Cluster 2 Proxmox VE 2.2 Servers

[TheReelaatiiv]

Hello,

i have two servers, which are directly connected with 1Gbit. So i decided to build a proxmox ve 2.2 cluster and already use drbd for another service but it seems like proxmox isn't compatible with drbd. Is there another way to cluster (replicate) an entire hard disk (/dev/sdb) over two servers?

 

[udo]

Hi,
why is pve not compatible with drbd?? I use on different pve-server drbd with good experiences. Look at the wiki!

Udo

 

[TheReelaatiiv]

Hi,

thank you for your answer. Do you mean this? http://pve.proxmox.com/wiki/DRBD What's about the Cluster Filesystem? On my other service i use ocfs2 but that seems not to be compatible with proxmox and so i thought drbd isn't brilliant to use. I also read on a blogpost about drbd's instability.

Whats about this config:

become-primary-on both;

Does this mean, i can mount the drbd filesystem on BOTH nodes at the same time and can access it also the same time?

 

[udo]

Hi,

thank you for your answer. Do you mean this? http://pve.proxmox.com/wiki/DRBD What's about the Cluster Filesystem? On my other service i use ocfs2 but that seems not to be compatible with proxmox and so i thought drbd isn't brilliant to use. I also read on a blogpost about drbd's instability.

Whats about this config:

become-primary-on both;

Does this mean, i can mount the drbd filesystem on BOTH nodes at the same time and can access it also the same time?

Hi,
it's depends on your usage. DRBD as lvm-storage is very usefull (you don't mount partitions/logical volumes) - but only usable for kvm.
I don't have experiences as shared storage for OpvenVZ (of course you need an cluster-fs for that)...

Most of my VMs are kvm - so DRDB fit very good. I have DRBD for years in production and must say it's very solid (only problems with dolphin-nic drivers - sometimes split brain - but after change to 10GB-Ethernet no one split brain condition occur).

Udo