Loss of connecitivity & 1GB file appears /arp.core

[5SpeedFun]

Any idea what I can do to troubleshoot this?  Just started happening after the latest update.  Has happened twice in 4 hours now.  Not sure what is going on.

I wonder if it's related to this?
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=242784

I also see in dmesg.today:

Code Select Expand

pid 99124 (arp), jid 0, uid 0: exited on signal 11 (core dumped)
pid 99531 (arp), jid 0, uid 0: exited on signal 11 (core dumped)

Restarting all services from the console doesn't fix it, but  reboot does.

My OPNSense version:

Code Select Expand

OPNsense 21.1.2-amd64
FreeBSD 12.1-RELEASE-p13-HBSD
OpenSSL 1.1.1j 16 Feb 2021

[crayon]

It's certainly possible; the patch referenced in the bug report has not been merged to OPNsense yet.

Could you open an issue at https://github.com/opnsense/src with the information you provided here?

[5SpeedFun]

I've opened an issue (https://github.com/opnsense/src/issues/109).  Hopefully it helps.

[franco]

Will be fixed in 21.1.3 this week, thanks a lot!


Cheers,
Franco

[5SpeedFun]

Thank you @Franco!

I am looking forward to see if 21.1.3 fixes the issue.  In the meantime I backed up my 21.1.2 config, restored an os backup of 21.1, reset config to defaults, and restored my 21.1.2 config to 21.1.1  I'm curious to see if the issue exists in the old OS with my newer config.

Yesterday I had 6 firewall crashes, so it was roll back to an pre-21.1.2 backup or try maybe the opnsense-patch method (which I only found out about after I did the restore). 

[5SpeedFun]

So it took about 10 hours, but it crashed again on 21.1.1.

To test the patch, is it best to upgrade to 21.1.2 and then run

Code Select Expand

opnsense-patch -p 3ea5beda50473a3071f9066610f94eee2408d771

and then reboot?

[5SpeedFun]

so I've lost connectivity twice in 20 minutes now....so looking for another solution.

This seems to fix the problem without a reboot

Code Select Expand

/usr/sbin/arp -d -a

[franco]

opnsense-patch only works on core.git and plugins.git mostly, and never for things that require compilation like the OS does :)

You can try the 21.1.3 sets...

# opnsense-update -bkr 21.1.3
# opnesnse-shell reboot

Edit: sorry, silly moment here. You need the base changes since arp is a program in userspace... Adjusted command accordingly.


Cheers,
Franco

[5SpeedFun]

Thanks @franco!  I did the update & it went ok except for named hanging.  Opening a new shell & doing kill -quit let it reboot.

The main UI says I'm still on 21.1.2 but if I go to system -> firmware -> update packages, it lists me as on 21.1.3

I'll let you know if I see anymore of the arp based crashes.  Thanks again for fixing this!

[5SpeedFun]

0 crashes in 12 hours.  Things looking much better!