[one-users] why doesn't opennebula judge the host capacity except for actions related to scheduler?

[Ruben S. Montero]

Hi

Here:

Deploy
https://github.com/OpenNebula/one/blob/master/src/rm/RequestManagerVirtualMachine.cc#L486

Migrate (Live or not)
https://github.com/OpenNebula/one/blob/master/src/rm/RequestManagerVirtualMachine.cc#L603

Check function test the capacity, here:
https://github.com/OpenNebula/one/blob/master/src/rm/RequestManagerVirtualMachine.cc#L224

Cheers

Ruben


On Fri, Mar 15, 2013 at 3:09 AM, cmcc.dylan <dx10years at 126.com> wrote:

> Thanks Ruben!
>     I read the source code related to cold migration, but i don't find
> judgement places about CPU. That is to say, the test_capacity function
> isn't called in the cold migration. Is it right? If you please, can you
> poit out how opennebula does?
>
>    Thank you very much!
>
>
>
>
> At 2013-03-14 17:13:17,"Ruben S. Montero" <rsmontero at opennebula.org>
> wrote:
>
> Hi
>
> About the first problem (-2). This should be fixed in OpenNebula 3.8.3,
> are you using that version? If so can you send the output of onehost sh ow
> -x <host_id> before and after the migration.
>
> About the second problem (capacity checks). We have included the
> '--enforce' option to enforce that the host capacity is not exceeded for
> onevm migrate and livemigrate. (The new onevm resize command also accepts
> that options so you can enforce that the host capacity is not overcommitted
> when re-sizing the VM capacity).
>
> Cheers
>
> Ruben
>
>
> On Thu, Mar 14, 2013 at 6:07 AM, cmcc.dylan <dx10years at 126.com> wrote:
>
>> Hi everyone!
>>    I doubt about this problem for a long time and i also prove it using
>> some test case.
>>    Firstly, let's describe the problem as follows:
>>    I manage two host machines using opennebula and they both have 4 cores
>> in total. I deploy vm1 which has 2 cores in host1 and deploy vm2 which has
>> 4 cores in host2. After that, I migrate vm1 to host2, to my surprise it
>> success and the free cpu of host2 becomes -2.
>>
>>    so i read the source code related to virtual machine's migration, and
>> i find this is no judgement of host resources' capacity in the whole
>> migration process. Source codes as follows:
>>                        vm->get_requirements(cpu,mem,disk);
>>                        hpool->add_capacity(vm->get_hid(),cpu,mem,disk);
>>    And i also seek what actions does opennebula to test host resources'
>> capacity and find only two places and they also related to scheduler.For
>> example:
>>    (1)void Scheduler::match()
>>    (2)int VirtualMachineXML::get_host()
>> Only the two place call the funstion - test_capacity(cpu,mem,dsk) to
>> judge the host resources.
>>
>>    i'm sorry if i don't describe the problem clearly. Hope someone to
>> help and communication with me.
>>
>>
>>
>> _______________________________________________
>> Users mailing list
>> Users at lists.opennebula.org
>> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>>
>>
>
>
> --
> Ruben S. Montero, PhD
> Project co-Lead and Chief Architect
> OpenNebula - The Open Source Solution for Data Center Virtualization
> www.OpenNebula.org | rsmontero at opennebula.org | @OpenNebula
>
>
>
>


-- 
Ruben S. Montero, PhD
Project co-Lead and Chief Architect
OpenNebula - The Open Source Solution for Data Center Virtualization
www.OpenNebula.org | rsmontero at opennebula.org | @OpenNebula
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opennebula.org/pipermail/users-opennebula.org/attachments/20130315/a40b18c8/attachment-0002.htm>