[ovirt-users] Re: nul interface error when engine is creating ovirtmgmt bridge post-deployment
On Wed, 10 Oct 2018 16:27:38 -0000
"Brad Riemann" wrote:
This is correct.
The physical interface (or bridge) which has the IP address used to add
the host is used for ovirtmgmt during the initial host setup.
For some reason oVirt does not detect the NIC or birdge which has the expected
IP address.
Is there an physical interface on the host with the IP address used to add
the host to oVirt?
Would you please share the line containing
[api.host] FINISH getCapabilities return
of /var/log/vdsm.log on the host and the IP address used to add the host to oVirt?
Ah.. crude.. yeah the node is behind a nat too, so it's probably looking at
the public ip and can't find it so errors out.
I'll send you an email directly with this line, I'm not to keen on letting our
subnets be shown on the net, but you can share among the rhel team if need be.
OVN will not work, because the ovn-controller on the host has to
connect to ovn-central on Engine's host.
To prevent ovn-controller to try to connect to the ovn-central, the
default network provider should be disabled in this cluster.
Based on your last comment in the response i'll bet that being behind the nat answers
my issue in general. Supplying the logs anyway for sanity but i'm betting that because
i'm initiating the update of the logical networks from the GUI it takes a different
approach to push down the changes than the initial setup? The more I think about it the
less sense it makes since it SHOULD use the same delivery methods, no?
And before anyone calls me ridiculous for having this connectivity go through a nat on
both sides, I know, I inherited the methodology and am trying to work within confines that
I am allowed.
Anywho, here goes, thanks for the reply appreciate the feedback GREATLY.