Please include ovirt-users list in your reply, to share the knowledge and experience with the community!

On Fri, Sep 11, 2020 at 12:12 PM Konstantinos Betsis <k.betsis@gmail.com> wrote:

Ok below the output per node and DC

DC01

node01

[root@dc01-node01 ~]# ovs-vsctl --no-wait get open . external-ids:ovn-remote

"ssl:OVIRT_ENGINE_IP:6642"

[root@ dc01-node01 ~]# ovs-vsctl --no-wait get open . external-ids:ovn-encap-type

geneve

[root@ dc01-node01 ~]# ovs-vsctl --no-wait get open . external-ids:ovn-encap-ip

"OVIRTMGMT_IP_DC01-NODE01"

node02

[root@dc01-node02 ~]# ovs-vsctl --no-wait get open . external-ids:ovn-remote

"ssl:OVIRT_ENGINE_IP:6642"

[root@ dc01-node02 ~]# ovs-vsctl --no-wait get open . external-ids:ovn-encap-type

geneve

[root@ dc01-node02 ~]# ovs-vsctl --no-wait get open . external-ids:ovn-encap-ip

"OVIRTMGMT_IP_DC01-NODE02"

DC02

node01

[root@dc02-node01 ~]# ovs-vsctl --no-wait get open . external-ids:ovn-remote

"ssl:OVIRT_ENGINE_IP:6642"

[root@ dc02-node01 ~]# ovs-vsctl --no-wait get open . external-ids:ovn-encap-type

geneve

[root@ dc02-node01 ~]# ovs-vsctl --no-wait get open . external-ids:ovn-encap-ip

"OVIRTMGMT_IP_DC02-NODE01"

Looks good.

 

DC01 node01 and node02 share the same VM networks and VMs deployed on top of them cannot talk to VM on the other hypervisor.

Maybe there is a hint on ovn-controller.log on dc01-node02 ? Maybe restarting ovn-controller creates more helpful log messages?

You can also try restart the ovn configuration on all hosts by executing

vdsm-tool ovn-config OVIRT_ENGINE_IP LOCAL_OVIRTMGMT_IP

on each host, this would trigger

https://github.com/oVirt/ovirt-provider-ovn/blob/master/driver/scripts/setup_ovn_controller.sh

internally.

 

So I would expect to see the same output for node01 to have a geneve tunnel to node02 and vice versa.

Me too.

 

On Fri, Sep 11, 2020 at 12:14 PM Dominik Holler <dholler@redhat.com> wrote:

On Fri, Sep 11, 2020 at 10:53 AM Konstantinos Betsis <k.betsis@gmail.com> wrote:

Hi Dominik

OVN is selected as the default network provider on the clusters and the hosts.

sounds good.

This configuration is required already during the host is added to oVirt Engine, because OVN is configured during this step.

 

The "ovn-sbctl show" works on the ovirt engine and shows only two hosts, 1 per DC.

Chassis "c4b23834-aec7-4bf8-8be7-aa94a50a6144"

    hostname: "dc01-node02"

    Encap geneve

        ip: "X.X.X.X"

        options: {csum="true"}

Chassis "be3abcc9-7358-4040-a37b-8d8a782f239c"

    hostname: "dc02-node1"

    Encap geneve

        ip: "A.A.A.A"

        options: {csum="true"}

The new node is not listed (dc01-node1).

When executed on the nodes the same command (ovn-sbctl show) times-out on all nodes.....

The output of the /var/log/openvswitch/ovn-conntroller.log lists on all logs

2020-09-11T08:46:55.197Z|07361|stream_ssl|WARN|SSL_connect: unexpected SSL connection close

Can you please compare the output of

ovs-vsctl --no-wait get open . external-ids:ovn-remote

ovs-vsctl --no-wait get open . external-ids:ovn-encap-type

ovs-vsctl --no-wait get open . external-ids:ovn-encap-ip

of the working hosts, e.g. dc01-node02, and the failing host dc01-node1?

This should point us the relevant difference in the configuration.

Please include ovirt-users list in your replay, to share the knowledge and experience with the community.

 

Thank you

Best regards

Konstantinos Betsis

On Fri, Sep 11, 2020 at 11:01 AM Dominik Holler <dholler@redhat.com> wrote:

On Thu, Sep 10, 2020 at 6:26 PM Konstantinos B <k.betsis@gmail.com> wrote:

Hi all

We have a small installation based on OVIRT 4.3.
1 Cluster is based on Centos 7 and the other on OVIRT NG Node image.

The environment was stable till an upgrade took place a couple of months ago.
As such we had to re-install one of the Centos 7 node and start from scratch.

To trigger the automatic configuration of the host, it is required to configure ovirt-provider-ovn as the default network provider for the cluster before adding the host to oVirt.

 

Even though the installation completed successfully and VMs are created, the following are not working as expected:
1. ovn geneve tunnels are not established with the other Centos 7 node in the cluster.
2. Centos 7 node is configured by ovirt engine however no geneve tunnel is established when "ovn-sbctl show" is issued on the engine.

Does "ovn-sbctl show" list the hosts?

 

3. no flows are shown on the engine on port 6642 for the ovs db.

Does anyone have any experience on how to troubleshoot OVN on ovirt?

/var/log/openvswitch/ovncontroller.log on the host should contain a helpful hint.

 

Thank you
_______________________________________________
Users mailing list -- users@ovirt.org
To unsubscribe send an email to users-leave@ovirt.org
Privacy Statement: https://www.ovirt.org/privacy-policy.html
oVirt Code of Conduct: https://www.ovirt.org/community/about/community-guidelines/
List Archives: https://lists.ovirt.org/archives/list/users@ovirt.org/message/LBVGLQJBWJF3EKFITPR72LBPA5A43WWW/