After update to 2.5 slow login and 504 on general setup.

[beermount]

In short, updated from 2.4.5p1 to 2.5. Hardware is x86_64 Celeron G3930 with 8GB RAM and SSD for storage.

Never had any issues with performance before, but after the update, getting to the dashboard takes 50 seconds (login page is fast). And the general setup I end up with a 504 gateway timed out error.

The only log entry I can find that seems related is this one, which would be 3 minutes after I clicked "General Setup" menu option:

Mar 13 18:40:22 pfSense nginx: 192.168.40.138 - - [13/Mar/2021:18:40:22 +0100] "POST /ifstats.php HTTP/1.1" 200 342 "http://192.168.40.1/" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/88.0.4324.190 Safari/537.36"
Mar 13 18:40:23 pfSense nginx: 2021/03/13 18:40:23 [error] 2445#100694: *2008416 upstream timed out (60: Operation timed out) while reading response header from upstream, client: 192.168.40.138, server: , request: "GET /system.php HTTP/1.1", upstream: "fastcgi://unix:/var/run/php-fpm.socket", host: "192.168.40.1", referrer: "http://192.168.40.1/"

When accessing the dashboard, I can see two processes of netstat running (using top -SH), at ~100%. When the page finally is done loading, those two processes disappear from top.

I do have some other pfSense installations, which have been upgraded to 2.4.5p1, that does not have this issue, but I was hoping to find out what would be causing this.

As a FYI: I'm also struggeling with FRR and a full IPv6 BGP table that is not installed into the kernel routing table any longer. ~38000 of the 110000 routes has been installed, which might be causing the slow loading of the dashboard?

I guess this could be related? https://redmine.pfsense.org/issues/11475#change-52321

EDIT: After testing out for a bit, it seems the slow dashboard loading issues, is caused by the "Gateways" widget.

[beermount]

If I remove the ~38000 IPv6 kernel routes I did get installed via BGP, then I'm able to access general setup.