LXC did not work on boot, I had to manually mkdir /run in order to start any LXC containers, and the the one container which needed to mount /sys/fs/group needed to have it’s setting changed to use cgroup2 instead of cgroup.
Also, although I did not have to change my mapping for usb devices into the container, the /dev tree was populated differently which required some reconfiguration of the app (homeassistant) which had previous saved a direct path to the device serial (which is no long being populated)
Lastly, I had to reconfigure my VLANs to be compatible with DSA.
Other than that, it has been running solidly and I have installed and configured docker natively without any issues.
@Pepe can you or some team member share how to map usb devices in lxc config in debian container for TOS 6 ? I will need it for Turris 1 as well as Omnia for USB Temper device and USB UVC camera.
Thank you
I want to apologize, but we will not be helping you in this case. The package list for LXC available in reForis is labeled for “Advanced users”. It means users who are not afraid to use SSH and configure containers to their needs. Our technical department supports things, which you can configure in reForis. Images for LXC are provided as it is, and we can not support each configuration done there or even provide steps on how to configure it. That would mean that we will support it somehow.
well okay, I understand that. I am not shy of using ssh console and currently using Turris 1 since the beginning and now have Omnia in state of let’s say migration on TOS6 HBD. I noticed that after migration from TOS5 there was need for update of LXC containers confing and USB and other things in LXC just stopped working under TOS6. I tried google how config has to be updated but it did not get me to make it running. Maybe someone else also faced this problem under TOS6 as there is newer version of LXC.
These days, I added more fixups and polished all the commits necessary to have OpenWrt daily snapshots in the HBD branch. My colleague reviewed the pull request approved it, so lets the party begin.
I give try to switch-branch hbl on Turris 1.
First there were not those packages:
inconsistent: Requested package samba36-client that is not available.
inconsistent: Requested package foris-pakon-plugin that is not available.
inconsistent: Requested package foris-netmetr-plugin-l10n-cs that is not available.
inconsistent: Requested package wireguard that is not available.
inconsistent: Requested package libaudiofile that is not available.
inconsistent: Requested package foris-openvpn-plugin-l10n-cs that is not available.
inconsistent: Requested package gst1-mod-alphacolor that is not available.
inconsistent: Requested package fcgi that is not available.
inconsistent: Requested package gst1-mod-deinterlace that is not available.
inconsistent: Requested package libwebcam that is not available.
inconsistent: Requested package libuvc that is not available.
I manually removed them from opkg-auto.lua
then I get an error even though i put repo.turris.cz.ip into hosts:
INFO:Queue removal of tc
Press return to continue, CTRL+C to abort
INFO:Downloading packages
line not found
line not found
line not found
WARN:Crash report of stack trace dump failed.
DIE:
download: Download of https://repo.turris.cz/hbl/turris1x/packages/packages/python3-decimal_3.9.10-3.9-1_powerpc_8540.ipk failed: Couldn’t resolve host ‘repo.turris.cz’
Aborted
turris ~ # ping repo.turris.cz
PING repo.turris.cz (217.31.192.84) 56(84) bytes of data.
64 bytes from repo.turris.cz (217.31.192.84): icmp_req=1 ttl=57 time=9.65 ms
Well, samba3.6 was removed from OpenWrt packages feed in favor of samba4. Foris and its plugins were removed in Turris OS 6.0, and there is only reForis. Regarding wireguard, the standalone package is no longer available as it is part of the Linux kernel.
libwebcam:
libuvc & fcgi:
libaudiofile:
alphacolor and deinterlace were disabled as there are some issues:
Since all of these packages were installed manually, then you need to take care of them yourself. We can also improve the Updater behavior for Foris, wireguard, and perhaps samba.
You should not do it since IP address of repo.turris.cz might change in the future as it happened in the past.
This is interesting. Would it be possible to paste a full log of Updater? Your snippet is so short, and it does not tell as many details what we would need to know what is going on or even in the first place to be able to reproduce it. Also, the diagnostics would be nice to have.
Today, the attempt to switch from HBT 5.3.6 to HBL 6.0 ends in error.
Turris 1.0
Downloading https://repo.turris.cz/hbl/turris1x/packages/turrispackages/Packages.gz
Updated list of available packages in /var/opkg-lists/turrisos_turrispackages
Downloading https://repo.turris.cz/hbl/turris1x/packages/turrispackages/Packages.sig
Signature check passed.
root@turris:~# pkgupdate
INFO:Target Turris OS: 6.0
line not found
line not found
line not found
line not found
line not found
ERROR:
inconsistent: Requested package sentinel-fwlogs that is not available.
And logs:
Mar 11 12:11:29 turris sentinel-dynfw-client[1430]: ipset v7.3: Error in line 1: Cannot open session to kernel.
Mar 11 12:11:29 turris sentinel-dynfw-client[1430]: 2022-03-11 13:11:29,944 - WARNING - Error running ipset command: [Errno 32] Broken pipe.
Mar 11 12:11:47 turris sentinel-dynfw-client[1430]: ipset v7.3: Error in line 1: Cannot open session to kernel.
Mar 11 12:11:47 turris sentinel-dynfw-client[1430]: 2022-03-11 13:11:47,889 - WARNING - Error running ipset command: [Errno 32] Broken pipe.
Mar 11 12:11:48 turris sentinel-dynfw-client[1430]: ipset v7.3: Error in line 1: Cannot open session to kernel.
Mar 11 12:11:48 turris sentinel-dynfw-client[1430]: 2022-03-11 13:11:48,751 - WARNING - Error running ipset command: [Errno 32] Broken pipe.
Mar 11 12:11:52 turris sentinel-dynfw-client[1430]: ipset v7.3: Error in line 1: Cannot open session to kernel.
Mar 11 12:11:52 turris sentinel-dynfw-client[1430]: 2022-03-11 13:11:52,193 - WARNING - Error running ipset command: [Errno 32] Broken pipe.
Any idea why the updater doesn’t start automatically? For a few days now, I have in my logo after the autostart:
turris updater-supervisor: Traceback (most recent call last):
File "/ usr / bin / updater-supervisor", line 33, in <module>
sys.exit (load_entry_point ('svupdater == 1.5.1', 'console_scripts', 'updater-supervisor') ())
File "/usr/lib/python3.9/site-packages/svupdater/__main__.py", line 87, in main
AttributeError: 'NoneType' object has no attribute 'next_window'
also my netmeter starts automatically even if the automatic measurement is switched off.
Is it normal for HBL to update daily even without my permission?
@ackstorm23 are you running docker inside the LXC container only, or you also run docker on TOS 6?
I tried to do so on TOS 6, but I get name resolution errors when trying to build:
WARNING: Retrying (Retry(total=0, connect=None, read=None, redirect=None, status=None)) after connection broken by 'NewConnectionError('<pip._vendor.urllib3.connection.HTTPSConnection object at 0xb55df070>: Failed to establish a new connection: [Errno -3] Temporary failure in name resolution')': /simple/requests/
ERROR: Could not find a version that satisfies the requirement requests (from versions: none)
ERROR: No matching distribution found for requests
The command '/bin/sh -c pip install requests pytz backoff psycopg2-binary' returned a non-zero code: 1