[SOLVED] Continuous replication using Ceph

[cloudguy]

Hello

I’m familiar with the storage replication capability for ZFS data stores. Is there a (supported) way to replicate disks stored on a ceph data store between PVE clusters - ie production and recovery scenario?

 

[aaron]

Have you seen this? https://pve.proxmox.com/wiki/Ceph_RBD_Mirroring

 

[cloudguy]

I have not. Thanks for sharing the link.

It would be a nice to have this integrated into PBS / gui. PBS would become a backup and replication product.

Perhaps something that could be included on the roadmap.

 

[aaron]

PBS got nothing to do with that, or am I missing something? Another approach would be to use two PBS instances in each site and remote sync the backups. Then you can (live) restore in the disaster case. Or do regular restores in the DR site.

Setting up replication between PVE clusters is something we will most likely address in the future with the upcoming Datacenter Manager.

 

[cloudguy]

Just noticed your reply. I think PBS should have this part of its data protection scope. Especially when you’re developing a Data Center Manager.

The use case is a VM that is replicated to another cluster via ceph RBD-mirror. DCM would help with failover scenarios such that the a VM running on the primary cluster would be stopped, replication stopped or direction reversed (from primary - secondary to secondary - primary) and Start VM at secondary cluster.