Expected Behaviour:
Working Pi-hole with dns service running as well
Actual Behaviour:
dns service down, high temp high load on the system. Temp over 105 degree celsius load just over 2 (got 2 cores)
Debug Token:
1ppw12usbv
Maybe an idea to post more info about the hardware and software used ?!
With only the debug token, only the developers can peek 
Gladly provide you with more info:
Thank you for stumbling upon this topic 
Running on Turris Omnia router
Operating system: debian 9 in lxc (latest packages)
hardware: 2 core ARMv7
2gb RAM
2tb HDD
8gb flash (debian is running on top of ext4 FS), while Turris OS runs on BTRFS
10gbit of switching capability internaly with SFP (2,5Gbps), 1x WAN + 5x LAN ports all GLAN.
There is suricata being run on not virtual userspace and other software. Normaly the load is only about 0.05 - 0.1 (meaning max 5% of CPU)
There was a process that awaited dns response and was agresively trying to do so. This process is a part of haas https://haas.nic.cz/ service. So I decided to uninstall this from my router because I love Pi-Hole as my DNS resolver for internal network.
I got a chance to switch to default kresd (Knot resolver), which is great and better than dnsmasq when it comes to privacy and security. But I love Pi-Hole.
I need DNS to be always up and running without it internal devices are not able to use internet at all. I am not willing to change DHCP info about DNS to include some unsecure DNS to mitigate this.
Are you running stock Pi-hole or the dev branch ?
If its dnsmasq you suspect not working, post results from below ones so we can have a look:
sudo systemctl status dnsmasq -l
tail /var/log/pihole.log
sudo netstat -nltup | grep 'Proto\|:53 \|:80 \|:67 \|:4711 '
And for the high load, the top command might give you a clue.
Hey,
stock pihole.
Just pushing enable in admin interface works . But I still do not know why and what mechanism caused dns resolving to be down?
Is it a feature?
I do have to add that I have started using DOH Cloadflare per advice of someone using Pi-hole.
Thats a good laugh ... thanks 
Do you mean the enable/disable buttons on the admin page ?
They just enable/disabling filtering and not the DNS service itself.
How have you determinded DNS is down ?
Any errors ?
Strange because it is enabling dns service while that dns service killed itself.
It was like there was +1 button then normally.
Determined by:
dns queries were not geting resolved...
admin interface said it in the status corner
Just disable it and on a clients pc run below one on a command prompt to test DNS resolution:
nslookup pi.hole <PIHOLE_IP_ADDRESS>
Solved now ?
My pihole is online, dns service is running. But when I have my scheduled cron auto update consisting of
pihole -g and pihole up daily at night
after that cron job dns service is down
caused by the machine encountering load of more than 2.0 in linux terms.
Next time its down, run these ones:
sudo systemctl status dnsmasq -l
tail /var/log/pihole.log
sudo netstat -nltup | grep 'Proto\|:53 \|:80 \|:67 \|:4711 '
EDIT: owh and a load of two on a two core system isnt that high.
Ive seen web servers with 4 cores have a load of 16 with no issues at all.
Load: 0.85 0.44 0.24 after doing pihole -g and pihole -up
I have decided to restart the machine this noon and loads are way lover to some extent related to HAAS being offline (whole ucollect actualy)
Things are stable now. I just fear of turning Ucollect back online.
You know, DNS service is fucking important to have online all the time. I am not willing to risk much. Many people relly on my decisions.
More info about that software here:
https://doc.turris.cz/doc/en/howto/turris_software
I will find a way to change DHCP setting to include some reserve dns then I will test it more thoroughly
Please remove the f.. ?
DNS is important but to reasure, Ive got an old Pi1 now close to a year running Pi-hole and no glitch at all.
It has worked perfect all that time.
This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.