spare server in cluster only part of time?

[rolfb]

Hi all,

I'm quite new to proxmox. Use it as platform for home-use: mailserver, NAS, web, media-servers run on a mix of VM's and containers on it.
My old server is now decommissioned, but technically OK. It houses 1 spinning disk of abt 2TB
So I was thinking to install proxmox on it too, and let it function as a backup/spare proxmox. But since it is based in my living, it's not my intention to have it running 24*7. (noise, electricity-costs, ...)

My idea:
install proxmox on it, config it as node in a cluster, and power it down. Then, during the night, wake it up with WOL, sync backups to its disk, and power down again.
When something happens to my main server, I can (temporarily) use the old server to spin up the servers from the synced backups on it (with some data loss between last backup and failure, that's fine for my purpose).

My question:
will this work? Will the freshly booted spare take its place in the cluster automatically? Do I need to config shared storage for this? Or should I script some scp to copy the backup files to the spare?
Any tips? Do's/don'ts, best practices?

 

[LnxBil]

As with any cluster, you have to have quorum to change stuff, so a two-node cluster does not work out-of-the-box. Maybe to independet nodes with pve-zsync may be the best solution for your problem?

 

[rolfb]

As with any cluster, you have to have quorum to change stuff, so a two-node cluster does not work out-of-the-box. Maybe to independet nodes with pve-zsync may be the best solution for your problem?

It took a while, but thanks for your reply. Unfortunately: I'm not using ZFS, so zsync is not an option.
Furthermore: quorum is necessary for a cluster with HA and automatic failover AFAIK. But that is not what I'm aiming at. I just want a spare machine to contain my backup VM's and containers, AND is is able to spin them up in case of hardware failure of main server.

What I'm looking at now:
My new server has 1TB NVME, and 250 GB SSD.
Big disk is LVM-thin, containing VM and containers.
250 GB is containing PVE, and backups of VM's and containers (and some iso etc).

Spare hardware has 2TB spinning disk, now divided (automatically during install) into
100 GB local for backups, iso, etc
1.9 TB LVM-thin for VM, containers.

If I scp the backups from pve-new to pve-old in the night, will I be able to restore the backups on pve-old? If so, I can script wake-on-lan, scp, and shutdown, so old server has freshly made backups every night?

Or should I create a cluster of these 2 nodes (not HA, not auto-failover), power on pve-old, create/move backups to the other "spare" leg of the cluster (how?) and power it down again?

Any thoughts?

 

[LnxBil]

Furthermore: quorum is necessary for a cluster with HA and automatic failover AFAIK.

No you need it to administer a two-node-cluster if one node is down. HA has nothing to do with that.

Or should I create a cluster of these 2 nodes (not HA, not auto-failover), power on pve-old, create/move backups to the other "spare" leg of the cluster (how?) and power it down again?

That is no real technical or organisatorical improvement over your scp solution.