Messages

OK, in that case I'll hold off on it.

Tausend Dank, Franco.

Kind regards,
Franklin

Thanks.  Looks like microcode plugin could be your issue as many others already reported.  Can you remove and retry? . . .

I deleted the microcode plugin and that has indeed resolved the issue, thanks for your help Franco. :)

Do you think it would be safe to reinstall the offending plugin now that the update is complete?

Kind regards,
Franklin

Hi Franco,

I tried upgrading via SSH with opnsense-update -ur 26.7. This time the system rebooted three times then hung. After a hard reboot the dashboard still showed 26.1.11_10 and the same issues were present. There is no upgrade log in the firmware section at all. My SSH session log only captured up to the reboot prompt, so I have nothing useful from the process.

However, after rolling back to my ZFS snapshot I got a crash report, which is odd. I did not think that would be possible after rolling back a snapshot. I guess it must be saved somewhere that is not wiped on a rollback. Either way, no crash report was generated on any of my many attempts at this yesterday, so this is new.

The crash report which seems to contain the update log is here, it is encrypted with a password which I will send you privately.

https://privatebin.net/?5937ca2153055fad#EwVcebaVQ3ZihNc7deSymEoyNJNcANzPmn17Jbhqhrqg

Thanks,
Franklin

OK, I'll try again first thing tomorrow, vielen dank. :)

Kind regards,
Franklin

[Kind regards,]

Thank you for replying so swiftly Franco. :)

To be honest, no, I am not sure if the upgrade completed properly and can only presume. In my frazzled state I neglected to check if the dashboard actually showed 26.7 after the hard reboot.

The upgrade certainly downloads and the web UI console output seems fine during the process. I get the message that it will reboot, it reboots, I get the system beep once the Ethernet ports light up, then a few moments later they go off again, then the port lights come back on and I get another beep. After that the VGA console output seems just the same as always and without errors, but then it just hangs indefinitely and the web ui never comes up. I leave it for 10 minutes just in case and then have to hard reboot and face the multiple issues I previously described.

The Protectli vault is running coreboot. Perhaps this is causing issues with such a major upgrade?

How can I check if the 26.7 core package actually installed correctly? Is there a way to force the core package to reinstall from the console, or to verify the patches you mentioned are in place? I am happy to try the upgrade again and run any commands you suggest before I hard reboot.

I am tempted to create a bootable OPNsense ISO, clean install, and restore my config, but I am very worried about ending up with the same issues and not being able to revert to a previous ZFS snapshot. I suppose I could still create another USB with a previous version of OPNsense and restore my snapshot to that if all else failed, but I really do not want to have to go through all of that.

I am pretty tired now so I will pick this up tomorrow with any suggestions you may have.

Kind regards,
Franklin

[What I have tried so far:]

I have attempted to upgrade to 26.7 approx 7 times and had to roll back each time due to issues.

I upgrade from the web UI. The system reboots as expected, then reboots again, which I assume is normal. After the second reboot the system hangs. Nothing notable appears on the console. I have to hard reboot to get back into the web UI.

After getting back in, I see these error messages even though none of my block rules have a gateway or reply-to set:

Code Select Expand



2026-07-15T14:45:25
Error
firewall
There were error(s) loading the rules: /tmp/rules.debug:330: route-to, reply-to and dup-to are not supported on block rules - The line in question reads [330]: block in log quick on pppoe0 reply-to ( pppoe0 redacted IP ) inet from {any} to {any} label "4b28d971-1a4d-46d1-ae84-900122150176" # block all IN pppoe
2026-07-15T14:45:25
Error
firewall
/usr/local/etc/rc.filter_configure: The command </sbin/pfctl -f '/tmp/rules.debug.old'> returned exit code 1 and the output was "/tmp/rules.debug.old:330: route-to, reply-to and dup-to are not supported on block rules /tmp/rules.debug.old:330: skipping rule due to errors /tmp/rules.debug.old:330: rule expands to no valid combination /tmp/rules.debug.old:336: route-to, reply-to and dup-to are not supported on block rules /tmp/rules.debug.old:336: skipping rule due to errors /tmp/rules.debug.old:336: rule expands to no valid combination /tmp/rules.debug.old:340: route-to, reply-to and dup-to are not supported on block rules /tmp/rules.debug.old:340: skipping rule due to errors /tmp/rules.debug.old:340: rule expands to no valid combination /tmp/rules.debug.old:345: route-to, reply-to and dup-to are not supported on block rules /tmp/rules.debug.old:345: skipping rule due to errors /tmp/rules.debug.old:345: rule expands to no valid combination /tmp/rules.debug.old:395: route-to, reply-to and dup-to are not supported on block rules /tmp/rules.debug.old:395: skipping rule due to errors /tmp/rules.debug.old:395: rule expands to no valid combination pfctl: Syntax error in config file: pf rules not loaded"



What I have tried so far:

• Temporarily changing the affected rules and saving. - Same result.
• Deleting and recreating the affected rules from scratch. - Same result.
• Importing my firewall rules after the upgrade. - Same result.
• Restoring my full backup config after the upgrade. - Same result.
• Checking the "Disable reply-to on WAN rules" option. - This removes the error but other problems remain (see below).

Other problems after the upgrade:

• None of my firewall rule descriptions or labels show up in the live log.
• Firewall rules do not seem to work correctly. Traffic is blocked where it should be allowed.
• I cannot get my WireGuard VPN to reconnect. My VPN routes all my internet traffic.

About my setup:

I use PPPoE for my WAN connection. I run a setup where all traffic is blocked in and out of pppoe because I want my VPN to carry everything, including traffic from OPNsense itself. This means I have to manually intervene if my vpn drops which it does from time to time or when I reboot but this isn't and issue as I work from home and quite like to keep things as secure as possible.

This means whenever I reboot I have to temporarily disable my WireGuard gateway for a few seconds. This allows my "allow WG connect out" rule to reach the external VPN server via pppoe. That rule is a floating rule on the opt6,wan interface group, direction out, UDP port [redacted] to my VPN server at [redacted IP], using the PPPOE gateway. This is a bit clunky but has worked perfectly until now.

Here is the rule:

Code Select Expand



bb60401b-a408-43e3-9ed2-d201ff5a628c  1  keep  6  pass  1  0  opt6,wan  out  inet  UDP  (self)  0  [redacted]  [redacted IP]  0  [redacted]  PPPOE_PPPOE  0  1  0  0  0  ALLOW WG CONNECT FLOAT


Some of my block rules use an interface group (opt6,wan) instead of a single interface. The affected block rules shown in the error also use this group.

I am running this on a protectli vault pro vp4630.

All rules work perfectly on 26.1.11_10 and are fine again once I revert to my previous snapshot.

Please help, I am becoming incredibly frustrated with this.

Kind regards,
Franklin