[Users] Starting With OpenVZ and Having Interface Issues

[Kir Kolyshkin]

On 02/05/2014 03:23 PM, Matt wrote:
> I just installed OpenVZ on CentOS 6.x 64 bit following this guide a
> few days ago.
>
> http://openvz.org/Quick_Installation_CentOS_6
>
> I am mostly interested in creating containers with veth interfaces so
> I can assign multiple public IPv4 and IPv6 addresses from inside the
> container.  I noticed that when I am logged into the CentOS physical
> server with putty through ssh I can start a container that has a veth
> interface with 'vzctl start 102' and it starts just fine.  I can ping
> its IPv4 and IPv6 addresses just fine.  However when I issue 'vzctl
> stop 102' my putty ssh connection to the OpenVZ server is dropped and
> I must reconnect.  I am able to reconnect just fine.  Again, I am
> logged into the OpenVZ master not the container.

I can't help much in here, since you are following the unofficial guide
(sorry for that, I just marked it as such on wiki), but it looks like it is
not correct to add your primary NIC device right to the bridge you are
using for containers. I don't remember details at the moment, but
it looks like in such setup you have the effect you are complaining about.
It's something about bridge state and the state of devices that are in
(or change in their state, i.e. ups and downs).

Again, I am not quite sure how to configure this to achieve what you want
(I'd just use venet and --ipadd/--ipdel to set up in-CT IPs), but I remember
having good experience with the virbr0 device which is pre-configured
on RHEL/CentOS (for use by KVM guests), with all bells and whistles 
including
a DHCP server. If this is not what you want, you have to dig deeper into how
bridged networking and bridge devices work.

Hope that helps.

>
> Can anyone tell me what is going on here?  After further testing I
> determined it does not always do it.  The first time I stop a
> container after I reboot the master, physical server, it seems most
> prone to act up.
>
> ######### just did reboot 30 minutes ago or so............
>
> [root at ex1 ~]# vzlist
>        CTID      NPROC STATUS    IP_ADDR         HOSTNAME
>         103         92 running   -               server.example456.com
> [root at ex1 ~]# vzctl start 102
> Starting container...
> Container is mounted
> Setting CPU units: 1000
> Setting CPUs: 1
> Configure veth devices: veth102.0
> Adding interface veth102.0 to bridge vmbr0 on CT0 for CT102
> Container start in progress...
> [root at ex1 ~]# vzlist
>        CTID      NPROC STATUS    IP_ADDR         HOSTNAME
>         102         19 running   -               server102.example.com
>         103         92 running   -               server.example456.com
> [root at ex1 ~]# vzctl stop 102
> Stopping container ...
>
> ############## Here is where I get kicked off putty/ssh......
> _______________________________________________
> Users mailing list
> Users at openvz.org
> https://lists.openvz.org/mailman/listinfo/users