Running OpenVZ Centos Containers on Proxmox

[shart]

Hi
We've been using OpenVZ on Fedora kernels for a while and they are very stable. We have installed Proxmox and like the GUI so we are testing moving containers to Proxmox.

However, when we move the containers which have been created with Centos appliance templates off of the Fedora based OpenVZ servers, on to the Proxmox server we get boot errors and the VEs start but do not complete the boot sequence.

Can we convert these containers somehow?

Thanks

 

[tom]

Hi
We've been using OpenVZ on Fedora kernels for a while and they are very stable. We have installed Proxmox and like the GUI so we are testing moving containers to Proxmox.

However, when we move the containers which have been created with Centos appliance templates off of the Fedora based OpenVZ servers, on to the Proxmox server we get boot errors and the VEs start but do not complete the boot sequence.

Can we convert these containers somehow?

Thanks

pls see http://pve.proxmox.com/wiki/Migration_of_servers_to_Proxmox_VE#Move_OpenVZ_containers_to_Proxmox_VE

if you have any issues come back with details (log files).

 

[shart]

Hi tom,
Thanks for the quick response. Yep followed all that stuff. I can manage the containers from the GUI quite happily using the pve.auto tip. As an aside I did have an initial issue with the original conf files containing a line DISK_QUOTA=no which Proxmox couldnt parse and so the containers didnt appear on the GUI. It's easily fixed by changing the line to DISK_QUOTA="no".

The issue we have now is more serious I think and to do with the original containers being build using a CentOS Release 4.5 template and running on a Fedora Core release 6 HN. Once we move these to the Proxmox server that's running Debian, the containers dont start properly. Here's a boot log:

Welcome to  CentOS release 4.5 (Final)
Initializing hardware... storage network audio done[ OK ]
Configuring kernel parameters: error: permission denied on key 'kernel.sysrq'
net.ipv4.ip_forward = 0
net.ipv4.conf.default.rp_filter = 1
net.ipv4.conf.default.accept_source_route = 0
error: permission denied on key 'kernel.core_uses_pid'
[FAILED]
hwclock is unable to get I/O port access: the iopl(3) call failed.
Setting clock : Wed Apr 29 12:32:48 BST 2009 [ OK ]
Setting hostname wiki: [ OK ]
Remounting root filesystem in read-write mode: mount: permission denied
[FAILED]
Mounting local filesystems: [ OK ]
/etc/rc.d/rc.sysinit: line 765: /dev/urandom: Permission denied
Enabling swap space: [ OK ]
INIT: Entering runlevel: 3
Entering non-interactive startup
Flushing firewall rules: [ OK ]
Setting chains to policy ACCEPT: mangle filter [ OK ]
Setting network parameters: error: permission denied on key 'kernel.sysrq'
error: permission denied on key 'kernel.core_uses_pid'
net.ipv4.ip_forward = 0
net.ipv4.conf.default.rp_filter = 1
net.ipv4.conf.default.accept_source_route = 0
[FAILED]
Bringing up loopback interface: [ OK ]
Bringing up interface venet0: [ OK ]
Starting system logger: [ OK ]
Starting kernel logger: [PASSED]
Starting MySQL: [ OK ]
Starting sshd:[ OK ]
If you encounter issues starting up Confluence Standalone, please see the Installation guide at http://confluence.atlassian.com/display/DOC/Confluence+Installation+Guide
Using CATALINA_BASE: /opt/confluence-std-2.6.0
Using CATALINA_HOME: /opt/confluence-std-2.6.0
Using CATALINA_TMPDIR: /opt/confluence-std-2.6.0/temp
Using JRE_HOME: /usr/java/jdk1.5.0_08
INIT: no more processes left in this runlevel

 

[tom]

I assume a lot of these warnings are also on your old host - but you do not have such a nice log to see them.

As far as I see the container starts up.

I suggest you compare the config file of your migrated machines with a fresh installed centos 4 downloaded via Proxmox VE gui.
http://pve.proxmox.com/wiki/CentOS_4_Standard

Before you start, change the RAM settings on your migrated containers (ram/swap) and click save. By doing this, the config file of the container is a bit adapted for Proxmox VE.

 

[shart]

I assume a lot of these warnings are also on your old host - but you do not have such a nice log to see them.

As far as I see the container starts up.

I suggest you compare the config file of your migrated machines with a fresh installed centos 4 downloaded via Proxmox VE gui.
http://pve.proxmox.com/wiki/CentOS_4_Standard

Before you start, change the RAM settings on your migrated containers (ram/swap) and click save. By doing this, the config file of the container is a bit adapted for Proxmox VE.

Well the VE I'm using is a wiki (as you can see from the log), the VE seems to start but the wiki java machine inside it does not.

Are you saying that a VE based on any template and from an HN that's running any kernel should work fully when restored to an HN running any other kernel?

Thanks for your help so far, btw