Cluster HA and Openvz Live migrate

[yena]

Hi,
i have a 3 node stage cluster, last proxmox version.
Two nodes manage Virtual Machines with a shared DRBD volume primary/primary.
KVM Live migration and HA works well.
If i try to create on my DRBD volume a new logical volume , ext4 formatted,
and on it i create one or more Openvz VM,
this Openvz conteiners can't migrate.

when i try to migrate i see the migrated conteiner on the second node (proxmox-2) but i can't start it.
if i do an ls -la on the proxmox that now manage the vps i see an empty private dir.

Oct 14 18:51:28 rgmanager Reconfiguring
Oct 14 18:51:28 rgmanager Loading Service Data
Oct 14 18:51:30 rgmanager Stopping changed resources.
Oct 14 18:51:30 rgmanager Restarting changed resources.
Oct 14 18:51:30 rgmanager Starting changed resources.
Oct 14 18:56:24 rgmanager Reconfiguring
Oct 14 18:56:24 rgmanager Loading Service Data
Oct 14 18:56:26 rgmanager Stopping changed resources.
Oct 14 18:56:26 rgmanager Restarting changed resources.
Oct 14 18:56:26 rgmanager Starting changed resources.
Oct 14 18:56:26 rgmanager Initializing pvevm:102
Oct 14 18:56:26 rgmanager pvevm:102 was added to the config, but I am not initializing it.
Oct 14 18:56:30 rgmanager Ignoring M_CLOSE for destroyed context 17
Oct 14 18:56:31 rgmanager Starting stopped service pvevm:102
Oct 14 18:56:32 rgmanager [pvevm] CT 102 is already running
Oct 14 18:56:32 rgmanager Service pvevm:102 started
Oct 14 18:57:05 rgmanager [pvevm] CT 102 is running
Oct 14 18:57:35 rgmanager [pvevm] CT 102 is running


Non active proxmox:
root@proxmox-1:/var/log/cluster# ls -la /var/lib/vz/private_drbd0/private/105/
total 88
drwxr-xr-x 22 root root 4096 Oct 14 19:25 .
drwxr-xr-x 3 root root 4096 Oct 14 19:25 ..
drwxr-xr-x 2 root root 4096 Jul 22 07:03 bin
drwxr-xr-x 2 root root 4096 Jun 2 22:29 boot
drwxr-xr-x 3 root root 4096 Oct 14 19:25 dev

Active proxmox:
root@proxmox-2:/var/lib/vz/private_drbd0# ls -la /var/lib/vz/private_drbd0/private
total 8
drwxr-xr-x 2 root root 4096 Oct 14 19:24 .
drwxr-xr-x 5 root root 4096 Oct 14 19:24 ..


On the abondoned node ( proxmox-1 ) i still see mounted dir:

dev/mapper/VG_DRBD0-openvz1 9.9G 593M 8.8G 7% /var/lib/vz/private_drbd0
/var/lib/vz/private_drbd0/private/105 4.0G 442M 3.6G 11% /var/lib/vz/root/105
tmpfs 52M 32K 52M 1% /var/lib/vz/root/105/run
tmpfs 5.0M 0 5.0M 0% /var/lib/vz/root/105/run/lock
tmpfs 205M 0 205M 0% /var/lib/vz/root/105/run/shm


Have i to use NFS ? ( mounted on my /var/lib/vz/private_drbd0 ) ?

Whats wrong ?

Thanks!

-- Yena --

 

[mir]

Did you remember to shotdown ct-102 before adding it to HA?
You cannot add a VM or CT to HA if it is running when added, CT and VM must be stopped first since rgmanager will start the CT or VM when you apply the HA changes. If the server is running at the time you will see errors like:
Oct 14 18:56:26 rgmanager Initializing pvevm:102
Oct 14 18:56:26 rgmanager pvevm:102 was added to the config, but I am not initializing it.

Above simply means that adding the CT 102 to HA failed since another process is owning it (another was responsible for starting the instance) in which case rgmanager cannot take ownership and therefore not add the instance to HA.