50x Error on System/Update

MarioCaires

Crash report begins.  Anonymous machine information:

amd64
16.0-CURRENT
FreeBSD 16.0-CURRENT #12 plus-RELENG_26_03_1-n256546-1d1bfd578383: Wed May 20 15:20:10 UTC 2026     root@pfsense-build-release-amd64-2.eng.atx.netgate.com:/var/jenkins/workspace/pfSense-Plus-snapshots-26_03_1-main/obj/amd64/fvF1vE9r/var/jenkins/workspace/

Crash report details:

PHP Errors:
[16-Aug-2026 13:08:21 Europe/London] PHP Fatal error:  Uncaught ValueError: gettext(): Argument #1 ($message) is too long in /usr/local/www/pkg_mgr_install.php:447
Stack trace:
#0 /usr/local/www/pkg_mgr_install.php(447): gettext()
#1 {main}
  thrown in /usr/local/www/pkg_mgr_install.php on line 447



No FreeBSD crash data found.

stephenw10

The webgui works OK other than that page?

Can you upgrade from the console?

Anything else logged? That error is pretty generic unfortunately.

MarioCaires

Hi stephenw10,

The webui works ok other than that page! The PHP Error Log shows the same info.
I haven't tried yet to upgrade as I'm currently on 26.03.1 and accessing it remotely. I will try to upgrade as soon as I am at home and report back.

MarioCaires

I managed to update my pfsense box from the console, and after rebooting the error no longer is there!

Primoz

I have the same issue after upgrading from CE 2.8.1 to 2.9.0. After installing the router hanged with the webpage displaying 50x with nothing working. After rebooting (once or twice) it started fine. Lately I've noticed that even though it works, the webgui hangs and becomes unreachable. I have to reboot the router multiple times to get it to work as the first reboot usually results in 50x.

The routing appears to be working underneath, after the initial upgrade and reboot Wireguard did not work though.

I tried SSHing to the router but SSH has always been spotty and very rarely worked for me so I can't check anything right now. Fixing SSH and this issue repeating (doesn't seem hard to reproduce, just a reboot) might be a way to give some more info. Or just accessing logs from the GUI. In the system log the main thing I noticed is Could not connect to /var/run/php-fpm.socket.

What should I check next?

EDIT: after rebooting and the GUI loading up normally, SSH also works.
EDIT2: GUI just crashed (is completely unreachable) and SSH also doesn't work. Internet (so routing) works.

Gertjan

@Primoz said in 50x Error on System/Update:

What should I check next?

The system log.
If GUI and SSH aren't cooperative, use the console access, and have a look : menu option 8, and then

cat /var/log/system.log

And typically this one :

cat /var/log/system.log | grep 'link'

Do you have a lot of link events per hour, or even minute ?

When I count mine, I see this :

cat /var/log/system.log | grep 'link' | wc -l

but that's fine for me, as the oldest entries in my system log are from April this year, so about 4 months worth of logging.

Primoz

Connected a PiKVM to the machine, no signal/image. And this is after the machine was off for a few hours (since the last post), so it's not directly connected to thermal issue (if it could cause direct problems). The first boot after cooling off also gave a 50x. I rebooted now and I get an image, I can SSH and can log into the webUI.

checking system.log (over SSH so I can actually scroll) all I get before the reboot is spammed "pfSense check_reload_status[498]: Could not connect to /var/run/php-fpm.socket" and kernel notifications of me connecting the PiKVM.

grep link gives (obviously due to the reboot and due to all the php-fpm spam from before):

Aug 30 14:41:26 pfSense kernel: tun_wg0: link state changed to UP
Aug 30 14:41:26 pfSense kernel: lo0: link state changed to UP
Aug 30 14:41:26 pfSense kernel: pppoe0: link state changed to DOWN
Aug 30 14:41:26 pfSense kernel: bridge0: link state changed to UP
Aug 30 14:41:29 pfSense kernel: igc3: link state changed to UP
Aug 30 14:41:29 pfSense kernel: igc0: link state changed to UP
Aug 30 14:41:29 pfSense kernel: igc0.3902: link state changed to UP
Aug 30 14:41:29 pfSense kernel: igc1: link state changed to UP
Aug 30 14:41:30 pfSense kernel: igc2: link state changed to UP
Aug 30 14:41:31 pfSense kernel: pppoe0: link state changed to UP
Aug 30 14:41:31 pfSense kernel: pppoe0: link state changed to DOWN
Aug 30 14:41:41 pfSense kernel: pppoe0: link state changed to UP
Aug 30 14:41:41 pfSense kernel: pppoe0: link state changed to DOWN
Aug 30 14:41:51 pfSense kernel: pppoe0: link state changed to UP
Aug 30 14:41:51 pfSense kernel: pppoe0: link state changed to DOWN
Aug 30 14:42:01 pfSense kernel: pppoe0: link state changed to UP
Aug 30 14:42:01 pfSense kernel: pppoe0: link state changed to DOWN
Aug 30 14:42:05 pfSense kernel: tun_wg0: link state changed to DOWN
Aug 30 14:42:07 pfSense kernel: tun_wg0: link state changed to UP
Aug 30 14:42:11 pfSense kernel: pppoe0: link state changed to UP
Aug 30 14:42:11 pfSense kernel: pppoe0: link state changed to DOWN
Aug 30 14:42:22 pfSense kernel: pppoe0: link state changed to UP
Aug 30 14:42:22 pfSense kernel: pppoe0: link state changed to DOWN
Aug 30 14:42:32 pfSense kernel: pppoe0: link state changed to UP
Aug 30 14:42:32 pfSense kernel: pppoe0: link state changed to DOWN
Aug 30 14:42:42 pfSense kernel: pppoe0: link state changed to UP
Aug 30 14:42:42 pfSense kernel: pppoe0: link state changed to DOWN
Aug 30 14:42:52 pfSense kernel: pppoe0: link state changed to UP

stephenw10

So those logs were all after the first boot and before the reboot?

Did you see the same link connects after the reboot when it was still responding normally?

Primoz

This is when it was responding normally after a reboot. Above the copied text it was all fpm.socket, repeated about 5 times per second so obviously the log got filled up real fast. so fpm.socket messages are there from before the reboot (unresponsive SSH and webUI) and the link state messages are from after a functioning reboot.

Another symptom I'm having is a more or less non-functioning wireguard with the routing functionality otherwise working at home. A different situation is the wifi AP not connecting or something (didn't do anything to it and rebooting the pfsense box remedies it) where any wifi device doesn't have internet.

Weird things.

I'll probably do a fresh install but I'm still deciding whether to install 2.7.2 that I still have an ISO on my network or try the new installer (should be fun with my special sauce PPPoE connection and setting that up to try to install over the internet...).
Honestly, the fact there are no more ISOs available anymore plus the fact the ZFS boot environments are paywalled makes me both sad I didn't push through with trying out OPNsense a while ago and makes me think about trying it out again... If I had boot environments I'd just roll back to 2.8.2 and see if it's the same. If it is, it's a HW issue on my side. Otherwise it's something with my install.

stephenw10

How special is your PPPoE? The installer can do PPPoE.

Also you can use ZFS BE snaps in CE it just doesn't do anything automatically.

Primoz

@stephenw10 it's setting a VLAN, MTU, username and password obviously, that's what I remember. I got the instructions off a local forum, set it up and forgot about all the details. It allows me to not use the ISP modem/router/AP at all.

I know you can, but for that you need to have it setup beforehand.

Anywho, I managed to get the 2.7.2 setup working with the memstick image afterwards and update a spare SSD to 2.8.1, then when I had that working, clear out the original SSD, do the same and update all the way to 2.9.0. Everything worked perfectly until just recently when I rebooted the router a few times. I was missing a speedtest widget that I had installed before and I can't setup again on a fresh 2.9.0 install (the package is too old) so I decided to do a fresh install again, install the widget and update to 2.9.0 to see if maaaaaaaybe the widget is the problem.

Looks like it's not. I'm having the same issues with an unresponsive webUI, no image when connecting PiKVM but the internet is working.

EDIT: okay, fresh install of 2.7.2 with the speedtest widget, updated to 2.8.1 on the old SSD. If there will be problems here, I'm switching SSDs. If there are problems there as well, it's something with a different part of the system (RAM maybe).

stephenw10

Nothing logged at the time it stops responding?

Primoz

@stephenw10 nothing that I could check other than what I already posted.

For what it's worth, I've now been running 2.8.0 without any issues for the past week.

Primoz

@stephenw10 if you tell me exactly what to look for or where to look for, I can go through the update process again and look at the issue immediately, reboot the router a few times, etc., to maybe catch something in the logs other than the php-fpm.socket spam. I just have to look at it immediatelly, there were about 5 of these messages per second, so I have about a minute to check the logs (at least the way it's like in the webGUI).

Gertjan

@Primoz said in 50x Error on System/Update:

grep link gives (obviously due to the reboot and due to all the php-fpm spam from before):

The other way around.
Because a(n) interface(s) got taken down, processes are restarted (DHCP server, client, DNS, web, PHP, etc etc).
If you can make this up down up down up down etc on an interface (every few seconds ?) stop, you'll see : your system log because silent, everything becomes stable, no more issues.

Your goal as an admin : make sure that :
Your WAN-ISP-upstream device is always powered, and always stays up.
Same thing for all your LAB connected devices, which are typical switches, not wired end-user-devices like PCs or printers, who can go in sleep mode, wake up, sleep again, etc.
If you have electricity issues : get an UPS.

Primoz

@Gertjan this up down was right after a reboot. A reboot to clear an unresponsive (50x error) gui. That's the topic of the discussion, not any of my connection issues, present or not.

I have an UPS and I have the connection part handled. It is not part of the discussion.

stephenw10

If you see a 50X error it's probably because PHP stops responding. So the first things is to try restarting PHP from the console menu.

Primoz

Okay I did the update to 2.9.0 on an otherwise perfectly working 2.8.1 system. The router rebooted twice (it beeps when it does, I didn't have a screen connected) and I was greeted with an error banner which linked me to crash reporter which says:

Crash report begins.  Anonymous machine information:

amd64
16.0-CURRENT
FreeBSD 16.0-CURRENT #12 RELENG_2_9_0-n256132-d8e3138ecf52: Mon Aug 17 18:50:13 UTC 2026     root@pfsense-build-release-amd64-2.eng.atx.netgate.com:/var/jenkins/workspace/pfSense-CE-snapshots-2_9_0-main/obj/amd64/XcuObCIF/var/jenkins/workspace/pfSense-CE-

Crash report details:

No PHP errors found.

No FreeBSD crash data found.

system.log from the moment I started the update to when I logged into the (updated) webUI is:

https://pastebin.com/d5U0JtDc
Hopefully I cleaned up all the URLs, IPs and MAC addresses and other specific data.

stephenw10

Sep  8 20:42:15 pfSense php[27756]: NOTICE The command '/usr/local/etc/rc.d/wireguardd stop' returned exit code '1', the output was 'ld-elf.so.1: Shared object "libxml2.so.2" not found, required by "php_wg"'

Ah, try uninstalling and reinstalling wireguard.

Primoz

@stephenw10 did it, I'll report back if anything goes wrong. I haven't had issues with 2.9.0 with the current update process.