[rhelv6-list] Using specific physical interface to migrate vms

[Gianluca Cecchi]

On Fri, Apr 15, 2011 at 9:44 AM, carlopmart <carlopmart at gmail.com> wrote:
> On 04/15/2011 12:57 AM, Gianluca Cecchi wrote:
>>
>> On Thu, 14 Apr 2011 12:44:39 +0200 carlopmart wrote:
>>>
>>> How can I configure cluster.conf file to assign eth1 interface when vm
>>> live migration will be required??

[snip]

>
> Thanks Gianluca.
>

You are welcome.
Keep in mind that you have also to:

a) set up ssh root equivalence on the hostnames you are going to use
for migration (hostlive1 and hostlive2 in my previous example)
otherwise a command such as
# clusvcadm -M vm:guest_vm -m node2
will get an error like this
Trying to migrate vm:guest_vm to node2...Failed; service running on
original owner
and in messages:
rgmanager[7861]: error: cannot recv data: Permission denied, please
try again.#015

b) setup your firewall rules on both hosts to accept tcp connections
in the range 49152:49215 from the other one ip chosen for migration
otherwise it will fail with something like this in log file:

Apr 15 10:48:56 host2 rgmanager[8387]: Migrate guest_vm to hostlive1 failed:
Apr 15 10:48:56 host2 rgmanager[8409]: error: operation failed:
Migration unexpectedly failed
Apr 15 10:48:56 host2 rgmanager[2793]: migrate on vm "guest_vm"
returned 150 (unspecified)

(node the hostlive1, while by default you will see node1, aka the
intracluster name)

HIH,
Gianluca