proxmox

Setting up Proxmox CLUSTER and STORAGE (Local, ZFS, NFS, CEPH) |

Setting up Proxmox CLUSTER and STORAGE (Local, ZFS, NFS, CEPH) | Proxmox Home Server Series

#Setting #Proxmox #CLUSTER #STORAGE #Local #ZFS #NFS #CEPH

“MRP”

In this video i will be setting up Proxmox Cluster of 3 Nodes and will go over each storage option you have [ Local, ZFS, NFS, CEPH ]
I hope this video will be helpful for you to decide if you want to “cluster” your Proxmox instances.

PVE Cluster…

source

 

To see the full content, share this page by clicking one of the buttons below

Related Articles

7 Comments

  1. Superb video, thanks. You make this look really easy, but I know from bitter experience, its not at all easy. I have exactly the same setup you have, but would like to add further USB SSD disk storage (1TB Drives) to two of my nodes. My guess is this can be setup as a separate ceph pool? I think there is little point adding it to the primary pool – replication 5 times seems overkill. If I cant, I may just try and do a separate zfs share for Proxmox backups
    One word of warning – I 'experimented' with pimox. I added one pimox node to the cluster, but ceph just didnt want to work properly. It automatically added my pooled storage to the new pimox node. I decided I didnt want that node, so tried to remove it. Its impossible. It ended up with a complete destruction of my entire cluster, and trashed my pooled data. I have backups luckily so suggest backing up too.
    One more thing. My nodes are mini pcs. Each has two 1GB NICs. I would somehow like to utilise both on each node – One for CEPH sync, the other for general management (assuming this will improve the bottlenecks) Do you have experience with this?

  2. The difference between and Ceph and NFS performance really needs to be considered in not only migration time but also normal read/write speeds when operating the container. Because the storage in ceph is local to the machine you should have MUCH more IO than you would going over the network to your NFS machine. NFS to me is not a good answer when talking about most use cases of HA.

  3. Awesome video, thanks a lot. I’ve been putting off redoing my 3 node cluster, because I didn’t have my head wrapped around Ceph yet, but I think I’m ready to try it out on a test cluster. Wish me luck. 😂

Leave a Reply