Can you Please check my debug log, I am having issues where pi-hole stop responding for up to minutes at a time, all devices lose connectivity, its in docker and the only files kept are the database and the /etc/pihole and /etc/dnsmasq.d folders
Your debug token is: https://tricorder.pi-hole.net/yEXiKHut/
could it be related to dhcp lease time, its like the lease is not renewed but its all devices at the exact same time
Can you please also post the compose file (or docker run command) used to start the container?
And the contents of external.conf?
I think I made changes to external.conf
I think i edit that file by copying it
sudo docker cp /docker/1.external.conf pi-hole:/etc/lighttpd/conf-enabled/
and all that is doing is ssl support and redirect everything thats not /admin to the block page
ssl support does need the use to press continue to page
sudo docker run --name pi-hole --network macvlan0 --hostname "pi-hole" --ip 192.168.1.252 -e WEBPASSWORD=password-v /docker/pi-hole-1/etc/dnsmasq.d:/etc/dnsmasq.d -v /docker/pi-hole-1/etc/pihole:/etc/pihole --cap-add=NET_ADMIN pihole/pihole &
lighttpd and SSLI asked for external.conf file because your SSL configuration is broken.
There are many messages about "unsupported protocol" and "expired certificate", like these:
2024-06-16 19:09:53: mod_openssl.c.3095) SSL: 1 error:14209102:SSL routines:tls_early_post_process_client_hello:unsupported protocol
2024-06-16 19:12:14: mod_openssl.c.3095) SSL: 1 error:14094415:SSL routines:ssl3_read_bytes:sslv3 alert certificate expired
There are many warnings likes this ("Broken pipe"):
[2024-06-16 21:18:23.009 465709/T465857] WARN: Could not write() everything in getStats() [src/api/api.c:133]: Broken pipe
I'm not sure what is causing it.
Maybe @DL6ER can explain it better.
It means that the PHP process was interrupted before FTL could finish sending all the requested data. This is likely hinting at another webserver issue - the PHP process should not exit/die before it is finished.
This may actually directly relate to the original problem
Hypothesis ahead: If PHP for some reason decides to keep the API blocked by not reacting (or whatever), this can theoretically cascade down into the DNS server blocking this one as well. Potential for something like this to happen is drastically reduced in the v6 beta but it is still existing in many places in the latest v5 release - typically it is not causing issues.
We can check this by setting DEBUG_LOCKS=true and then checking the last lines in /var/log/pihole/FTL.log when it happened. If the lock is continuously held in the API thread, then our hypothesis is true.
Either way, it seems of significance to fix the lighttpd issues.
the ssl configuration has probably been broken a long time
all it needed to do was allow https to redirect to the block page, always said the connection was not private which its going to because the certificate was for my website not every website in the world
this is a reply to both messages @rdwebdesign @DL6ER
could this be the script I have deleting queries, that get too frequent and are pointless like my raspberry pi updating openweathermap every few seconds, every 15 minutes then restarting the dns
do I set debug_locks in pihole-FTL.conf
strangely this has not happened today, but it was doing it all the time yesterday
do it just not setup properly with the bootup script I have, where I delete and recreate the container
the system is set to restart 9AM every Monday in crontab, so maybe that fixed the issue, and also maybe it didn't, I havn't really been on my PC much today
every time I have an issue, it magically fixes itself every time I ask for help