Messages

@MK4L just an outside observer on this; seems like the distilled version would be a great add to "Tutorials and FAQ"

as a data point, I see this pretty regularly (on 25.7). Probably a couple times a week, especially when
- accessing a system across a large geographic distance (cross-ocean)
- a table that has a high rate of change (e.g. firewall live view on public/wan interface)

I'll keep an eye on this once I update to 26.7 (when first point release is available)

This was a timely nugget of information. 

I'm trying to reproduce a multi-site wireguard site-to-site issue and am using VMs to mimic the environments. My VM template disk was too small, but with `touch /.probe.for.growfs` I was back up and running in minutes...

And also a big thanks to @Maurice for the aarch64 images!

[C-a d]

So the issue went away after the connected host was updated. Thought it was gone for good, but.... It re-occurred again this morning.

So I have figured out what's causing it; 100% repro.  Not sure if on the host side or opnsense side, but here's the repro and work-around (debian host, opnsense 15.7.10)

using

Code Select Expand

screen /dev/ttyUSB0 115200When exiting:
- C-a d (detach) - when reattaching the console is corrupted and has the character loss issue
- C-a K (kill) - the console exits cleanly and can be reconnected to without any issues

Recovery
1. reboot host :(
2. Reset host port using

Code Select Expand

echo "<devinfo>" | sudo tee /sys/bus/usb/drivers/usb/unbind
sleep 1
echo "<devinfo>" | sudo tee /sys/bus/usb/drivers/usb/bind
The host port devinfo can be retrieved using:

Code Select Expand

udevadm info --name=/dev/ttyUSB0 | grep DEVPATH and using the last numbers before the `:`
i.e. if udevadm returns
DEVPATH=/devices/pci0000:00/0000:00:01.1/0000:01:00.1/usb1/1-2/1-2.1/1-2.1:1.0/ttyUSB<X>/tty/ttyUSB0
use 1-2.1

Good data point. There must be something site specific, but I haven't been able to figure it out.

EDIT: the tunable that needed to be set then was hint.uart.0.at="isa"

I've tried the tunable. No change in behavior, so I've reverted it.

I tried another terminal emulator, picocom. With it I see more of the missing characters, but still have a significant loss.

Protectli support recommended the opencore bios update as well. Will update the bios over the weekend.

After that, I'll try:
- replacing the protectli usb-c/usb-a cable with a different usb-c/usb-a cable
- use a different host to for the console connection

The install is remote, so a bit of delay for anything needing physical access. (and the need for a reliable console connection...)

As always, welcome other ideas to help troubleshoot

I'm having a strange issue when trying to access the console on a protecli V1211 (2 port firewall with corebios) running Opnsense 25.7.10

I'm able to connect from a linux host using `screen /dev/tty<name> 115200` However characters are randomly dropped from the output and many times not echoed when typed. Boot messages are garbled and only parts of the opnsense-shell are shown.  I am able to type `8` to get to the shell and run commands (with the same dropped character issues).

My current console settings are in the screencap.  An example of the corrupted output is:

```
# ls -al
30 08:08 .history
-rw-------   1 root wheel   20 Dec 29 03:03 .wheel 1174 Dec 17 02:56 .shrc
drwx------   3 root wheel    4 De
```

and partial/truncated opnsense-shell default menu

```
FreeBSD/amd64 (fw) (ttyu0)
 :::        :::
|  Code:        https://github.com/opnsense    root@fw:~ #
```

I've tried the sc console driver, tried the usb setting, etc. Nothing seems to change the behavior.

Any ideas/thoughts on next steps to troubleshoot this?

Observations:
- I don't remember seeing the dropped character issues on the factory installed 25.7 version. I booted, confirmed connectivity, then applied 25.7.10. My next console access had the issue.
- I also had this on the VP6650 with ami bios. I never was able to get usb-c console working. I was able to use the rj45 console instead, but the V1211 does not have one.


- I've also opened a ticket with Protectli. Will update here if any progress on that front.