Hi,

We have installed and added a new hypervisor into the ovirt cluster but this time with disabled HT.

I migrated a RedHat 5.10 machine to it and immediately the qemu-kvm process running the vm (freshly installed, just basic packages) began to consume 20-40% CPU as showed running top on the hypervisor.

Now that I have a hypervisor to run tests in, what would you suggest the next step is?

Thank you.

Xavier

On Thu, Oct 30, 2014 at 12:38 PM, Xavier Naveira <xnaveira@gmail.com> wrote:
Yeah, that's definitively worth a try. The problem is that we have been running on ovirt for some months now and haven't realized the problem with the RHEL 5.x vms until recently, when we began to import them (until now we mainly run RHEL 6.x vms), so the ovirt hypervisors are production machines and I don't have a lot of margin doing tests on them. Hopefully next week we'll be able to decomission another of the "old ones" and then we can do some testing before adding it to the ovirt cluster... The other thing that we're going to test is disabling HT.

X

On Thu, Oct 30, 2014 at 12:33 PM, Gianluca Cecchi <gianluca.cecchi@gmail.com> wrote:
On Thu, Oct 30, 2014 at 12:25 PM, Xavier Naveira <xnaveira@gmail.com> wrote:
These hardware is exactly the same, no changes in bios, reinstalled from scratch.

The versions in the "old ones" are:

RHEL 5.6
kvm-83-224.el5
libvirt-0.8.2-15.el5_6.1
kmod-kvm-83-224.el5
kvm-tools-83-224.el5
libvirt-0.8.2-15.el5_6.1
Kernel: 2.6.18-238.1.1.el5 #1 SMP Tue Jan 4 13:32:19 EST 2011 x86_64 x86_64 x86_64 GNU/Linux

I the ovirt "new" ones:

RHEL 6.5
qemu-img-rhev-0.12.1.2-2.415.el6_5.14.x86_64
qemu-kvm-rhev-0.12.1.2-2.415.el6_5.14.x86_64
qemu-kvm-rhev-tools-0.12.1.2-2.415.el6_5.14.x86_64
libvirt-0.10.2-29.el6_5.12.x86_64
libvirt-client-0.10.2-29.el6_5.12.x86_64
Kernel: 2.6.32-431.23.3.el6.x86_64 #1 SMP Wed Jul 16 06:12:23 EDT 2014 x86_64 x86_64 x86_64 GNU/Linux

I haven't gone into the details but the command line generated by our old software is like 2 lines of text while in the ovirt nodes for the same type of machine it is 10 :)




So one "important" thing is that hypervisor OS was RHEL 5 with plain Qemu/KVM and instead RHEL 6 with oVirt....
You could check if plain Qemu/KVM with RH EL 6.5 generates the same problems for RHEL 5.x VMs..
Not that it would solve the problem itself, but it would help to put the different OS version of hypervisor as a possible cause