dec 2011 21:14:29 GMT

Does anybody know what this means:

tscigen03:/tmp # /ulm/bin/pssh -f /tmp/cpu_power_list -t 30 "grep update_cpu_power /var/log/messages |tail -

1 "

tscca01003: Dec 5 21:32:28 tscca01003 kernel: [155256.148861] update_cpu_power: cpu_power = 155402658341731

; scale_rt = 18428885077858411363

tscca01009: Dec 5 21:32:26 tscca01009 kernel: [155222.708028] update_cpu_power: cpu_power = 155369217508907

; scale_rt = 18428885044417578539

tscca01020: Dec 5 21:32:25 tscca01020 kernel: [155347.969448] update_cpu_power: cpu_power = 155494978928854

; scale_rt = 18428885170178998486

tscca02031: Dec 5 00:20:16 tscca02031 kernel: [7484447.113265] update_cpu_power: cpu_power = 18014398378591

008; scale_rt = 18446744073578660640

tscca04005: Dec 4 08:59:07 tscca04005 kernel: [2309230.201270] update_cpu_power: cpu_power = 18014397739493

047; scale_rt = 18446744072939562679

tscca51006: Dec 5 17:53:46 tscca51006 kernel: [148699.344034] update_cpu_power: cpu_power = 148836353494934

; scale_rt = 18428878511553564566

tscca51016: Dec 5 18:51:08 tscca51016 kernel: [152091.991889] update_cpu_power: cpu_power = 152221501310966

; scale_rt = 18428881896701380598

tscca51007: Dec 5 21:32:31 tscca51007 kernel: [161527.329230] update_cpu_power: cpu_power = 161656338692678

; scale_rt = 18428891331538762310

tscca51017: Dec 5 17:10:15 tscca51017 kernel: [146291.375071] update_cpu_power: cpu_power = 146421884503415

; scale_rt = 18428876097084573047

tscca51020: Dec 5 17:07:25 tscca51020 kernel: [145774.801587] update_cpu_power: cpu_power = 145904290759344

; scale_rt = 18428875579490828976

tscca51027: Dec 5 15:15:25 tscca51027 kernel: [138937.614517] update_cpu_power: cpu_power = 139066623969726

; scale_rt = 1842886Dec 5 15:45:20 tscca51027 kernel: [140762.854972] update_group_power: cpu_power = 30024

63983

tscca51035: Dec 5 20:11:13 tscca51035 kernel: [156729.766900] update_cpu_power: cpu_power = 156860776360816

; scale_rt = 18428886535976430448

tscca64058: Dec 5 21:23:32 tscca64058 kernel: [150584.381722] update_cpu_power: cpu_power = 15071388279993p

_power: cpu_power = 4292094025

tscca51036: Dec 5 13:52:37 tscca51036 kernel: [133830.382275] update_cpu_power: cpu_power = 133962891678329

; scale_rt = 18428863638091747961

tscca51028: Dec 5 21:32:32 tscca51028 kernel: [161469.474712] update_cpu_power: cpu_power = 161600484193567

; scale_rt = 18428891275684263199

tscca01006: timed out

tscca51021: timed out

Re: BL460c G6 "update_cpu_power"

Matti_Kurkela — Thu, 08 Dec 2011 11:27:19 GMT

Looks like a Parallel SSH command: it runs the command specified within quotes (i.e. "grep update_cpu_power /var/log/messages | tail -1") in all hosts listed in file /tmp/cpu_power_list with a timeout of 30 seconds.

The quoted command finds the latest message that includes the words "update_cpu_power" in the system logfile /var/log/messages.

The identifiers "update_cpu_power", "cpu_power" and "scale_rt" can be found in the Completely Fair Scheduler (CFS) routines of the Linux kernel (in file <kernel_source_root>/kernel/sched_fair.c). Are you running some sort of debugging kernel that prints these values out?

If you're asking "what is the overall meaning of this action?" then my first guess would be that someone is gathering some CFS debugging information in some group of systems or perhaps in a Linux computing cluster.

Re: BL460c G6 "update_cpu_power"

itadmin_tk — Thu, 08 Mar 2012 11:48:56 GMT

Hi,

This sounds like a problem with a Linux Kernel bug in VMware environments

http://communities.vmware.com/message/1886118

That means, after a while of uptime some buffers will be overflow and causing a Linux VM becomes unsresponsive to the vmware host.

Affects linux kernel 2.6.32

Workaround till problem is fixed: Reboot machine before Uptime of 200 days