cloudstack-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Melanie Desaive <m.desa...@heinlein-support.de>
Subject Physical Network Setup and Labels for mixed XenServer / KVM Infrastructure
Date Wed, 14 Aug 2019 14:21:07 GMT
Hello,

we plan to change from XenServer to KVM virtualisation for Apache
CloudStack 4.11 with advanced networking.

I am currently installing an KVM proof of concept und would like to
integrate two KVM hosts for testing as a new cluster to CloudStack.

I got some questions concerning the networking setup for the zone.

I plan to use OpenVSwitch on the KVM virtualization hosts.

I understand, that I can now, once define the labels for the physical
network of the zone for KVM. I do now have the opportunity to decide
newly about the division between the different network types (e.g.
public, guest, management, storage). I do not have to keep the
decisions made for XenServer. Is this correct?

For XenServer we did use one LACP trunk, that carries all traffic. On
this LACP trunk we have one bridge/network, named "LACPTRUNK" which is
used for all four different traffic categories.

----
uuid ( RO)                : ca88f7f8-2ce0-b3ea-b218-26336ee6496e
          name-label ( RW): LACPTRUNK
    name-description ( RW): 2 x 10GBit/s LACP Dynamic über eth2 und
eth3 via openvswitch
              bridge ( RO): xapi1
----

I think that it could be a good idea, to handle this differently in 
KVM and would like to ask your opinion.

I think about preparing four bridges on the KVM hosts. One bridge for 
each out of (public, guest, management, storage). Problem is, that in 
our actual hardware configuration I do only have one physical (lacp bond) 
port on the virtualization hosts. I would like to try to use the syntax suggested in 

http://docs.cloudstack.apache.org/en/latest/installguide/hypervisor/kvm.html#configure-the-network-using-openvswitch

preparing several "fake bridges" on the main LACP bond interface. Never
nested ovs bridges up to now, do I have to expect something unexpected?

By splitting up the bridges, I could open the opportunity to later set
up virtualization hosts with more physical network interface and
dedicate interfaces to traffic types. Using a different OVS
configuration for future KVM clusters. Is that correct?

What do I have to keep in mind regarding the surrounding network
infrastrukture and VLAN configuration for the switch ports the hosts
are attached to, when assigning more labels for KVM hosts, than I used
with XenServer?

Do I have to expect different requirements for the switches? I would
expect the same switch port setup beeing valid for KVM and XenServer. 

Thank you all and best greetings,

Melanie 



---

-- 
-- 
Heinlein Support GmbH
Schwedter Str. 8/9b, 10119 Berlin
  
https://www.heinlein-support.de
 
Tel: 030 / 40 50 51 - 62
Fax: 030 / 40 50 51 - 19
  
Amtsgericht Berlin-Charlottenburg - HRB 93818 B
Geschäftsführer: Peer Heinlein - Sitz: Berlin

Mime
View raw message