Open Stack

There was an issue with the order of the agents starting up that led to
some issues with L3 HA that was fixed with a systemd notification here:
https://review.openstack.org/#/c/254920/ However, I assume there had to be
a corresponding packaging patch that would make the the other agents wait
for that notification before starting.

I think a way to test if that would fix your issue is to restart the DHCP
agent and L3 agent and see if that fixes your broken resources.

On Mon, Feb 29, 2016 at 8:48 AM, Sergio Morales Acuña <semoac at gmail.com>
wrote:

> Hi all.
> I'm facing a problem with "Neutron/OVS/DVR" in "Liberty".
>
> With 3 network/compute nodes I created, successfully, 100 tenants, 100
> networks (1 per tenants), 3 DHCP agents per network and 100 routers (1 per
> network), 100 VM (1 per tenant) and 1 FIP per VM. All working and
> responding.
>
> The problem occurs after restarting the servers. Now i have some routers
> without communications, dhcp agents not connected to the tenant network and
> VM running but not connected to the router/fip. This happen randomly for
> any router/dhcp agent/ or VM.
>
> There's a way to force a port binding check in neutron? How i can make the
> restart process more effective?
>
> I'm using Neutron 7.0.1 (DVR), RHEL 7.2, Kolla with RDO Liberty and  OVS
> 2.3.2.
>
> I tried neutron 7.0.4 (from stable/libery) with OVS 2.3.2 and OVS 2.4.0.
>
> I would appreciate any comments, suggestions or patches under review.
>
> __________________________________________________________________________
> OpenStack Development Mailing List (not for usage questions)
> Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:unsubscribe
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20160229/7add1497/attachment.html>