Home Server Setup - Application data sync of ZFS pool binded data
Current situation:
- On server with a ZFS storage pool that several docker container use as a persistant storage (each in another folder) from an LXC container
=> works, but is not HA
Intention: Have a HA cluster with 3 nodes where 2 nodes are used for redundency (server/storage), the 3 node is not holding VMs or storage, just used for quorum
Setup:
- 3 node cluster
- group created to just use node 1/2
- zfspools (same name) created for node 1/2 to keep in sync (CEPH overkill for home cluster)
- For the new HA setup of the docker containers I would need to spin-up kubernetes (prefered) or separate them in single LCXs
Question:
- Can I achieve (neartime) HA by using storage replication on the ZFS pools? Or do I need to change my setup?
- I was confused by having to select the VM, that makes sense for the VM-File, but what about the application data in a folders on the ZFS pool?
- Wouldn't make sense to configure it for every VM, as they all use the same pool (just different folders) on the ZFS pool
- THAT IS MY MAIN ISSUE OF UNDERSTANDING HOW TO SETUP REPLICATION FOR THE APPLICATION DATA -> How to configure it for replication?
- Extra question: if I distribute the VMs in normal operation, can the ZFS pools be syncronised from both nodes or is this done anyway in both directions?
Please, also tell me, if I am completely wrong with my setup (home cluster!) as I am new to that topic.
Current situation:
- On server with a ZFS storage pool that several docker container use as a persistant storage (each in another folder) from an LXC container
=> works, but is not HA
Intention: Have a HA cluster with 3 nodes where 2 nodes are used for redundency (server/storage), the 3 node is not holding VMs or storage, just used for quorum
Setup:
- 3 node cluster
- group created to just use node 1/2
- zfspools (same name) created for node 1/2 to keep in sync (CEPH overkill for home cluster)
- For the new HA setup of the docker containers I would need to spin-up kubernetes (prefered) or separate them in single LCXs
Question:
- Can I achieve (neartime) HA by using storage replication on the ZFS pools? Or do I need to change my setup?
- I was confused by having to select the VM, that makes sense for the VM-File, but what about the application data in a folders on the ZFS pool?
- Wouldn't make sense to configure it for every VM, as they all use the same pool (just different folders) on the ZFS pool
- THAT IS MY MAIN ISSUE OF UNDERSTANDING HOW TO SETUP REPLICATION FOR THE APPLICATION DATA -> How to configure it for replication?
- Extra question: if I distribute the VMs in normal operation, can the ZFS pools be syncronised from both nodes or is this done anyway in both directions?
Please, also tell me, if I am completely wrong with my setup (home cluster!) as I am new to that topic.