Every ~24hrs, my pihole will become unresponsive until I physically disconnect it and power it back on. During that window, I cannot ssh to the pihole, and after rebooting, the query log shows that there were no queries during that time.
I’ve tried searching and a variety of debugging steps I found online with no luck. At this point, I think the problem is unfortunately outside of my expertise / knowledge to know both what to fix, and what to check to continue finding potential fixes.
Expected Behaviour:
My pihole should operate consistently, without requiring a hard reboot.
Actual Behaviour:
Every ~24hr, I have to physically unplug my pihole and reconnect it to get it working again.
So far, I have not been able to find any smoking guns in my logs via journalctl or by looking in the pihole or FTL logs. My only remaining theory, and I don’t know if this is possible or not, is that when my DHCP lease is being renewed (I still have my router doing DHCP) this causes some issue with the service?
I have already configured the pihole with a static IP both at the router level, and at the OS level, and neither seem to be able to fix the issue.
Without ssh, is there a way I can connect to the pihole during one of these blips to try and provide more diagnostic information? Or, is there any additional debugging output I could try to provide here?
Okay, I will try to do that next time. Aside from that, is there any additional debugging or information you would recommend I try to compile, or perhaps any other potential root causes I might be able to try and investigate?
Mine becomes unresponsive when I try to browse the query log page; a chicken and the egg thing, when it does this, I cannot even ssh into the unit to do deeper troubleshooting.
It is almost like the PiHole becomes resource-constrained and it locks up.
The device running Pi-hole should always have a static IPv4 address on all connections used (eth0, wlan0, etcetera).
Are using log2ram? if so, run: echo "Filesystem Type Size Used Avail Use% Mounted on" && df -hT | grep log2ram && echo && sudo du -hs /var/log/* | sort -hr | head -n 5 && echo && sudo ls -lah /var/log/pihole/pihole.log*
and ensure there is sufficient free space to rotate the pihole.log file. Sufficient space is two times the size of the pihole.log file. If not, edit /etc/log2ram.conf.
Your Pihole.log files should rotate at midnight each day. If the latest log file hasn't been rotated: sudo logrotate -f /etc/pihole/logrotate.
I ran into this problem with a pi5 and zero. Using the RaspberryPiOS seemed the most prone to it, dietpi seems much better with it, but I keep a cron script that runs every 5 minutes to check network connectivity and bring the interface down and up immediately if there’s a network stack issue. Haven’t had that particular issue happen except on power outages where I need to physically unplug it and plug it back in for whatever reason.
I’ve never seen rate limiting errors in the logs after the issue occurs, and temperatures generally remain quite constant for my hardware. There haven’t been any errors at all, which makes all of this even harder to diagnose.
My pihole became entirely unresponsive today, failing to recover or boot back online after the issue occurred. I’m going through a full deep reinstall using an entirely different OS (DietPi this time) to see if that helps.
Unfortunately I can’t say for sure yet. Since the issue occurs somewhat inconsistently and infrequently, it’ll just be a matter of waiting to see if the error reoccurs or not after my reinstall.
If I haven’t had a hard freeze in a week or two, I’ll come back and resolve the thread with a full OS reinstall as the apparent solution.