Turris OS 6.0 is now in HBT (Testing branch)

At the end of last month we moved Turris OS 6.0 (based on OpenWrt 21.02) from the HBL to the HBK branch. We are now moving it to HBT for final testing before a stable release.

This only affects users who opted in for testing and switched their branch to HBT.

What’s new?

:boom: Breaking Changes

  • Based on the latest OpenWrt 21.02.4 release
    • LuCI supports L2 and L3 configuration
  • LTS kernel 5.15
  • Foris removed (use reForis exclusively now)
  • Lighttpd uses number prefixes for configuration files to ensure a predictable order

:rocket: New Features

  • Morce (integrated IDS)
  • reForis Nextcloud plugin
  • Wi-Fi 6 cards supported
  • Turris Auth (login gateway)
  • PaKon moved to a separate page
  • New LEDs driver for all routers
  • Dashboard in LuCI
  • NetMetr supports measurement with IPv6 address

:bug: Bug Fixes

  • Fixed Ethernet ports numbering for Turris MOX in reForis

:pushpin: Updates

  • Python version 3.7 ➔ 3.9
  • Lighttpd version ➔ 1.4.67
  • Nextcloud version ➔ 22.2.10

:package: Other Changes

  • Obsolete SIDN package feed dropped

Known issues

High:

Medium/Low:

What’s going to be addressed in upcoming days:

  • Rainbow settings migration is not available yet which can result in losing the LED settings. Fixed in RC2.
13 Likes

Requested package ddns-scripts_no-ip_com that is not available.

Do you plan to provide ddns-scripts package?

Also iputils-ping6 is missing, but the command is now in iputils-ping and ping -6 works as well. (I expect there may be multiple packages with similar kind of changes.)

In my case the Omnia froze during update. I could only get it to answer to ping over wire, but no WiFi, no routing, no reForis. It seems OK after reboot (which lost logs about that, etc.)

What is the default content of the file /etc/config/netmeter?

@OK2SLC you should have it in /etc/config/netmetr-opkg

Just for information, with the self signed certificate WebSocket doesn’t work with Chrome.
To bypass the NET::ERR_CERT_INVALID warning I had to type ‘thisisunsafe’ but after that WebSocket will be “blocked”, I was wondering why Test connection was working.

No problem with Firefox or with Chrome in http. But is not a problem but a behaviour of Chrome.

Edit : I have to same problem with a valid certificate :confused:
Edit2: Correction, it works with a valid certificate and lighttpd correctly configured.
Edit3: It works one time and randomly when launching with : lighttpd -f /etc/lighttpd/lighttpd.conf

Finally very looooooooooooong waited kernel (5) in the horizon :grin:

Why it take ages for major releases don’t know exactly but cz.nic definitely needs more developers. OpenWrt is already developing 23 I think.

1 Like

Yeah mine froze too after autoupdate. No routing, no foris, no luci, no login over sshd, no automatic reboot. Works fine after manual reboot, linux containers (lxc) with devices.allow broke though

Loving this. Thank you devs. Looks great on first impressions. Now digging deeper for bugs etc.

1 Like

Thanks, but file /etc/config/netmetr-opkg is empty except “config settings ‘settings’”. But it doesn’t matter, I have already found it on the forum.

1 Like

On http://192.168.1.1/reforis/network-settings/interfaces WIFI interfaces are located in Unassigned group:

Unable to change - grayed out:

WIFI and network is working fine though. It is only Reforis glitch.

Omnia, 4LED reset+reflash, HBT, went through the Initial Wizard

1 Like

Will it be possible to put here correct /etc/config/rainbow for all new LED naming in rainbow config for Omnia and Turris 1x ?
Even after I reinit it from scratch I am unable to set leds correctly on both routers and in LuCi I see some color names in english and some in Czech. Either way it is not reinitialized correctly from LuCI.

Could you send the diagnostics to our support?

sorry I’m not at home now, but it seems this issue will be easily reproducible, I performed multiple 3LED and 4LED resets and the issue is persisting no matter what I set in LUCI or etc/config/network.

It didn’t happen in my case.

Hi @xsys,

It is indeed visual bug and wifi should be working fine regardless of it, thanks for reporting this!

I agree that it is quite easily reproducible, see the issue in our gitlab. Feel free to add aditional details to the issue, in case we missed something.

I guess that you performed simple upgrade to TOS 6.0 with already configured wifi, right?

Yes, WiFi was configured before upgrade. Though I think I changed some minor setting around the upgrade.

I will add it to known bugs, this was reported here before: WebSocket connection to 'wss://192.168.1.1/foris-ws' failed (#387) · Issues · Turris / reForis / reForis · GitLab

1 Like

Since I updated to TOS 6.0, NetMetr is scheduled every 2hours. Last time that NetMetr was scheduled was in June 2020.
In ReForis Autostart is disable.

344 | time      |    11.04MiB | 2022-10-09 01:05:01 +0200 | Snapshot created by cron
346 | pre       |    11.76MiB | 2022-10-10 16:18:02 +0200 | Automatic pre-update snapshot (TurrisOS 6.0)

cat /mnt/snapshot-@344/etc/cron.d/netmetr 
MAILTO=""
00 	* 	* 	* 	* 	root	netmetr --autostart --dwlhist --rwait 1800 &

cat /mnt/snapshot-@346/etc/cron.d/netmetr 
MAILTO=""
00 	* 	* 	* 	* 	root	netmetr --autostart --dwlhist --syslog &

cat /mnt/snapshot-@344/etc/config/netmetr
config settings 'settings'
	option control_server 'control.netmetr.cz'
	option max_history_logs '10'
	option client 'HW-PROBE'
	option uuid 'xYxxYYxY-xYYx-YxYx-xxYY-xYYxYYxxxYYx'
	list hours_to_run '1'
	list hours_to_run '3'
	list hours_to_run '5'
	list hours_to_run '7'
	list hours_to_run '9'
	list hours_to_run '11'
	list hours_to_run '12'
	list hours_to_run '14'
	list hours_to_run '16'
	list hours_to_run '18'
	list hours_to_run '20'
	list hours_to_run '22'
	list hours_to_run '23'
	option autostart_enabled '0'
	option sync_code 'xYxYYxYxxYxx' 

diff /mnt/snapshot-@344/etc/config/netmetr  /mnt/snapshot-@346/etc/config/netmetr
21a22,23
> 	option autostart_delay '1858'
> 	option protocol_mode 'prefer_6'
1 Like