I installed Pi-hole on a fresh install of Debian 11 x86-64. The only other thing I have installed so far is isc-dhcp-server, which I installed before Pi-hole. I expect it to launch at when the system starts.
On startup, Pi-hole fails to start. I find this in the system log:
Aug 03 09:42:45 gateway pihole-FTL: install: cannot create regular file '/run/pihole-ftl/pihole-FTL.pid': No such file or directory
Aug 03 09:42:45 gateway pihole-FTL: install: cannot create regular file '/run/pihole-ftl/pihole-FTL.port': No such file or directory
Aug 03 09:42:51 gateway pihole-FTL: install: cannot create regular file '/run/pihole-ftl/pihole-FTL.pid': No such file or directory
I can see that the init script is supposed to create the directory /run/pihole-ftl and those files, as well as /run/log/pihole-ftl, and it assigns ownership to pihole:pihole but this isn't happening. It is my understanding that the script is run as root, so it should have permission to do this. If I manually create those directories and assign ownership as expected, it works. /run/pihole is created by something but doesn't contain any files. It is owned by pihole:pihole.
The fact that this interface has your public IP is a concern. There should be no exposure of the Pi to the WAN - it should be completely NAT'd behind your router. No DMZ, no port forward, etc from the router.
Maybe that's an issue of best practices, but this is my edge device. That's where I have it running right now. Maybe I'll move it if I build another machine but for now, it seems unlikely that that could be preventing the init script from creating a directory.
I've done that before but it brings up another issue that I forgot to mention earlier. This shows up occasionally, yet it actually is running. The dashboard shows it as active and the clients are able to use it.
Interesting. There was no thought process. I never knowingly changed that manually. I migrated the config files from a previous install on a Manjaro system. I never messed with paths there but if there are differences in what the installer does on different distributions, it might come from that. That's the only way I can imagine it got changed. That seems like a weird place to put that and I would never have done it intentionally. I guess I should have started from scratch instead of trying to reuse my old config.
rob@gateway:~$ pihole status
[✗] DNS service is NOT running
[✗] DNS service is NOT running rob@gateway:~$ pihole restartdns
[✓] Restarting DNS server rob@gateway:~$ pihole status
[✓] FTL is listening on port 53
[✓] UDP (IPv4)
[✓] TCP (IPv4)
[✓] UDP (IPv6)
[✓] TCP (IPv6)
[✓] Pi-hole blocking is enabled
You appear to have some errors in your query database (about 12 minutes after Pi-hole restart):
-----tail of FTL.log------
[2022-08-03 16:20:24.744 45689M] SQLite3 message: no such column: aliasclient_id in "SELECT aliasclient_id FROM network WHERE id = (SELECT network_id FROM network_addresses WHERE ip = ? AND aliasclient_id IS NOT NULL GROUP BY ip HAVING max(lastSeen));" (1)
[2022-08-03 16:20:24.744 45689M] getAliasclientIDfromIP("10.0.1.29") - SQL error prepare: SQL logic error
[2022-08-03 16:20:25.084 45689M] SQLite3 message: no such column: aliasclient_id in "SELECT aliasclient_id FROM network WHERE id = (SELECT network_id FROM network_addresses WHERE ip = ? AND aliasclient_id IS NOT NULL GROUP BY ip HAVING max(lastSeen));" (1)
[2022-08-03 16:20:25.084 45689M] getAliasclientIDfromIP("10.0.1.48") - SQL error prepare: SQL logic error
You can run a quick check on this database with the following command: