So, i decited to upgrade to 25.04.0, and it dident go as planned.
My “Virtulization” is now “Instances”, and i tried to figure that out.
I coudent get it working, until i stopped my pihole, and then i tried importing the drives, but they gave an error.
But no pihole, no internet. So.
Question:
How to fix the v. 25 DNS Pihole thing, any input here?
I never figured out how to change IP’s for my pihole app, so I have been running like this for some time.
And what does the VM service need port 53 for?
I have reverted to 24.x and have things working there.
But i will need to upgrade at some point in time.
Yes i found out about the “Experimental” part, but when my pihole stopped working, i decided to go back.
It did bind to my IP, that is what the whole issue is. And it started the “Incus” on port 53.
I will try to explain better what happend, but the whole thing is that port 53 was in use:
1: I upgraded, Pihole stared as always.
2: Instances was unable to run, as port 53 was in use.
3: Stop piehole, instances works and i an set up the VM’s again, but.
4: No pihole = no internet.
5: Downgrade.
6: Ajust VM’s again, and all is working.
Did you change the network settings in instances? You should almost certainly leave them at the default (they should especially NOT match the truenas IP)
Honestly it sounds like weirdness that just a reboot might have resolved, or something funky with the pihole config (maybe listening mode all has a weird interaction?)
Sorry that my explenation wasent clear. I’m trying.
Port 53, on my TrueNAS, the IP of my TrueNAS.
As i described in “2:”.Instances was unable to run.
Only after stopping Pihole, was i able to start Instances.
(This is from memory).
When going to instances, you need to do some kind og setup, before it starts.
You can se a IPv4 and IPv6.
I just tried pressing the OK, or apply or what it was, but when i did that it says
“Error port 53 is in use”.
So i tried typing in an unused IP, and it still says “Error port 53 is in use”
No matter what i did, or tried, it says “Error port 53”
So part “3:” I stopped Pihole and boom, it started.
But i was unable to start pihole, as port 53 is now in use.
I did a reboot (24 years in IT, networking, servers, VMware virutulization, firewalls… So a reboot is kind of second nature )
But it was the same.
My conclution is, that maby instances isent ready to be “in the wild” and the release as 25x as a stable is a bit of a stretch.
So, maby wait to upgrade, until it is not in “Experimental”, and is FINAL and working as “Virtual” is today.
I have the same issues with Pi-hole not working. This release could have done with some more testing I spent hours trying to get the instances to work. But each instance I create (from scratch) has a very small disk of a few GB, despite explicitly giving it a bigger disk. So the boot images don’t expand their main disk to the max available disk space. Stuff like this makes it really hard to use.
Yeh ive ended up deleting pihole, which is a shame, but the only way pihole starts is if i change its port from 53 to something else (just throws an error otherwise) but then it doesnt seem to get any hits, even when the nas ip is in my router. had same thing setting up adguard
I am digging in the same hole…
I have the pihole app in use since EE and run into this issue after upgrade to Fangtooth as I did my first trials with Instances (Incus) but did not even succed with the very first Global Setting (Pool = existing pool, Default Network Bridge = automatic):
(“dnsmasq: failed to create listening socket for 10.84.194.1: Address already in use”)
Incus behaviour:
After having stopped pihole I could finish the initial Setting of Instance. Due to “automatic” it used port 53 on a randomly choosed 10.0.0.0 network:
I removed the Instance config again (by setting the pool = disabled) and the port 53 listeners disappeared.
pihole behaviour:
Before restarting pihole again I tried to bind the pihole port 53 listener to the exiting (external) interfaces of my TrueNAS config:
Result:
pihole allocates all IPv4 addresses (0.0.0.0:53) and all IPv4 addresses (:::53) for its DNS listener what explains the initial “collission” with Incus:
Question:
For me it looks like a pihole issue rather than an Incus implementation issue. Any idea to configure pihole in a different way to avoid this problem?
But back to the root cause of this incompatibility between pihole and Incus:
As long as pihole uses 0.0.0.0:53, it will collide with Incus anyway.
No idea how to limit pihole to a singe ip (192.168.178.74:53)…
It looks like that I can select existing external host interface IPs for the management access (each HTTP and HTTPS with a specific high port to avoid confilcts with other GUIs) but for the DNS listener it takes all (0.0.0.0:53) for both tcp and udp. In other words, the configGUI is a bit missleading…
For my understanding docker will allow to restrict the listener to a specific host IP (as it does for the pihole management access), but for port 53 it is not implemented in the current pihole app.
I spent hours trying to get the instances to work. But each instance I create (from scratch) has a very small disk of a few GB, despite explicitly giving it a bigger disk. So the boot images don’t expand their main disk to the max available disk space. Stuff like this makes it really hard to use.
