I have been self hosting on and off since about 2000. When I got my fiber link (4 or so years ago) I also started to do home automation and the self hosting got more serious ("national security infrastructure" as my wife calls it, her to-go question is "what will happen with that shit if you die tomorrow?").
I run everything in docker containers, except Wireguard and the backup.
I had an idea to "document everything" but the realized there would be no-one to maintain the services (neither my wife not my children are interested in IT). I then decided to document how to untangle themselves from the self-hosting (moving to well-established services, dismantle the home automation, ...). Thi sis going to make more sense in the long term.
Since I do not plan to die tomorrow but I do not know for my server, I tried to do a disaster recovery from scratch. It was surprisingly easy - after starting a VM on my laptop (simulating the bare-metal rebuild of my server), I was up and running in about an hour. Docker is simply fantastic for that.
I took some notes about the disaster recovery, which I promptly lost so I will have to run this again soon (maybe next week). Actually I will copy the encryption key for my remote backup right away as this is the only thing I won't be able to recover. EDIT: done :)
TL;DR:
- use docker containers for your services, it is a life-saver and makes everything easier
- use a docker-compose script to have everything on file
- document either your maintenance or exit strategy
I run everything in docker containers, except Wireguard and the backup.
I had an idea to "document everything" but the realized there would be no-one to maintain the services (neither my wife not my children are interested in IT). I then decided to document how to untangle themselves from the self-hosting (moving to well-established services, dismantle the home automation, ...). Thi sis going to make more sense in the long term.
Since I do not plan to die tomorrow but I do not know for my server, I tried to do a disaster recovery from scratch. It was surprisingly easy - after starting a VM on my laptop (simulating the bare-metal rebuild of my server), I was up and running in about an hour. Docker is simply fantastic for that.
I took some notes about the disaster recovery, which I promptly lost so I will have to run this again soon (maybe next week). Actually I will copy the encryption key for my remote backup right away as this is the only thing I won't be able to recover. EDIT: done :)
TL;DR:
- use docker containers for your services, it is a life-saver and makes everything easier
- use a docker-compose script to have everything on file
- document either your maintenance or exit strategy
- do a quick DRP test on a VM