OpenVZ Forum


Home » General » Support » OpenVZ on OVH vRack (Multiple subnet on same IF)
OpenVZ on OVH vRack (Multiple subnet on same IF) [message #51441] Thu, 05 June 2014 12:43 Go to next message
loki is currently offline  loki
Messages: 6
Registered: June 2014
Location: Canada
Junior Member
Hi,

We have servers on OVH using vRack technology with OpenVZ and SolusVM.

With our xen servers, we can use these blocs without doing any changes to eth1, and specify IP Addr with correct subnet / gateway on each VM. That allows us to move any vm to any server and keep the same public address.

I have been trying for 2 days , to setup 2 FO IP Blocs for use by our SolusVM servers.

So it looks like :

/28 |--\
/29 |---------OVH-vRack1.5 --------eth1----SolusVM(OpenVZ)Server #1


In short, 2 IP blocs, assigned to the same interface, on the same server (eventualy more servers as we scale on that technology).


So, i can make things work without issues if i setup an IP address from either the /28 or /29 bloc to eth1 , and then bind vz to eth1. but it has the following problems:

Cannot use more then 1 blocs with OpenVZ
Has to waste 1 ip from every bloc i would eventually add

Does anyone have a suggestion / idea of what could be a setup for that ? If we start OpenVZ witouth an IP assigned to eth1 device, of course, we get the following starting vm's : vps-net_add ERROR: Unable to get source ip [ETH1]


Thanks!
Re: OpenVZ on OVH vRack (Multiple subnet on same IF) [message #51500 is a reply to message #51441] Sun, 29 June 2014 21:10 Go to previous messageGo to next message
devonblzx is currently offline  devonblzx
Messages: 127
Registered: December 2006
Senior Member
What is NEIGHBOR_DEVS set to in /etc/vz/vz.conf?

http://static.openvz.org/userbars/openvz-user-2.png
ByteOnSite President
Re: OpenVZ on OVH vRack (Multiple subnet on same IF) [message #51501 is a reply to message #51500] Mon, 30 June 2014 01:30 Go to previous message
loki is currently offline  loki
Messages: 6
Registered: June 2014
Location: Canada
Junior Member
For some reason one of the routers at the DC flushe ARP cache every 4 hours.

Now i have scripted something to do arp_send for every running ct's (similar to what happens in the vzctl start process)

This script will run on an hourly cron, never lost connectivity after that.
Previous Topic: Dream Request - IO Limiting per process
Next Topic: After upgrade, host reboots when trying to start container
Goto Forum:
  


Current Time: Fri Nov 08 01:11:05 GMT 2024

Total time taken to generate the page: 0.03403 seconds