Hello, I am getting "An error occured while loading the data. Presumably your log is too large to be processed." when opening Query log from the admin panel. My log file is very small (500kb). It also happens when log file is cleared from the panel. How do i fix this?
It looks like the log file is not cleared from dashboard. Maybe it doesn't have access to read/write it.
But looks like you got other major issue if ownership has changed from "dnsmasq:root" into "nobody:root".
Are you sharing your "/var/log" folder via Samba by chance ?
That's not one of our supported configurations. Can you post your configuration file for nginx and hopefully another user that has tried running under that daemon can help?
Web server config is file AFAIK. I think the problem could be that dnsmasq runs as nobody user. I have no idea how to make it create a user for itself and use that user.
CentOS 7 with lighttpd functions through test without issues or changes to the users or permissions. I'm pretty sure your stock permissions are correct.
Well, what could be the problem? Pihole status is unknown on admin panel, tail doesnt show anything and can't access queries page. This looks like a permission problem. What do you think?
There appear to be some issues with the testing environments and the install script. (It goes back to the NetworkManager status.) It may take a bit of time to fix the installer and then get us to a place that we can check for solutions to this issue.
Just wanted to update and not leave the issue hanging without a status.
@DanSchaper ,
I posted below to comfort, Discourse derailed (cloudflare error) and my post was gone ?
Luckely my editor kept a copy
I dont have a solution for you because Pi-Hole was designed to run on top of dnsmasq and lighttpd.
But you dont need the web GUI if your only interested in getting the ads to be blocked.
Most of the magic is done by dnsmasq resolving known blacklisted ad domain names to that of the Pi-Hole IP address eg:
So if you can get this bit to work, you are done for 90%.
The other part involves Pi-Hole to return a blank page instead of the ad via its own lighttpd daemon:
$ tree /var/www/html/pihole/
/var/www/html/pihole/
+-- blockingpage.css
+-- index.js
+-- index.php
For that to work, lighttpd got these configuration options:
So if you want nginx to do the same, you got to find the 404 error handler for nginx and configure yourself.
But Pi-Hole also depends an a few cron jobs to keep the blacklists updated and do some maintenance:
Another alternative is to have both web daemons running and have them listen to different IP addresses so they wont conflict (below is for Raspbian/Debian/Ubuntu):
I know I dont have a ready solution but at least you have a bit of an idea whats going on under the hood.
Sorry, we had a glitch with Discourse this morning on a software update and some posts made right before that process started were lost. So nothing was edited out or removed. I appreciate the assistance with this, so please continue with what you are finding.
Noted in the installer script though is that dnsmasqs user is set to nobody for CentOS installs.
Hmmm reading up a bit, I dont see any issue running it as user dnsmasq if it would solve the problem. Might need to create a user with /usr/sbin/nologin.
$ groups dnsmasq
dnsmasq : nogroup
In "/etc/dnsmasq.conf" you seem to be able to run as any user:
# If you want dnsmasq to change uid and gid to something other
# than the default, edit the following lines.
#user=
#group=
Just did a reverse proxy from nginx to lighttpd. Everything is working as expected but Query log is still showing "An error occured while loading the data. Presumably your log is too large to be processed." while the log size is a few kb.