[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

[virt-tools-list] live migrate ok with virsh, ko with virt-manager



hello,
using 2 hosts with F11 x86_64 and its virt components:
# rpm -qa|grep virt
virt-viewer-0.0.3-6.fc11.x86_64
virt-top-1.0.3-4.fc11.x86_64
libvirt-0.6.2-13.fc11.x86_64
python-virtinst-0.400.3-8.fc11.noarch
libvirt-python-0.6.2-13.fc11.x86_64
virt-df-1.0.64-2.fc11.x86_64
virt-manager-0.7.0-5.fc11.x86_64

# rpm -qa|grep qemu
qemu-common-0.10.5-3.fc11.x86_64
qemu-kvm-0.10.5-3.fc11.x86_64
qemu-system-x86-0.10.5-3.fc11.x86_64
qemu-img-0.10.5-3.fc11.x86_64


the 2 hosts are configured witch qemu+tls connections between them.
vmA created on host1 using virt-manager
vmB created on host2 using virt-manager

both vmA and vmB have one network card that is on br0 (physical
bridge) configured on both hosts
I'm able to do:
- a live migration of vmA from host 1 to host2 using virsh
(command like "migrate --live dom_id qemu+tls://host2.mydomain.it/system")
and verify that is indeed a live migration being connected from an
external client....
- a live migration of vmB from host2 to host1 using virsh

I'm not able to do:
- the same live migration as above if using virt-manager. I get this
error inside a window
libvirtError: invalid argument in only tcp URIs are supported for KVM
migrations

- after having done a live migration of vmA with virsh from host1 to
host2, if I try the reverse on the same vm, the commands never
completes (tried only once actually). I can see that:
an strace of the virsh pid on the source host gives:
 # strace -p 5995
Process 5995 attached - interrupt to quit
restart_syscall(<... resuming interrupted call ...>

Can I do a ctrl-C of the migrate command that seems blocked (or
waiting for something)?

Questions:
- when migrating from virsh and keeping virt-manager windows open, I
can see that on the host from where the domain has been live migrated,
the line with the vm remains with status shutoff and without a domain
id. Is this correct/expected or should the line go away at all from
the source host?
- if the answer to the question above is yes, when/if I re-migrate a
vm to its original host, is it supposed to mantain the same previous
domain id?
How are these domain ids managed internally between hosts?
- could it help enabling the fedora-virt-preview repo?

Thanks in advance and sorry for the many topics, but they seem sort of
concatenated...

Gianluca


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]