Hi,
I would like to share my troublesome experience with you.
I setup two Proxmox VE 3.3-5/bfebec03 installed on Debian Wheezy. They are connected with a Gigabit LAN via a crossover cable and there is a drbd blockdevice where qemu disks are located, with HA enabled, failover domain properly configured, a quorum disk on a third server and fencing setup. Up to here, everything works great: I can live migrate, rgmanager works as expected.
Now, in order to play with some HA openvz container, I tried different setups with no luck.
I'm using a basic container, derived from ubuntu-12.04-standard_12.04-1_i386.tar.gz template.
I tried with a glusterfs, set on the same two machines, via the same LAN, backupping and restoring containers on it, but I can't properly start the containers from there and I get weird errors when trying to migrate them between my two nodes.
I tried to mount the glusterfs as NFS, with the same result: containers seem to hang at startup, with no external errors. Stopping them takes more time than the usual, they appear stuck.
What could cause this behavior? What could be the best configuration to have the containers on a shared stored, able to live migrate from one to ther node, even in case of failure?
Thank you in advance.
I would like to share my troublesome experience with you.
I setup two Proxmox VE 3.3-5/bfebec03 installed on Debian Wheezy. They are connected with a Gigabit LAN via a crossover cable and there is a drbd blockdevice where qemu disks are located, with HA enabled, failover domain properly configured, a quorum disk on a third server and fencing setup. Up to here, everything works great: I can live migrate, rgmanager works as expected.
Now, in order to play with some HA openvz container, I tried different setups with no luck.
I'm using a basic container, derived from ubuntu-12.04-standard_12.04-1_i386.tar.gz template.
I tried with a glusterfs, set on the same two machines, via the same LAN, backupping and restoring containers on it, but I can't properly start the containers from there and I get weird errors when trying to migrate them between my two nodes.
I tried to mount the glusterfs as NFS, with the same result: containers seem to hang at startup, with no external errors. Stopping them takes more time than the usual, they appear stuck.
What could cause this behavior? What could be the best configuration to have the containers on a shared stored, able to live migrate from one to ther node, even in case of failure?
Thank you in advance.
Last edited: