[SOLVED] Debian container unable to start after migration from OpenVZ to LXC on Proxmox 4

[brickmasterj]

After migrating a Debian 7 OpenVZ CT from Proxmox 3 to a LXC on Proxmox 4 it does not appear to start correctly. Fristly, migration was done by stopping the container on Proxmox 3, backing up to external harddisk, then loading that harddisk in Proxmox 4, restoring that backup and setting the network config. After that, the container appears to start correctly according to Proxmox but is a not able to be connected to through either console (tty), gives a blank output, no login or whatsoever, or SSH (connection refused). Restarting both the CT and host does not seem to have any effect, and starting the CT with the `lxc-start -n 101` with a `-o` log file gives the following output:

lxc-start 1445190074.912 ERROR    lxc_commands - commands.c:lxc_cmd_rsp_send:237 - failed to send command response -1 Broken pipelxc-start 1445190074.913 ERRORlxc_commands - commands.c:lxc_cmd_rsp_send:237 - failed to send command response -1 Broken pipe

Any ideas on how to diagnose where this problem may be coming from or how to go about fixing it? EDIT: All other containers migrated the exact same way worked perfectly out of the box...

 

[dietmar]

what is the output of

# pveversion -v

 

[brickmasterj]

what is the output of# pveversion -v

proxmox-ve: 4.0-16 (running kernel: 4.2.2-1-pve)pve-manager: 4.0-48 (running version: 4.0-48/0d8559d0)pve-kernel-4.2.2-1-pve: 4.2.2-16lvm2: 2.02.116-pve1corosync-pve: 2.3.5-1libqb0: 0.17.2-1pve-cluster: 4.0-22qemu-server: 4.0-30pve-firmware: 1.1-7libpve-common-perl: 4.0-29libpve-access-control: 4.0-9libpve-storage-perl: 4.0-25pve-libspice-server1: 0.12.5-1vncterm: 1.2-1pve-qemu-kvm: 2.4-9pve-container: 1.0-6pve-firewall: 2.0-12pve-ha-manager: 1.0-9ksm-control-daemon: 1.2-1glusterfs-client: 3.5.2-2+deb8u1lxc-pve: 1.1.3-1lxcfs: 0.9-pve2cgmanager: 0.37-pve2criu: 1.6.0-1zfsutils: 0.6.5-pve4~jessie

 

[brickmasterj]

Looking at the boot log (dmesg) of the LXC, I seem to be getting some weird errors similar to this:

[ 9741.362095] audit: type=1400 audit(1445199000.045:102): apparmor="DENIED" operation="mount" info="failed flags match" error=-13 profile="lxc-container-default" name="/" pid=10171 comm="mount" flags="rw, remount"

For a more detailed version of the bootlog, see http://pastebin.com/2H0iAJPn Does this mean it might have something to do with permissions?

 

[DJSnoopy]

Hi all.

I´m fighting with the same behavior during my migration tests @home.
My consoles starts working after 2-5 minutes of waiting. Same with the network interfaces.
It seems that there is a timeout or something until the the network starts running for backup migrated machines.

Just for test purposes i installed a blank new debian CT and this console starts directly, also network interfaces.

kind regards
Michael

 

[brickmasterj]

*bump*This is still very much an issue

 

[sigxcpu]

Are you sure that is the boot log of the container? It looks like host's one.

What is the output of

lxc-start -n <container id> -F -l debug -o /dev/stdout

 

[brickmasterj]

Are you sure that is the boot log of the container? It looks like host's one.What is the output of

lxc-start -n  -F -l debug -o /dev/stdout

Thank you so much, this enabled me to diagnose the problem. Turns out, it was to do with something other than LXC itself. The booting container got stuck on apache2 asking for a password for a certificate... Not exactly sure how this caused LXC to crash while in OpenVZ this was never an issue, but I managed to fix it in apache2 itself.