Instances cannot ping each other and cannot ping virtual-router

wodel youchi wodel.youchi at gmail.com
Mon Sep 13 12:07:53 UTC 2021


Hi,

After some reading, I found out that I needed to specify the physical port
that will be used to connect my infrastructure to the external world.
In my configuration I created two ovs bridges over two bonds :
br0 : is used for storage and storage management networks.
br1 : is used for api, tenant and external networks.

So I added this to my network-environment.yaml file :
*Neu*tronBridgeMappings: 'datacentre:br1'

And it did fix the majority of my connectivity problems, now the instances
can ping each other, the instances can ping the internet. I can ping the
external vrouter interface, but when I associate a floating IP with an
instance, I cannot ping or ssh that instance from the external network.
I have any to any security group rules for icmp and ssh (for test). How can
I debug that?

I have some other questions :
What does mean technically these variables ?

1) What is the difference between these two lines?
NeutronNetworkVLANRanges: 'datacentre:1:4000'
NeutronNetworkVLANRanges: 'datacentre:1:1000,tenant:500:1000'

2) What is the difference between NeutronNetworkType and NeutronTunnelType

Regards.

Le mer. 8 sept. 2021 à 17:13, wodel youchi <wodel.youchi at gmail.com> a
écrit :

> Hi,
>
> I deployed OpenStack Train using TripleO using this tutorial :
> https://kdjlab.com/deploying-rdo-in-a-cohesive-manner/ and the
> documentation of TripleO.
> I deployed it with DVR.
>
> In my deployment I am using virtual machines with nested-kvm.
>
> The deployment went well, I am using network isolation like this :
> - nic1 : provisioning
> - nic2 and nic3 (bond0) storage and storage mgmt networks, each one in
> it's VLAN
> - nic3 and nic5 (bond1) tenant, api and *external* (10.0.2.0/24 VLAN2100)
> networks, each one in it's VLAN
>
> In my physical host (the bare metal KVM) I created a bridge which handles
> the provisioning, tenant, api and external networks.
>
> I created a private tenant network (172.16.100.0/24).
>
> openstack network create private
> neutron subnet-create private 172.16.100.0/24 --name private-sub --dns-nameserver 172.16.0.252
>
>
> I created a public network and I attached it to the external network using
> the same VLAN tag (10.0.2.0/24 VLAN 2100, pool: 10.0.2.100-10.0.2.120) :
>
> *openstack network create --provider-network-type vlan --provider-physical-network datacentre --provider-segment 2100 --external public*
> neutron subnet-create public 10.0.2.0/24 --name public-sub --disable-dhcp --allocation-pool=start=10.0.2.100,end=10.0.2.120 --gateway=10.0.2.1 --dns-nameserver 172.16.0.252
>
>
>
> I created a vrouter, one port in the public network and the other in the
> private network.
> I created two cirrus instances, each one got it's ip address from the
> private network.
>
> I found :
> cirrus-1 : 172.16.100.81
> cirrus-2 : 172.16.100.103
> vrouter : 172.16.100.1 private
>             : 10.0.2.101  external
> neutron:dhcp : 172.16.100.2
>
> The problems :
> - The instances cannot ping each other.
> - The instances cannot ping the vrouter.
> - I cannot ping the public vrouter interface.
>
> But both instances can ping neutron:dhcp
>
> Could someone help me dig into this.
>
> Thanks in advance, Regards.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-discuss/attachments/20210913/351a6b2a/attachment.htm>


More information about the openstack-discuss mailing list