On 01/16/2018 08:37 AM, Satish Patel wrote:
Thanks Brian,
I may having difficulty to understand that example, if I have only /24 public
subnet for cloud and have 200 compute node then how does it work?
The intention is to have a second subnet on the external network, but
only have it usable within the datacenter. If you create it and set the
service-type to only certain types of ports, like DVR, then it won't be
used for floating IP as the other one is.
-Brian
On Jan 15, 2018, at 9:47 PM, Brian Haley <haleyb....@gmail.com> wrote:
On 01/15/2018 01:57 PM, Satish Patel wrote:
I am planning to build openstack on production and big question is
network (legacy vs DVR) but in DVR big concern is number of Public IP
used on every compute node, I am planning to add max 100 node in
cluster in that case it will use 100 public IP for compute node, Ouch!
You can reduce this public IP consumption by using multiple subnets on the
external network, with one just for those DVR interfaces. See
https://docs.openstack.org/ocata/networking-guide/config-service-subnets.html
Example #2 for a possible configuration.
As for what type of L3 configuration to run, it seems like you have a good idea
of some of the trade-offs with each.
-Brian
If i use legacy compute node then it could be bottleneck or failure
node if not in HA.
what most of company use for network node? DVR or legacy?
_______________________________________________
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
_______________________________________________
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
