OpenVZ Forum


Home » General » Support » Strange top %us values on 55.10 kernel
Strange top %us values on 55.10 kernel [message #46729] Wed, 06 June 2012 10:37 Go to next message
lazy
Messages: 16
Registered: January 2008
Junior Member
on containers inside 2.6.32-042stab055.10 kernel most of the time %id is 0%, it dosn't make a difrence that load is 0

outside %id %us and %sy values look normal


it seems that machines are getting there share of cpu (openssl speed runs as expected)

top - 12:20:00 up 11 days,  9:18,  0 users,  load average: 0.00, 0.01, 0.06
Tasks:  30 total,   1 running,  29 sleeping,   0 stopped,   0 zombie
Cpu0  : 92.4%us,  7.6%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
Cpu1  : 91.3%us,  8.7%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
Cpu2  : 89.9%us, 10.1%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
Cpu3  : 88.7%us, 11.3%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
Mem:   2621840k total,  2561584k used,    60256k free,        0k buffers
Swap:        0k total,        0k used,        0k free,  1194520k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND                                                                                       
    1 root      20   0  1988  632  536 S    0  0.0   0:07.23 init                                                                                           
    2 root      20   0     0    0    0 S    0  0.0   0:00.00 kthreadd/548                                                                                   
    3 root      20   0     0    0    0 S    0  0.0   0:00.00 khelper/548                                                                                    
  362 daemon    20   0  1772  460  356 S    0  0.0   0:00.00 portmap                                                                                        
  374 statd     20   0  1832  620  516 S    0  0.0   0:00.00 rpc.statd                                                                                      
  453 root      20   0 29044 1240  808 S    0  0.0   0:34.46 rsyslogd                                                                                       
  467 root      20   0  5280  872  516 S    0  0.0   0:03.79 sshd                                                                                           
  507 root      20   0  2704 1232 1004 S    0  0.0   0:00.00 mysqld_safe                                                                                    
  544 mysql     20   0  237m 109m 2840 S    0  4.3  91:31.54 mysqld                                                                                         
  545 root      20   0  1636  532  456 S    0  0.0   0:00.00 logger 


even if container is doing absolutely nothing top shows 90% + user utilization, this doesn't seem right

on rhel5 based kernels total cpu utilization numbers reflected the sum of cpu utilization and %id reflected free cpu

is this reflecting some change in cpu usage reporting inside the containers or is it a bug?

vps config follows
CPULIMIT="100"
PHYSPAGES="0:321536"
NUMIPTENT="200"
CPUUNITS="29153"
NOATIME="yes"
CPUS="4"
SWAPPAGES="0"


thanks in advance for any answers

Regards,
Michal Grzedzicki
Re: Strange top %us values on 55.10 kernel [message #46732 is a reply to message #46729] Wed, 06 June 2012 18:21 Go to previous messageGo to next message
disaster is currently offline  disaster
Messages: 158
Registered: February 2006
Senior Member
I see the same on 055.12
Re: Strange top %us values on 55.10 kernel [message #46733 is a reply to message #46729] Wed, 06 June 2012 18:28 Go to previous messageGo to next message
disaster is currently offline  disaster
Messages: 158
Registered: February 2006
Senior Member
But i also see this with 053.10
Re: Strange top %us values on 55.10 kernel [message #46761 is a reply to message #46729] Sun, 10 June 2012 21:11 Go to previous messageGo to next message
disaster is currently offline  disaster
Messages: 158
Registered: February 2006
Senior Member
This is fixed with 042stab056.11
Re: Strange top %us values on 55.10 kernel [message #46772 is a reply to message #46761] Tue, 12 June 2012 05:04 Go to previous messageGo to next message
lazy
Messages: 16
Registered: January 2008
Junior Member
thank You for the information,
but on 56.11 idle is still almost all the time 0%

top - 07:02:01 up  1:18,  0 users,  load average: 0.04, 0.03, 0.00
Tasks:  26 total,   1 running,  25 sleeping,   0 stopped,   0 zombie
Cpu(s):100.0%us,  0.0%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
Mem:  4294967292k total,   175132k used, 4294792160k free,        0k buffers
Swap:        0k total,        0k used,        0k free,   113240k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND                                                                                       
 1350 www-data  20   0 39104 8472 3236 S    0  0.0   0:03.72 apache2                                                                                        
 1482 www-data  20   0 39104 8332 3288 S    0  0.0   0:01.12 apache2                                                                                        
 1483 www-data  20   0 39104 8648 3244 S    0  0.0   0:02.47 apache2                                                                                        
 1510 www-data  20   0 39104 8192 3180 S    0  0.0   0:00.99 apache2  


Regards,
Michal Grzedzicki
Re: Strange top %us values on 55.10 kernel [message #46773 is a reply to message #46729] Tue, 12 June 2012 06:19 Go to previous messageGo to next message
disaster is currently offline  disaster
Messages: 158
Registered: February 2006
Senior Member
Just checked my VPS again and they're fine now. Do you run a custom kernel build or the official releases?
Re: Strange top %us values on 55.10 kernel [message #46774 is a reply to message #46773] Tue, 12 June 2012 06:32 Go to previous messageGo to next message
lazy
Messages: 16
Registered: January 2008
Junior Member
I'm running a custom build kernel config attached,


I have rechecked it and now top is switching from

100%us 0%id to 0%us 100%id
  • Attachment: cfg
    (Size: 24.99KB, Downloaded 202 times)
Re: Strange top %us values on 55.10 kernel [message #46775 is a reply to message #46729] Tue, 12 June 2012 06:35 Go to previous message
disaster is currently offline  disaster
Messages: 158
Registered: February 2006
Senior Member
perhaps you should try with an original config
Previous Topic: ploop yum update problem
Next Topic: Enquiry on CentOS 6 HW + CentOS 5 guest?
Goto Forum:
  


Current Time: Sat Apr 27 15:56:52 GMT 2024

Total time taken to generate the page: 0.02452 seconds