Please follow the below template, it will help us to help you!
Expected Behaviour:
Name resolution should occur all the time
pi.hole should load the web interface all of the time
Known ads domains should be blocked
Actual Behaviour:
Sometimes name resolution will occur however mostly it isn't. When it works it will usually work for a period of time before failing.
Most of the time pi.hole fails to resolve and is only accessible by IP address.
Sometimes ads from known blocked domains will show/play
Debug Token:
p8a5rgahj3
Other info
Running hostapd with wlan0 bridged to eth0 - have run this for a while without issue. Clean install several days ago (as decided to change wifi to 802.11n/ac after started having speed issues). Since clean install worked fine for a while but issues have got progressively worse. DNS server is unbound. Dig to unbound works instantly and returns the expected result. Dig on pihole port returns temporary failure in name resolution error. NB after running pihole -d dig to pihole is returning normal response again - I would expect this to fail after a period of time as it usually does.
After much fiddling I've narrowed down the cause and believe for sure it is Pi-hole for the following reasons.
Doing dig directly on the Raspberry Pi via unbound resolves instantly while the same domain via Pi-hole fails with 'connection timed out; no servers could be reached'. Performing the same dig on my Mac (which has the Pi-hole as it's DNS server) also resolves instantly.
Now here is where it gets interesting. The failed dig on the RPi via Pi-hole will actually resolve about 3 minutes later (at least this is when it appears in the logs). Eg perform dig at 13:10:16 which times out in the terminal but will show 3 minutes later in the logs with timestamp as 13:13:20. Something else I noticed is that stopping pihole-FTL and restarting it means I can get one successful dig from the RPi via Pi-hole before each subsequent request times out.
This is as far as I've got - I believe queries from other devices are resolving most of the time but from time to time they don't (or perhaps they are taking 3 minutes or so as well) as the device will have issues resolving a website or Netflix starts having a fit and throwing up Netflix errors.
Not really sure what to look at next to work out why this is happening though. Any suggestions welcome and appreciated.
Interestingly the debug ran straight through. Normally it pauses for a while the *** [ DIAGNOSING ]: FTL version (I think) part before it goes through.
Likely because I rebooted the Pi (and possibly the router) several times as part of trying to fix it??? I thought ULA was standard in Pi-hole now? My router option for ULA is enabled anyway.
It's been fine since reverting to the standard lists. I've since enabled lists slowly and without issue - am now back to the previous set of lists and all seems to be working fine. Will keep monitoring.
So it worked for 24 hours before it decided to play up again. I removed one list and it worked for about half an hour before it had issues again. Have now removed all but the default and will see if any issues occur over the next few days or so.