[one-users] Host information

[Tino Vazquez]

Hi Rangababu,

comments inline,



On Thu, Jan 14, 2010 at 9:02 PM, Rangababu Chakravarthula
<rbabu at hexagrid.com> wrote:
> Thanks Tino for the information. What we have observed is when we have only
> one host and we submit a VM allocation request with memory value more than
> the value given under free but less than  value given under free with
> buffers/cache,  opennebula  is failing to provision the VM.
> 1) Is it expected behavior?

OpenNebula's scheduler looks at the maximum memory available in the
host and then subtracts  the memory consumed by the VMs present in the
host. If the memory requested by the VM is less than this value, the
VM is submitted (look at include/HostShare.h).

>
> Also in the VM information that is displayed in onevm list and onevm show:
> 2) The memory value that is displayed there, is it allocated memory or used
> memory?

Memory shown in onevm list is the used memory, and memory shown in
onevm show under Template is the allocated memory.

>
> There was an email  last week regarding opennebula displaying 0 value for
> CPU for which your team clarified that the reason for that is libvirt is not
> reporting CPU value for KVM hypervisor.
> 3) How do we get CPU usage and memory usage of VM?

onevm list

> 4) Is it possible only if we have snmp inside the VM and use some NMS tool
> to gather that?

See above

> 5) Since VM is a process in the underlying hypervisor, if we get cpu and
> memory usage of the vm process, is it a true representation of VM and CPU
> usage of VM?

This is true with KVM

>
> Thanks

Hope it helps,

-Tino

>
> Ranga
>
> Tino Vazquez wrote:
>>
>> Hi Ranga,
>>
>> The free memory is calculated with the following line:
>>
>> $ free -k|grep "buffers\/cache"|awk '{print $4}'
>>
>> As you can see, this will get the 854328 figure out of the free output
>> provided, similar to the 853588 shown in the output.
>>
>> This value is the free memory counting the cached pages that can be
>> flushed.
>>
>> Regards,
>>
>> -Tino
>>
>> --
>> Constantino Vázquez, Grid Technology Engineer/Researcher:
>> http://www.dsa-research.org/tinova
>> DSA Research Group: http://dsa-research.org
>> Globus GridWay Metascheduler: http://www.GridWay.org
>> OpenNebula Virtual Infrastructure Engine: http://www.OpenNebula.org
>>
>>
>>
>> On Mon, Jan 11, 2010 at 11:14 PM, Rangababu Chakravarthula
>> <rbabu at hexagrid.com> wrote:
>>
>>>
>>> Hello
>>>
>>> onehost list gives me this information of a host
>>> HID NAME                      RVM   TCPU   FCPU   ACPU    TMEM    FMEM
>>> STAT
>>>  5 10.10.20.215                7       400       376      400
>>>  8185632
>>>  853588   on
>>>
>>> However when I logon to the host and execute free command, I see the
>>> following result.
>>>
>>>               total               used           free     shared
>>>  buffers
>>>    cached
>>> Mem:       8185632    7900788     284844          0       6688     562796
>>> -/+ buffers/cache:    7331304     854328
>>> Swap:            0          0          0
>>>
>>>
>>> The free memory seems to be very less than what onehost list is giving on
>>> the management node.
>>> Where does one get the value that it is displaying? Is it something it is
>>> not updating frequently?
>>>
>>> Ranga
>>> _______________________________________________
>>> Users mailing list
>>> Users at lists.opennebula.org
>>> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>>>
>>>
>
>



--
Constantino Vázquez, Grid Technology Engineer/Researcher:
http://www.dsa-research.org/tinova
DSA Research Group: http://dsa-research.org
Globus GridWay Metascheduler: http://www.GridWay.org
OpenNebula Virtual Infrastructure Engine: http://www.OpenNebula.org