VM migration use small amount of available bandwidth

Latest response

Hello,

We have RHEV on 10Gbps network infrastructure. Usually network transfers run cca 800-1000 MB/s, but when I migrate VM from one hypervisor to another, network usage is about 50-70 MB/s only...

I tried to set migration_max_bandwidth = 500 in vdsm.conf, but IMHO VDSM ignores this value...

Do you have any idea to speed up live migration on RHEV 3.2? Why is VM live migration too slow?

Thanks,
Milan

Responses

What is the version of vdsmd and libvirt running on hypervisors?

Did you restart vdsmd after setting migration_max_bandwidth = 500?

Yes, I restarted vdsm after vdsm.conf change.

[root@rhevh09 ~]# rpm -qa|grep vdsm
vdsm-cli-4.10.2-24.1.el6ev.noarch
vdsm-hook-vhostmd-4.10.2-24.1.el6ev.noarch
vdsm-4.10.2-24.1.el6ev.x86_64
vdsm-python-4.10.2-24.1.el6ev.x86_64
vdsm-xmlrpc-4.10.2-24.1.el6ev.noarch
vdsm-reg-4.10.2-24.1.el6ev.noarch

I t is the latest RHEV-H iso based on EL 6.4.

Please take a look at this https://access.redhat.com/site/solutions/108813 You have to maintenance the host and activate.

Sadique, thank you for this article, now it's clearer for me. I didn't know that the migration_max_bandwidth is in Mbites (I thought it is in MBytes :-) ). I will try to set it to higher value by the manual...

I noticed, a small VM (up to 4 vCPUs) migrates faster (about 200-300 MB/s), but when I add more vCPUs (16), the migration runs much slower (50-60 MB/s).

Thanks

Close

Welcome! Check out the Getting Started with Red Hat page for quick tours and guides for common tasks.