No, I want the whole server, so I can restore this backup onto a naked PC.If you only ask about /etc/pve
No, I want the whole server, so I can restore this backup onto a naked PC.
But I see, that here is no official way to get a backup and I have to handle it on my own. Sad.
I have eg. also installed nut to let the server shut itself off, when the UPS is running from battery, I installed alsa for audio to get the speaker and an attached USB speaker of the server to work in my Lyrion server container.that only holds configurations and ISO install
I have eg. also installed nut to let the server shut itself off, when the UPS is running from battery, I installed alsa for audio to get the speaker and an attached USB speaker of the server to work in my Lyrion server container.
So there are many thing that have to be restored and I don't know where all these things are.
There cannot be a one-shoe-fits-all-approach, because you can change whatever you want in the underlying Debian and the backup tool should be able to include everything. That's not going to happen.But I see, that here is no official way to get a backup and I have to handle it on my own. Sad.
You can however use e.g. Rear for backup and bare-metal-restore, which works (to the ordinary limits of live backup) well with PVE if you're not using ZFS. It you're using ZFS, just replicate your pool.
No, I would not backup cluster node (no point, as you already pointed out), this was for a single node.Did you test this e.g. multiple times during heavy corosync traffic?
I suspect this too, therefore the ordinary limits of live backup.I suspect you might back up the config.db middle of SQLite checkpointing. I would definitely want to .dump it before actual backup run.
There cannot be a one-shoe-fits-all-approach, because you can change whatever you want in the underlying Debian and the backup tool should be able to include everything. That's not going to happen.