IP-handling

[TErxleben]

How about a possibility to supply PVE/PBS permanently with an IP via DHCP?
The idea behind is to operate a backup device in another remote network and, in the event of an error, to simply physically move and plug this external server into the target network in order to get rid of a lame Internet connection in case of restauration?

 

[Moayad]

Hi,

The DHCP in a server It's not a common practice. In PVE/PBS you can change the IP easily, by editing the `/etc/network/interfaces` and then `ifreload -a`

 

[TErxleben]

We're not in the 80's anymore. With PVE I can see problems with DHCP (bridge settings etc.). I don't see such problems with a baremetal-PBS. In my opinion, it would be extremely helpful to operate a baremetal PBS in another remote network as a sync-partner and be able to carry and use the same in the main network in order to be able to start the restoration without annoying fiddling with CLI network settings.
Of course it's just a suggestion.

 

[pveuser113]

We're not in the 80's anymore.

we are not. Yet your idea of disaster recovery is to throw your precious backup server on a truck/train/ship/plane/all of the above, and the only thing that seems to complicate this plan - a text file update to change the IP?

In addition to faking IP, you are going to need to fake the SSL/SSH authentication that PBS/PVE uses to communication. Its infinitely easier to just add a new PBS location that uses new static IP, if your server made the journey in one piece.

 

[TErxleben]

The PBS servers (4 bay mini housing) are located at different locations connected via VPN. Travel time maybe 1 hour. Copy time of 2TB at 100mbit/s 45h.
Since it's all about the restore, I thought SSL/SSH doesn't matter. I am not sure about that. I wanted to bypass the manual IP change of the corresponding machine, e.g. to avoid typos. Of course it's whining at a relatively high level.

 

[TErxleben]

OK, new idea.
Due to the correct objection by @pveuser113 it is a danger to transport the complete server:

- export the last backup of the backed up VMs on a PBS as dumps to an external USB disk every day.
- instead the complete server transport this medium to where there was an accident.
- connect this medium to the target PVE machine and restore the expanded transported VMs there.
- I prefer the whole thing to be as automatic as possible.

Idiotic idea, wet dream on my part or worth considering?

It's mostly about having remote locations fast available with lame internet connections.