After upgrading to Unraid 7 - 500 Internal Server Error on both web gui and local GUI Safemode

[jedimstr]

Upgraded a few of my servers successfully including a security/backup server and a homelab testing server, so decided to pull the trigger on my main media server.  That primary server is now getting Internal 500 Server Error in both the Web GUI and in local SafeMode GUI (just in case it was one of the plugins that prevented nginx from working as others have seen with this error in the past).  But even in the local Safemode GUI instance I get an error, this time a 502 in the local browser.  All my docker containers like Plex, Grafana, the **arrs, Tautulli, etc all are working fine and plugin based UIs like Unbalanced is working, so it just seems like the main webGUI is broken.  Trying to access dashboard, plugins, docker views via their direct URLs all result in 500 errors.  SSH works perfectly fine as well.  Because of the multiple restarts in trying to diagnose the issue I see in my Slack notifcations that its performing a Parity Check so assuming the Array in general is fine.  

 

Trying to restart or manually stop and start with these commands doesn't fix anything, still 500 errors:
/etc/rc.d/rc.php-fpm restart
then tried these:
/etc/rc.d/rc.php-fpm stop
/etc/rc.d/rc.php-fpm start

Attached is my Diagnostics zip file.  I'm hoping there's some simple fix here that I missed modifying a config or playing with something on the /boot/config/plugins or in the go file to fix this.  I'd hate to rebuild the flash from scratch on this server and trying to recover all my settings, plugins, and whatever customizing I did over the years.

holocron-diagnostics-20250109-2352.zip

[JorgeB]

Try typing

 use_ssl no

 

And then accessing with http

[jedimstr]

Quote

5 hours ago, JorgeB said:

Try typing

 use_ssl no

 

And then accessing with http

 

Nope, still getting 500 internal Server Error on http with use_ssl set to no.  I use my own certs and domain and other services that use those certs are still working like Grafana. My two other servers have pretty much the same plugins installed and also use my wildcard cert & domains and they're working fine with 7.

[JorgeB]

You can try redoing the flash drive, backup the current one first and then recreate it using the USB tool and just restore the bare minimum, like the key, super.dat and the pools folder for the assignments, also copy the docker user templates folder (\config\plugins\dockerMan\templates-user), if all works you can then reconfigure the server or try restoring a few config files at a time from the backup to see if you can find the culprit.    

[jedimstr]

I ended up following the instructions here for manually downgrading back down to 6.12.14 and that works fine including switching back to HTTPS mode with my own certs.

Will need more time to diagnose and try upgrading again to 7 at a later date, but for now I'll stick to the working WebGUI config, or I'll wait for a later revision to try again.  If anyone catches any glaring potential sources of the issue in my Diagnostics Zip, let me know.

Edited Friday at 05:59 PM by jedimstr

[jedimstr]

After a bit of experimental cleanup of some plugins and removing a DVB Driver and Unraid Connect's beta plugin, tried upgrading to 7.0 again.  Same issue, 500 internal server error.  Manually downgraded back to 6.12.14 again and WebGUI is working fine again.  Def something is conflicting with my setup on this one server and 7.0.  My two other servers as mentioned above are perfectly fine with 7.0.

 

Captured a second Diagnostics.zip after the second attempt to help narrow down what it could be.

 

holocron-diagnostics-20250110-1514.zip

Edited Friday at 08:34 PM by jedimstr

[ljm42]

You have a very complex go script, and it looks like it messes with nginx/php config:

cp /boot/config/www.conf /etc/php-fpm.d/

I'd recommend commenting all of your go script mods and re-implementing them one at a time in Unraid 7

[jedimstr]

1 hour ago, ljm42 said:

You have a very complex go script, and it looks like it messes with nginx/php config:

cp /boot/config/www.conf /etc/php-fpm.d/

I'd recommend commenting all of your go script mods and re-implementing them one at a time in Unraid 7

That was it!!!  The custom www.conf settings were based on an older version of the www.conf that worked fine for 6 but not 7.  The original reason for the custom www.conf was to raise the pm.max_children to 100 from the default of 50 but it may no longer be necessary. 

The actual breaking line within the www.conf itself was this in the 6.x version:
 

listen = /var/run/php5-fpm.sock

vs
 

listen = /var/run/php-fpm.sock 

in the 7.x version.

 

I'm keeping both the www.conf and sudoers cp lines commented out in the /boot/config/go since those old mods probably are no longer necessary for what I was trying to do originally.

For reference, here's the diff of that www.conf file:
https://www.diffnow.com/report/bjr2l

 

Problem solved and running 7.0 on my media server now! Thanks!!!

Edited Friday at 11:55 PM by jedimstr