I've moved this topic to Community Help, hopefully someone will see this and be able to provide you with the answers you have concluded are the only answers possible.
1 Like
That is a deliberate insinuation on your behalf.
Look, I have given you the hooks how Pi-hole could be related to your issue in my very first reply.
However, you denied any possibility that Pi-hole could be involved by these hooks during our exchange, and you did also preclude that your hardware could be at fault.
I cannot know your network configuration, nor your skills.
Guiding you through potential network issues is out of scope for Pi-hole, but even to judge Pi-hole's involvement, I have to rely on precisely what information you provide.
You are making it hard to help you if you reinterpret my requests for information to be questioning your skills, preferring to just repeat the same observations again and again instead, forcing me in turn to repeat them up to three times before a reaction is received.
In our following exchange, you didn't provide any additional leads to DNS, and lately, you've also ruled out that Pi-hole's DHCP server functionality is enabled.
None of the information you've provided supports that Pi-hole would be involved, and I've presented that conclusion as late as in my last post.
I am usually open to reassess on additional information, but I am certainly much less now you're resorting to personal accusations.
Pihole is maybe not 'active' on my network but it's certainly active on my Rpi4. If the program crashes or whatever happens I don't expect my complete wired network to go down with it. And with down I mean every single device doesn't receive an IP. When I powercycle the RPI my network comes back almost immediately. To me it that sounds it's doing something on my network.
So yes you are right. Pihole does not have a role on my network, except doing dns requests, but it certainly plays a role on my network to whatever happens at the moment when it happens.
And to be fair I'm not really waiting for answers like 'if xxxx can't help me then i'm on my own'. That only adds salt to the conversation. I only want to stop pointing fingers, you say it isn't pi-hole and I say it is somewhat related. All I want is to isolate this further because this is not normal to what is happening. It was a clean Debian install and a new pull docker request. From a container I would expect it works out of the box just fine. And I can safely say that this happened from the beginning. Only reason I didn't noticed it so fast was because I restarted the pi on a daily basis back then because of home assistant etc.
If you don't want to help me that's fine. I thanked you numerous times. All I want is that someone can point me in the right direction or give me an idea. I tried so many things which I hardly can remember. The problem usually occurs after a few days (since last reboot) which makes it a pain to troubleshoot properly.
To do steps;
tcpdump capture on the rpi
Install pihole via bash script and turn off docker version.
How do you apply Pi-hole as DNS for your network? Is it the routers upstream or provided as DNS server via DHCP (or manually)?
What is the upstream DNS of your Pi-hole?
Also check Pi-hole's query and application logs (in /var/log/, although not sure how to do with the Docker container) for abnormalities and quantities, and also dmesg (on the RPi host) for any network related errors or limit hits and such. Also check your routers logs, especially after the crash happened.
Probably a DNS loop is causing this, or Pi-hole's reverse DNS attempts are involved.
Another thing to rule out or verify other applications on that RPi being the issue could be to systemctl stop pihole-FTL and use another upstream DNS via router while keeping the RPi at the network and see if it still crashes or survives.
1 Like
Atm I only let my main-pc go through pi-hole. I don't use Pihole DHCP only my router.
Now that I see this there was a time I filled in the alt. DNS server also google 8.8.8.8. In case my rpi went down or was rebooting I still had internet on my PC. Could that be an issue?
Router is using my ISP DNS settings for the moment.
I use Google a upstream DNS
Thanks, I can work with those suggestions when it happens again.
Okay, a DNS loop is then basically ruled out.
I'm not sure how Windows uses the fallback, but I don't see a way how it could cause an issue.
You mentioned tcpdump earlier. When only one client uses that Pi-hole instance, that could be good as well. When it's about days and week, clever filtering/logging would need to be done to not have a too long list to look through. But probably some hours are enough to check for unexpected requests floating around.
Do you have IPv6 enabled in your home network/router? I have no suspicion about it but it's advertisement-based function sometimes causes unexpected issues, and since it usually does not serve any purpose in LANs, it should be save to disable and see if that changes things. I don't mean to disable IPv4 on the clients, just in the router so that it does not send out router advertisements (simplified the IPv6 variant of DHCP).
No, ipv6 is turned off on my router
I just experienced the problem again.
-
Ethernet devices keep last known IP address
-
Cannot ping ANY eth devices
-
Cannot reach my router from my main pc (eth)
-
With laptop I can reach router
The moment i Turn off my RPI my network is INSTANTLY back.
From what I've seen so fast I could not see anything weird in pihole query log but also not sure if it recorded everything due shutdown. I will check later when I'm done with work if can see see, once again, any abnormalities.
For now I turned off docker Pihole and installed Pihole via de regular bash script.
I had a similar issue once that was driving me crazy. I finally found out that a port on my router was flaky and causing the entire network to go down. It's a pretty long thread, but have you tried plugging the Pi into a different port on the router if one is available? If not, can you try and connect your pi to the wireless network instead of a physical connection? That was you can determine if the issue is resolved or moves to the wireless wireless network with your pi?
That's not a bad suggestion.
The pi is the only device that is plugged into a port straight into my router. All other devices are connected via a switch or wireless.
Although this shouldn't explain why this is only happening when I have pihole running. But as I said earlier I'm open to anything to rule everything out.
Ill add it to my list. Today/tomorrow I reach the usual ~2 day mark when shit hits the fan after a restart.
Just for the record.
Since I stopped the docker container and switched to the pihole bash installation method I haven't experienced any problems as of 'yet'.Usually it happens with a day or 2-3 after a reboot.
So far my suspicion that it was related to my debian/docker/pihole setup is being confirmed. Although I have no explanation why, seeing that I used the docker-compose file that is on the pihole docker instruction page. And the docker container itself is pre-made and I haven't altered it any way. But this was all the reason for me to post this in the docker section and request help from here.
Perhaps there is a conflict in my docker network although I don't see why. Pihole was on his own subnet.
For now I'll keep using this 'method' of Pihole for at least another week just to be sure that it doesn't happen again. Then I could try/test the other suggestions or maybe even assign a dedicated 'internal' IP to pihole/docker container. Or I might even drop this issue and just use the normal installation. There aren't any big advantages to use pihole docker in my case anyway.
This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.