[one-users] Creating new VM fails
Javier Fontan
jfontan at fdi.ucm.es
Mon May 25 02:51:50 PDT 2009
Hello,
On May 22, 2009, at 8:07 PM, cmasmas cmasmas wrote:
> That was the problem. Thanks. I was requesting too much RAM. The
> point is that each node has around 1.5 GB of free RAM. However,
> opennebula detects about 100MB free on one noe, and 2 MB on the other.
>
> HID NAME RVM TCPU FCPU ACPU TMEM
> FMEM STAT
> 1 pc-GridLab00 0 200 199 200 2086912
> 2048 on
> 3 pc-GridLab01 1 200 100 180 2069504
> 24576 on
>
> Why is this?
This happens because Dom-0 of xen is taking the rest of the ram. You
can decrease the ram it is taking form the command like using this
command:
$ xm mem-max 0 256
It will change the amount of ram used by dom-0 to 256 Mb and the rest
will be free for VM's.
If you want to set this parameter automatically on system startup /
boot/grub/menu.lst can be modified so the kernel parameter looks
something like this:
kernel /xen-3.2.gz dom0_mem=256M
Just add the dom0_mem parameter, leave the path of your xen hypervisor
as it is in your menu.lst.
> Another thing is that although one shows that the VM is running, xen
> reveals that is really blocked.
>
> ID NAME STAT CPU MEM HOSTNAME TIME
> 21 vm-examp runn 0 102400 pc-GridLab01 00 00:21:34
That is normal, blocked does not mean that the machine is stopped but
that in the moment that you call xm command is not in the cpu
executing instructions. It is like processes, you can have a bunch of
processes running but only a number of them can be in the CPU
executing instructions (the same as the number of cores you have in
the computer). Also your machine is idling as you can see in the CPU
value from monitoring so catching it doing something will be hard.
Bye
--
Javier Fontan, Grid & Virtualization Technology Engineer/Researcher
DSA Research Group: http://dsa-research.org
Globus GridWay Metascheduler: http://www.GridWay.org
OpenNebula Virtual Infrastructure Engine: http://www.OpenNebula.org
More information about the Users
mailing list