Hi all,
This is my first post so hello to the community and congratulation for this wonderful software.
I use proxmox for while (1.8 version in production in a SME) and I'm testing the 2.0 beta3 for HA feature.
What I have done so far is :
- installing / configuring Proxmox 2.0 beta3 in cluster mode on two physical machine
- install DRBD on a second hard drive on both machine and configure one VM (KVM) on a lvm storage on top of DRBD.
Migration from one to other are working (sometime needs a reboot although the "live" migration is used but I will dig into that later.
Now I want to test the following scenario:
I'm running my VM on node 1 of my cluster
I switch off the node 1 (simulation of a fault on this machine)
Assuming I have a problem on node 1 I imagined I could restart the VM on node 2 based on the up to date data of DRBD storage (Is that hypothesis correct ??)
To test this I simply start VM on node 1 and disconnect cable between node 1 and 2.
DRBD detected a spli brain and I 'm testing in Standalone on node 2
Since then acessing to the web interface of Proxmox on node 2 display my VM still on node 1 and migration on node 2 is impossible (pop up open but no target server available)
Did I miss something or am I using the DRBD feature the wrong way ?
Thank you for answers in advance and congratulations again to the development team of this very nice software.
Snarf
This is my first post so hello to the community and congratulation for this wonderful software.
I use proxmox for while (1.8 version in production in a SME) and I'm testing the 2.0 beta3 for HA feature.
What I have done so far is :
- installing / configuring Proxmox 2.0 beta3 in cluster mode on two physical machine
- install DRBD on a second hard drive on both machine and configure one VM (KVM) on a lvm storage on top of DRBD.
Migration from one to other are working (sometime needs a reboot although the "live" migration is used but I will dig into that later.
Now I want to test the following scenario:
I'm running my VM on node 1 of my cluster
I switch off the node 1 (simulation of a fault on this machine)
Assuming I have a problem on node 1 I imagined I could restart the VM on node 2 based on the up to date data of DRBD storage (Is that hypothesis correct ??)
To test this I simply start VM on node 1 and disconnect cable between node 1 and 2.
DRBD detected a spli brain and I 'm testing in Standalone on node 2
Since then acessing to the web interface of Proxmox on node 2 display my VM still on node 1 and migration on node 2 is impossible (pop up open but no target server available)
Did I miss something or am I using the DRBD feature the wrong way ?
Thank you for answers in advance and congratulations again to the development team of this very nice software.
Snarf