Thanks Daniel for reply. I'll go through directions of cgroup. I was just curious to know, what happens if I spawn 2 LXC containers, and load 1 container with lots of CPU intensive processes causing it's CPU usage to exceed beyond some threshold which may lead to 'denial / delay of service' type of attack to processes on 2nd LXC container. Whether this thing is possible at present ?<br>
<br>thanks<br>Sagar<br><br><div class="gmail_quote">On Tue, Nov 9, 2010 at 7:22 AM, Daniel Lezcano <span dir="ltr"><<a href="mailto:daniel.lezcano@free.fr">daniel.lezcano@free.fr</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
<div><div></div><div class="h5">On 11/08/2010 10:22 PM, Sagar Dixit wrote:<br>
<blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
Hi,<br>
<br>
I am curious to know if there is fine grained resource control in LXC<br>
containers. For ex. Can I allocate some fixed part of CPU (some %) to a<br>
container and control this configuration while creating LXC container ? Or<br>
can I update it dynamically ? Similarly, can I specify such control<br>
parameters for memory, disks etc ?<br>
<br>
If no, can any body give me direction as to where to look in the code to<br>
start implementing it ?<br>
I am keen to explore this feature in LXC.<br>
<br>
</blockquote>
<br></div></div>
You have to look at the direction of the cgroups:<br>
<br>
The existing cgroups are:<br>
<br>
* memory + swap<br>
* disk io<br>
* scheduler<br>
* cpuset<br>
* devices whitelist<br>
* cpu accounting<br>
* network bandwidth<br>
* freezer<br>
<br>
There is not cpu usage thresold and disk quota per container is not yet implemented.<br>
</blockquote></div><br>