I'm trying to understand how netplan is working exactly and I get stuck at the following technical design concept:
From netplan introduction:
During early boot, the netplan “network renderer” runs which reads
/{lib,etc,run}/netplan/*.yaml and writes configuration to /run to hand
off control of devices to the specified networking daemon.
This is because of the netplan key goal:
No persistent generated config, only original YAML config
I do not understand, why netplan does follow this design concept. What is the reason and/or advantage of a non persistent back-end configuration, that is generated every time, the system does boot?
And in addition to that: What do I need to take care about, if I would work on a network configuration by editing the back-end configuration using one of the well known methods?
A reason for a mix of using netplan along with a back-end configuration tool could be e.g., that most of the network configruation is covered by using netplan, but for a specific part there would be the need to edit the configuration without netplan, because netplan does not support the needed capability (yet).
Would be the only way to resolve such a situation by leaving netplan alone and start just using pure networkd, even though one of a couple of connections/interfaces would be affected only?
This question is about understanding netplan only! I do not have any technical issues with netplan currently!
Thank you!