Messages

I'm running a mini-PC headless, as a result if  I get any errors during an update I can't see them. Given the location of the PC it's not really feasible to attach a monitor to it in situ so I need to power down the PC & move it before hooking up a monitor, thus loosing any messages that were displayed.   I have set up a RS232-Ethernet server so I can view the terminal output of the comport com terminal remotely.

However, I have a quick question, when I open a terminal, I see the same screen as if I attach a monitor directly to the Opnsense PC which shows basic LAN & WAN parameters then the 0 – 13 menu options, including 8) Shell etc.

My question is am I better laving the terminal on the first screen showing the menu options while performing the update (via the GUI) or going into a shell session? I'm not sure which session will show the most useful information & any potential error messages.

My thoughts are leave it on the first screen without going into the shell but would welcome any advice.

My goal in this setup is to be able to see any errors or issues or even re-boot in the event that the web interface and SSH are unavailable.

After a few updates I should be able to figure this out myself and will post back my results but any guidance welcome. 

PS I have considered getting ethernet – KVM but haven't been able to find anything reliable at a reasonable price. Once the JetKVM is available I may get one of those but they aren't available yet so this is a temporary solution. 

Since upgrading to 25.7 I feel that the browsing experience has improved. I don't have any hard and fast metrics and speed tests are unchanged it was maxing my line out anyway. However general browsing on all my devices, Win 11 desktop, iPad & Mac Book all feel faster so not sure if Unbound is resolving addresses faster or if it's just my imagination but the general browsing experience feels sharper some how, has any one else noticed an improvement?

Same here patch worked perfectly.

Fantastic verry quick response, a big tank you for all your hard work it really is appreciated.

+1

Uninstalled the os-cpu-microcode-intel microcode & rebooted to make sure it took effect.

Ran the upgrade to 25.7 device hung, power cycled it and it came up OK so removing the microcode did improve the situation. Unfortunately I'm running this router headless so could not see any errors on screen - I have to take it out of the rack to attach a monitor to it.

I'm going to continue running it without the microcode installed and see how it goes. Following on from Franco said earlier I will look it there is a BIOS update available, it's a small fan less PC so may need to do some research to find it.

PS big thank you to Franco for you extremely rapid and helpful suggestion, I love this community.

sorry should have said it is intel

"Intel(R) Celeron(R) J4125 CPU @ 2.00GHz (4 cores, 4 threads)"

I think I do I have this

"os-cpu-microcode-intel (installed)" listed under plugins is that the one you are referring to. I can't remember why I installed it?

Do you suggest disabling / uninstalling it and trying the update again?

Upgrading from 25.1.12 (system running fine since install ing 25.1 earlier this year) to 25.7

GUI appeared to download all packages and install everything without any errors as far as could tell & system went on to reboot but never came back up.

Hooked up a monitor and rebooted to get the following


(the following is manually typed from a photograph of the screen using mobile phone so there could be some typo's in there but I think it's correct.)


>>> Invoking import script 'importer'
Configuring crash dump device: /dev/ada0p3
swapon: adding /var/ada0p3 as swap device
.ELF ldconfig path /lib /usr/lib /usr/lib/compact /usr/local/lib /usr/localpsec /usr/local/lib/perl5/5.40 mach/core
32-bit compatability ldconfig path:
done.
chown: /var/run/booting: No such file or directory

>>> Invoking early script 'upgrade'
>>> Invoking early script 'configd'
Generating configuration configd socket missing <@var/run/configd.socket>
>>> Error in early script '15-templates'
>>> Invoking early script 'backup'
>>> Invoking backup script 'captiverportal'
>>> Invoking backup script 'dhcplease'
>>> Invoking backup script 'duid'
>>> Invoking backup script 'netflow'
>>> Invoking backup script 'rrd'
>>> Invoking backup script 'carp'
CARP event system: configd socket missing <@var/run/configd.socket>
>>> Error in early script '90-carp'
Launching init sysm...done.
Initalizing................done.

(system hangs at this point. Hitting Ctrl C gives)



Enter full pathname of shell or RETURN for /bin/sh:
root@:/ #

power cycling still gave the same issue.

As a result I have used "bectl activate snapshot_name" to revert to a previous snapshot and everything is running fine.

Can anyone shed any light on what the problem may be  please?

Following up on my own post because I have solved the issue and this may help someone else.

It turns out clock was about 25 seconds slow the Windows 11 PC I was viewing the dashboard on. I have no idea why the clock was wrong as windows was configured to set the time automatically and always was set that way.

Anyway set it to manual then straight back to automatic and the time immediately corrected it's self which resulted in the dashboard working perfectly.

So if anyone else notices the traffic graphs on the dashboard taking ages to load and then lagging check the time on your computer vs the Opnsense box.  Just to clarify it was the Windows box running slow Opnsense box clock was correct (and also set to automatically)

Not a fault with Opnsense as such but just can't get my mind around it. Running latest version 24-7-8 and recently noticed that on my main Win11 machine the Traffic Graph on the Dashboard takes ages to populate. Didn't notice this before but recently verry noticeable the rest of dashboard loads quickly as expected over a 1Gig LAN connection.

As a result I have tried various devices and browsers - as a bench mark the under reporting the traffic graphs start to populate withing 3 seconds of clicking the reporting->traffic (on all devices and browsers) so in effect immediately as the 3s includes drawing all the screen from scratch.

Using various pieces of kit all hard wired except the iPad & Mac Book with the following results on the Lobby Dashboard Traffic Graph

No VPN running ipv4 & 6 dual stack with two interfaces LAN & WAN

Win11
Opera 26s
Edge 26s

Win10
Opera 8s
Edge 8s
Chrome 8s

iPad
Safari 4s
Chrome 4s

Mac Book Air
Safari 4s
Chrome 4s

Given that everything works as expected on the two apple devices I assume it's some sort of windows issue rather than an Opnsense issue but given the CPU widget is so responsive on all devices I just don't get it.


It's so noticeable in that the the CPU graph is drawn immediately and has plotted all the way across the widget before the traffic even start to populate at all.

If the data rate WAN or LAN changes before the graph is populated with the trace the scale of the graphs changes as though it's adjusting the scale to match the data so it seems to know there is data there it's just not drawing it on the screen.

Both windows PC's and browsers work fine on every other part of Opnsense with no noticeable delay both work fine on locally hosted web pages, local speed tests and remote speed tests.

What I can't understand why the there is only a delay on the Dashboard Traffic Graph.

If anyone can explain what appears to me to be extremely strange behaviour or has any solutions I would be eternally grateful or of anyone else is seeing this please let me know.

Yes it's a shame there isn't a way to select which ones to show or conversely select which ones to hide via some check box or something. Not sure of the correct way to suggest this if anyone knows of the appropriate way to request a feature pleas point me in that direction. It's obviously a minor enhancement but it would be cool to add this feature.

I'm running 24.7.2 and the first graph on the Packets and traffic tabs is Ipsec. I'm not using Ipsec so graphs show zero no surprise there. However is there a way to make the Ipsec graphs dissapear?

I have read in the forum that if I delete the rrd files ipsec-packets.rrd & ipsec-traffic.rrd which are stored in the /var/db/rrd/ directory, they don't show on the relevent tabs so I did this and haypresto they dissapeared from the relevent tabs leaving the graphs for WAN & LAN which is what I want.

Unfortunatly they re-appeared several days ago.

Is there a way to make them disappear perminantly?

I was running 24.7.1 at the time which was an upgrade from 24.1.10 so not sure if the upgrade was an issue I did have issues with the health graphs that worked fine and occasionaly stopped updating, resetting RRD data brought them back to life. This is what promted me to do a fresh install of 24.7 I havn't tried deleting the rrd files since the fresh install.