pfSense webConfigurator and Console crash after login

testcb00

Hi everyone, sorry for my English.

I am now using pfSense 2.5.1 with AMD R3 Pro 4350G + Asrock B450M Pro 4 + 16GB DDR4 ECC UDIMM (8GB x2). The boot device are 2x 16GB Intel Optane NVMe SSD in mirror.

I find that the pfSense webConfigurator and Console will crash after login. The situation is:

I boot my Router, I can login to the webConfigurator and I can control via Console.
After some days, I try to login to the webConfigurator, the Console shows "login success" but the webConfigurator and Console are crash after this login. The Console will not response, I cannot access webConfigurator anymore.

Before the installation, I had installed pfSense 2.5.1 on a USB stick in this PC and I had run this settings over two weeks without any problem. The difference between the new installation and the old installation are:

BIOS Upgraded from 5.00 to 5.10
Boot is in mirror in new Installation
New installation add the pfBlockerNG-devel package

What will be the problem? Hardware? Software? Wrong pfSense Setting?

stephenw10

When you reboot do you see a crash report?

Anything in the logs at the time it stopped responding?

Is it actually completely unresponsive? At the console does it respond to ctl+t ? That can often respond when nothing else does.

Steve

testcb00

@stephenw10

Sorry for my late reply.

May I know how to get the crash report?
After remove the pfblockerNG-devel, the webConfigurator and console still crash under similar condition:
After the removal of pfblockerNG-devel package, I rebooted pfSense on Monday.
I still able to login after reboot. Today I try to login to pfSense and find that the webConfigurator and console crash.

The webConfigurator show
504 Gateway Time-out

However, I find that there is error messages in console (last time there is no error message):

maxproc limit exceeded by uid 0 (pid xxxx): see tuning(7) and login.conf(5)

There are lots of this error message with different pid in console

Ctrl + T show:

load: xxx.xx cmd: sh 63009[runnable] xxxxxx.xxr 0.00u 0.01s 0% 2844k

Also, the system.log which I capture last time.
Router_Bug.txt

I am keeping the pfSense in this condition as it takes time to reproduce the condition.
Waiting for reply and suggestion.

Gertjan

@testcb00

The log you showed is rather clean.

With the famous exception : You have a realtek based NIC - and that means "troubles".
I saw lines like :

Jun  7 11:42:04 pfSense check_reload_status[438]: Linkup starting re0
Jun  7 11:42:04 pfSense kernel: done.
Jun  7 11:42:04 pfSense kernel: re0: link state changed to DOWN

The Realtek NIC comes up, and goes down a split second later.
( and this probably several times, until the system goes brain dead ).
The solution is easy : do not use realtek NIC's (called "re0" in your log) - change it for some other NIC, 'intel' based is preferred.

testcb00

@gertjan
Noted with thanks.
I will reboot and disable the Realtek NIC.
Hope there are no more problem.

stephenw10

If you're actually hitting the max processes limit something is amiss.
If you are though you should see it in Status > Monitoring. Do you see a slow rise in the processes number or a spike just before it crashes?

Steve