I don’t know about tailscale, but it seems pihole has got you covered with local DNS, if you’re willing to set the local DNS records manually.
I use pihole as selfhosted DNS server for all my servers and clients. I don’t have many local DNS records (only 2), so if you handle a great amount of ever-changing DNS records, this might not be for you.
To me it seems like:
I’d try Proxmox VE and, if you’re also searching for a Backup Server, Proxmox Backup Server.
I recommend these because:
I personally run a Proxmox VE + Proxmox BS setup in 3 companies + my own homelab.
It’s not magic, Proxmox VE is literally Debian 13 + qemu + kvm with a nice webui.
So you know the tech is proven, it’s just now you also get an easy to use interface instead of virsh console commands or virt-manager.
I personally like a stable infrastructure to test and run my important and experimental tuff upon. That’s why I’m going with this instead of managing even the hypervisor myself with Arch.
Sadly, that did not solve my problems either.
There’s still no voice output except for the first sentence of all of the navigation apps and when I test the voice navigation with osmand development plugin.
GLaDOS voice is working though, just not when navigating. So, when you actually need it.
Uh yeah, GLaDOS voice! I’ll try SherpaTTS tomorrow.
I’ll report back tomorrow, thank you very much!
Sadly, I’m back to square one now.
I am able to hear he first sentence and that’s it :(
My only thought is maybe your tts software?
Thank you for your input! I, too, suspect the culprit being RHVoice. I have no proof, though. Yet.
While digging through GrapheneOS’s system settings, I found something which may be of interest. Text-to-speech settings were set to use RHVoice and use system settings for the language. There are 2 sliders to alter speech speed and pitch and there’s a button to test the new settings. Neither could I use the sliders, nor the button, all of them were greyed out.
After setting the spoken language to english (united states), the sliders and button became usable. I did start a test-navigation with Osmand and it now also tells me the first thing I need to do on my drive (turn left in 400m), which it didn’t do before.
Not sure how that setting changes anything of relevance though. RHVoice has no german voice (my system language), but all the navigation apps mentioned above come with their own voices. I’ll test this out tomorrow on my way to work.
Will report back. Thanks again!
Of course you can always build a good PC or server.
I could have done that too, but I wanted my first real homelab-do-it-all-yourself setup to be a little more on the cautiously small side. I didn’t want to have too much noise in my apartment and also didn’t want to stress my electricity-bill and wallet too much, so I opted to build small and reuse what I had lying around.
I already had 2 Mini-PCs and a raspberry pi from earlier experiments with selfhosting. I just bought some disks and RAM. If you don’t have any mini-PCs, they’re relatively cheap in comparison with full PCs. Or you could use some older PC you still have but do not use.
My motto more or less was you can always spend more money and build bigger later
The final Hardware
What I host on my Proxmox VE
The 2nd Mini PC (some old intel NUC with 4 cores and 16 GB RAM) + a USB HDD is my Proxmox Backup Server for all this. And what’s really important (my data from nextcloud + some configs) gets backed up to my Hetzner Storage Box with restic.
The raspberry pi is now my WiFi Access Point :)
Conclusion
Homelab doesn’t need to be big or small, it can be whatever you want it to be or whatever you can afford or are willing to have and maintain. From my experience, if you’re not hosting anything CPU-intensive, older or smaller machines will do just fine.
For example, my nextcloud could easily use more resources than the whole Zotac ZBox could house, if there were more users. But as my services are only used by me, most of them are idle most of the time.
Tip at the end about your opsense-VM on Proxmox
I tried letting Proxmox host my pfSense too, but that got old pretty fast. Whenever Proxmox needed a Reboot, my internet was gone too for that time, as the pfSense VM on Proxmox was the gateway to my ISP-modem. In the end, I just bought a real Netgate pfSense appliance.
I did not run OPNSense, but I have a direct comparison for pfSense as VM on Proxmox VE vs pfSense on a ~400€ official pfSense physical appliance.
I do not feel any internet-speed or LAN-speed differences in the 2 setups, I did not measure it though. The change VM -> physical appliance was not planned.
Running a VM-firewall just got tiring fast, as I realized that Proxmox VE needs a lot more reboot-updates than pfsense does. And every time you reboot your pfSense-VM-Hypervisor, your internet’s gone for a short time. Yes, you’re not forced to reboot. I like to do it anyway, if it’s been advised by the people creating the software I use.
Though I gotta say, the pfSense webinterface is actually really snappy and fast when running on an x86 VM. Now that I have a Netgate 2100 physical pfSense appliance, the webinterface takes a looooong time to respond in comparison.
I guess the most important thing is to test it for yourself and to always keep an easy migration-path open, like exporting firewall-settings to a file so you can migrate easily, if the need arises.
[EDIT] - Like others, I also would advice heavily against using the the same hypervisor for your firewall and other VMs. Bare-Metal is the most “uncomplicated” in terms of extra workload just to have your firewall up and running, but if you want to virtualize your firewall, put that VM on its own hypervisor.