Difference between revisions of "util-vserver:Cgroups"

From Linux-VServer

Jump to: navigation, search
(New page: == Draft - Cgroups and vservers == to make cgroups works with vserver your kernel must have the following settings: <pre> CONFIG_CGROUPS=y # CONFIG_CGROUP_DEBUG is not set # CONFIG_CGROU...)
 
Line 33: Line 33:
 
<br/>
 
<br/>
 
Then you must create on you guest /etc/vservers/<guest>/cgroup/ all the files put here will be copied to the guest cgroup when created. The doc for cgroup is not on the great flower page but in the source of the latests utils. You can have a default cgroup directory for all guest that can be overrided by each guest's cgroup directory.
 
Then you must create on you guest /etc/vservers/<guest>/cgroup/ all the files put here will be copied to the guest cgroup when created. The doc for cgroup is not on the great flower page but in the source of the latests utils. You can have a default cgroup directory for all guest that can be overrided by each guest's cgroup directory.
 +
 +
== Draft - Distributing cpu shares with cgroups ==
 +
 +
This is simply done by adjusting the cpu.shares. Just do:
 +
 +
echo '512' > /dev/cgroup/<guest name>/cpu.shares
 +
 +
The share you get is equal to the guest's share divided by the sum of the cpu shares of all the guest. So for exemple :
 +
 +
<pre>
 +
vserver guest 1 => 512 
 +
vserver guest 2 => 512
 +
vserver guest 3 => 2048
 +
vserver guest 4 => 512
 +
</pre>
 +
 +
then you get :
 +
 +
<pre>
 +
vserver guest 1 => 512 / 3584 = 14%  cpu
 +
vserver guest 2 => 512 / 3584 = 14%  cpu
 +
vserver guest 3 => 2048 / 3584 = 57% cpu
 +
vserver guest 4 => 512 / 3584 = 14%  cpu
 +
</pre>
 
<br/>
 
<br/>
 +
 +
 +
Note that this is fair scheduling and this will not enfore HARD limit (as far as i know).
  
 
Regards,
 
Regards,
 
Ghislain.
 
Ghislain.

Revision as of 09:01, 27 February 2009

Draft - Cgroups and vservers

to make cgroups works with vserver your kernel must have the following settings:

CONFIG_CGROUPS=y
# CONFIG_CGROUP_DEBUG is not set
# CONFIG_CGROUP_NS is not set
CONFIG_CGROUP_FREEZER=y
CONFIG_CGROUP_DEVICE=y
CONFIG_CGROUP_SCHED=y
CONFIG_CGROUP_CPUACCT=y
CONFIG_CGROUP_MEM_RES_CTLR=y


Or in the menu:

[*] Control Group support                               [ ]   Example debug cgroup subsystem
[ ]   Namespace cgroup subsystem
[*]   control group freezer subsystem
[*]   Device controller for cgroups
[*]   Cpuset support
[*] Group CPU scheduler
[*]   Group scheduling for SCHED_OTHER
[ ]   Group scheduling for SCHED_RR/FIFO
      Basis for grouping tasks (Control groups)  --->
[*] Simple CPU accounting cgroup subsystem
[*] Resource counters
[*] Memory Resource Controller for Control Groups


Then you must create on you guest /etc/vservers/<guest>/cgroup/ all the files put here will be copied to the guest cgroup when created. The doc for cgroup is not on the great flower page but in the source of the latests utils. You can have a default cgroup directory for all guest that can be overrided by each guest's cgroup directory.

Draft - Distributing cpu shares with cgroups

This is simply done by adjusting the cpu.shares. Just do:

echo '512' > /dev/cgroup/<guest name>/cpu.shares

The share you get is equal to the guest's share divided by the sum of the cpu shares of all the guest. So for exemple :

vserver guest 1 => 512   
vserver guest 2 => 512
vserver guest 3 => 2048
vserver guest 4 => 512

then you get :

vserver guest 1 => 512 / 3584 = 14%  cpu
vserver guest 2 => 512 / 3584 = 14%  cpu
vserver guest 3 => 2048 / 3584 = 57% cpu
vserver guest 4 => 512 / 3584 = 14%  cpu



Note that this is fair scheduling and this will not enfore HARD limit (as far as i know).

Regards, Ghislain.

Personal tools