RBD Cache to offset consumer NVME latency for an uptime prioritized cluster (data consistency lower priority)

gogito

Member
Jan 12, 2022
11
1
8
25
Hi everyone, so I have a proxmox cluster with zfs replication on consumer NVMEthat I'm planning to change into Ceph.

The cluster host multiple VMs that require high uptime so users can log in and do their work, the user data is on an NFS (also on VM). The data is backup periodically and I am ok if needed to restore from the previous backup.


I understand that consumer NVME lack PLP so I will have terrible performance if I run Ceph on them and put my VMs on top. However my plan is to have a cache layer on top so all data read write will go to the local cache and then flush to Ceph later. This cache can be ssd or more preferably, ram.


I see that we have Ceph RBD cache on client side which seems to be doing this. Is that right? Can I expect fast data read/write with the redundancy/ease of migration/data access from multiple server with Ceph?


As title, I don't mind if I lose some data if hosts are down before data from cache is flushed to Ceph, that would be worst case scenario and is still acceptable. For daily usage, I expect it to be as fast (or almost) as local storage due to the cache but when a host is down/shutdown, I can still migrate/start VM on another nodes and at worst only lose the data not flushed to Ceph from the cache.

Is this doable?
 

About

The Proxmox community has been around for many years and offers help and support for Proxmox VE, Proxmox Backup Server, and Proxmox Mail Gateway.
We think our community is one of the best thanks to people like you!

Get your subscription!

The Proxmox team works very hard to make sure you are running the best software and getting stable updates and security enhancements, as well as quick enterprise support. Tens of thousands of happy customers have a Proxmox subscription. Get yours easily in our online shop.

Buy now!