Pi-hole keeps going down

I'm running Pi-hole on a raspberry with 4gb of ram and a 32gb SD card.

I have been running Pi-hole for about a month now with no problem. Last night I was suddenly unable to access any new webpages and google was giving me DNS errors. Assuming something with Pi-hole was acting up I tried to log into the UI and it just wouldn't load.

I am able to SSH over to the raspberry just fine, but Pi-hole is not working at all. I am using Pi-hole as my DHCP as well so it really messes everything up when it goes down.

After the restarting the the raspberry, Pi-hole will stay operational for various amounts of time before just breaking again.

https://tricorder.pi-hole.net/7pfhp23v5t

Can you somehow narrow down the time when the issue happened?
What are the last few lines in the two log files

  • /var/log/pihole.log
  • /var/log/pihole-FTL.log

?

pihole.log

Feb 24 14:55:01 dnsmasq[659]: query[AAAA] beacons4.gvt2.com from 192.168.0.5
Feb 24 15:00:02 dnsmasq[659]: gravity blocked beacons4.gvt2.com is ::
Feb 24 15:00:02 dnsmasq[659]: query[A] beacons4.gvt2.com from 192.168.0.5
Feb 24 15:00:02 dnsmasq[659]: gravity blocked beacons4.gvt2.com is 0.0.0.0
Feb 24 15:00:02 dnsmasq-dhcp[659]: DHCPREQUEST(eth0) 192.168.0.5 04:d9:f5:f2:fd:77
Feb 24 15:00:02 dnsmasq-dhcp[659]: DHCPACK(eth0) 192.168.0.5 04:d9:f5:f2:fd:77 Richard-Desktop
Feb 24 15:00:02 dnsmasq[13111]: query[A] mail2.gamedash.com from 78.128.113.130
Feb 24 15:00:02 dnsmasq-dhcp[659]: no address range available for DHCPv6 request via eth0
Feb 24 15:00:02 dnsmasq-dhcp[659]: RTR-SOLICIT(eth0) 32:96:bb:74:60:34
Feb 24 15:00:02 dnsmasq[659]: query[AAAA] wpad.lan from 192.168.0.5

pihole-FTL.log

[2021-02-24 14:02:10.772 9637/F659] Remapping "FTL-strings" from 114688 to (118784 * 1) == 118784
[2021-02-24 14:13:17.457 10571/F659] Resizing "FTL-dns-cache" from 28672 to (2048 * 16) == 32768 (/dev/shm: 3.8MB used, 2.0GB total, FTL uses 3.8MB)
[2021-02-24 14:13:17.659 659/T10572] Remapping "FTL-dns-cache" from 28672 to (2048 * 16) == 32768
[2021-02-24 14:24:07.515 659M] Resizing "FTL-strings" from 118784 to (122880 * 1) == 122880 (/dev/shm: 3.8MB used, 2.0GB total, FTL uses 3.8MB)
[2021-02-24 14:34:03.774 11819/F659] Resizing "FTL-dns-cache" from 32768 to (2304 * 16) == 36864 (/dev/shm: 3.8MB used, 2.0GB total, FTL uses 3.8MB)
[2021-02-24 14:34:04.628 659/T11820] Remapping "FTL-dns-cache" from 32768 to (2304 * 16) == 36864
[2021-02-24 14:40:31.000 659M] Resizing "FTL-strings" from 122880 to (126976 * 1) == 126976 (/dev/shm: 3.8MB used, 2.0GB total, FTL uses 3.8MB)
[2021-02-24 14:47:14.810 12593/F659] Resizing "FTL-dns-cache" from 36864 to (2560 * 16) == 40960 (/dev/shm: 3.8MB used, 2.0GB total, FTL uses 3.8MB)
[2021-02-24 14:47:15.612 659/T12594] Remapping "FTL-dns-cache" from 36864 to (2560 * 16) == 40960
[2021-02-24 14:48:12.218 12589/F659] Remapping "FTL-dns-cache" from 36864 to (2560 * 16) == 40960

I'm noticing in the pihole.log file it is saying "no address range available for DHCPv6 request via eth0". Might that have anything to do with this issue?

Is it expected that a public IP can query your Pi-hole ?


Something is really off with your Pi-hole:

Why did resolving this take (almost) exactly five minutes? This is definitely way off. How does your CPU and IO load look like when this happens?

I'm such an idiot! No, no one should be able to query my Pi-hole from a public IP, but I failed to setup my firewall correctly. I have now corrected my mistake and no one should be able to query it anymore.

Thanks to my retardation, I'm guessing my DNS was being used for some nefarious purposes and that might have been what was causing my issues. Do you agree?

Here is a picture of my performance incase you still want it:

Capture.1PNG

This is what my Pi-hole usually runs like through the whole day, so I don't see why it could be a load issue.

Maybe. We can only tell if things are stable now. Even if it was used for nefarious purposes, this shouldn't have caused the 5 minutes delay. We'll see how things evolve from here now that you are (again) shields up.

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.