OpenVZ tech issue

[tyaakow]

I'm installing OpenVZ on my dedi server.
I followed all the instructions in the OpenVZ wiki quick install guide.

I installed thru yum (my host box is on centos6)
I need to notice that yum didn't update grub.conf as supposed, but I needed to do it manually.
It all went fine, however after rebooting server into openvz kernel, and trying to start a container,
I always got this output (host is amd64, and i tried 64 and 32 bit templates, debian & centos)

Code:

    [root@cloud2 ~]# vzctl start 86
    Starting container ...
    Container is mounted
    Container start failed (try to check kernel messages, e.g. "dmesg | tail")
    Container is unmounted
    [root@cloud2 ~]# dmesg | tail
    [ 1973.401596] CT: 86: failed to start with err=-105
    [ 2107.113850] Failed to initialize the ICMP6 control socket (err -105).
    [ 2107.155523] CT: 86: stopped
    [ 2107.155543] CT: 86: failed to start with err=-105
    [ 6348.282184] Failed to initialize the ICMP6 control socket (err -105).
    [ 6348.330348] CT: 86: stopped
    [ 6348.330361] CT: 86: failed to start with err=-105
    [45184.024002] Failed to initialize the ICMP6 control socket (err -105).
    [45184.072086] CT: 86: stopped
    [45184.072099] CT: 86: failed to start with err=-105
    [root@cloud2 ~]#

Is anyone savy enough to give me an advice?

I really like the container technology and I really want to make it work.

[f8ll]

Is the ipv6 module loaded or compiled into the kernel?

[ninetailfox97]

search the tutorials, deathproxy made a tutorial about this i believe.

[DeathProxy]

My tutorial was for Xen VPSes. It's a little bit different for dedicated/KVM.

[tyaakow]

I believe the issue was with sysctl -p problem that gets raised by default in centos6
Because it returns status code that messes up something in openvz while starting it.
I think.

Because this time after uncommenting those lines in sysctl.conf it's okay.

Anyway, I'm able to start containers now, I installed OpenVZ Web Panel in ruby/ext.js

However, container, tho assigned it's public IP, can't connect (ping google).

How does networking go with this?

If anyone has any advice, throw it here freely

And if there's any other good management console, also.

Thanks for your replies so far

[TheHackBox]

I would make sure you have ipv4 forwarding enabled on the box.

[ShEkhAr]

I think the ipv6 and ipv4 both adress are not compiled fully restart your dedicated and re compiled the adresses i think this reolved the issues.

[f8ll]

I don't know anything about the panel you are using, but in general see http://wiki.openvz.org/Common_Networking_HOWTOs and http://wiki.openvz.org/Virtual_network_device

[Boltersdriveer]

Try doing a ifconfig and see what comes up on the IP's.

[tyaakow]

(03-21-2012 07:11 AM)f8ll Wrote:  Is the ipv6 module loaded or compiled into the kernel?

I resolved the issue, it was standard CentOS 6 issue with sysctl.conf file,
there are 3 lines that need to be commented out because they return error code when sysctl is being run.

---

The issue that needed to be resolved, for posterity:
I commented out these lines in CentOS sysctl.conf file, like this:

#net.bridge.bridge-nf-call-ip6tables = 0
#net.bridge.bridge-nf-call-iptables = 0
#net.bridge.bridge-nf-call-arptables = 0