I self host a lot, including some business related infrastructure for my home office (some employees also work from here).
However services designed to be accessed by a wider public audience (eg; websites, emails, Nextcloud) are hosted on a rented dedi.
As other comments have pointed out, kids suck up every minute of your life, and when the internet or Plex don't work I'll know about it real quick (DAAAAD!).
The important question for me is; how fast can I rebuild all of this if all the drives were wiped?
* Rundeck container for various infrastructure based automation jobs
* Zabbix + Grafana containers for monitoring
* PacketFence VM for NAC
* Shinobi container for NVR
* Snapcast container for multi-room audio
I also have plenty of containers and VM's for various testing apps, or dev projects.
All of my non-critical containers (eg; Plex) self update daily at early morning hours.
For the critical stuff, I have tried to automate updates as much as possible.
The majority of things can be recreated from ansible, docker compose, or Nomad scripts, all of which are backed up to an offsite Nextcloud instance.
I use a lot of services on Opnsense, but I think one of the most important for me is the Traffic Shaper, allowing for bandwidth control.
I have about 14 VLAN's, and am in the process of setting up VXLAN's for further isolation.
I use restic for encrypted backups, stored in Wasabisys.com (no cost to download, unlike Backblaze).
I will admit there is still one design flaw that I'm yet to spend time overcoming; if everything is powered on at the same time, there is a chance some devices won't get an IP because OpnSense is not yet ready.
My current flawless work-around is to boot Opnsense 30 seconds before everything else.
As other comments have pointed out, kids suck up every minute of your life, and when the internet or Plex don't work I'll know about it real quick (DAAAAD!). The important question for me is; how fast can I rebuild all of this if all the drives were wiped?
Here is my home setup;
Hardware:
* OPNSense - old i5 desktop, 70GB hdd, 20GB RAM
* TrueNAS - AMD A10-5800K, 16TB hdd, 1TB ssd, 8GB RAM
* OpenNebula Frontend - Random Intel NUC, 128GB ssd, 4GB RAM
* OpenNebula KVM Node - Xeon ?, 4TB hdd, 64GB RAM
* OpenNebula KVM Node - Xeon Silver ?, 2TB ssd, 128GB RAM + GTX 1070Ti
* Unifi EdgeSwitch ES-48
* Unifi WAP's
Software:
* OPNSense for routing
* OpenNebula for VM's
* Nomad for docker containers
* NetMaker for Wireguard based VPN
Main Apps:
* Home Assistant VM
* Plex (+rtorrent+sonarr+radarr+jackett+ombi) containers
* FusionPBX container for VoIP
* Kasm for remote desktops
* Rundeck container for various infrastructure based automation jobs
* Zabbix + Grafana containers for monitoring
* PacketFence VM for NAC
* Shinobi container for NVR
* Snapcast container for multi-room audio
I also have plenty of containers and VM's for various testing apps, or dev projects.
All of my non-critical containers (eg; Plex) self update daily at early morning hours.
For the critical stuff, I have tried to automate updates as much as possible.
The majority of things can be recreated from ansible, docker compose, or Nomad scripts, all of which are backed up to an offsite Nextcloud instance.
I use a lot of services on Opnsense, but I think one of the most important for me is the Traffic Shaper, allowing for bandwidth control.
I have about 14 VLAN's, and am in the process of setting up VXLAN's for further isolation.
I use restic for encrypted backups, stored in Wasabisys.com (no cost to download, unlike Backblaze).
I will admit there is still one design flaw that I'm yet to spend time overcoming; if everything is powered on at the same time, there is a chance some devices won't get an IP because OpnSense is not yet ready. My current flawless work-around is to boot Opnsense 30 seconds before everything else.