Turris OS 5.0.0 is going to be in HBK

I mean I tried first to update from the Turris OS 3.0, but that was a dead end.
As I do not have any valuable data despite of the configuration itself on the turris, anything else stays on an attached external ssd (LXC container with ArchLinux).

So, I used the variant ā€œRe-flash routerā€ described in here: https://docs.turris.cz/hw/omnia/rescue_modes/
Instead of using the HBS, I first tried with the OS 5.0 HBK medkit (https://repo.turris.cz/hbk/medkit/omnia-medkit-latest.tar.gz). But sadly that does not work. Currently, I do not have a UART adapter, otherwise I would have digged into why the medkit of HBK does not work for the Turris Omnia.
Consequently, I used now the OS 4.0 HBT medkit (https://repo.turris.cz/hbt/medkit/omnia-medkit-latest.tar.gz), which worked for my Turris Omnia. As soon as 4.0 was installed, I sshed into the router and switched to OS 5.0 with:

switch-branch hbk

Result:
Bildschirmfoto vom 2020-03-14 20-17-00

Thanks to @Pepe for the kmod-tun links. Will try it out.
EDIT: after a ā€œopkg updateā€, opkg was able to find the kmod-tun

Hi,
I have switched to hbk version recently.

Model:			Turris Omnia
CPU:			ARMv7 Processor rev 1 (v7l)
Release:		TurrisOS
Release Revision:	74a8e36
Release Codename:
Release Target:		mvebu/cortexa9
Release Description:	TurrisOS 5.0.0 74a8e36975a7148b4a901b849facbb6fa309ea97
Turris OS Version:	5.0.0
Kernel:			4.14.172

I am having samba shared folder, nextcloud, adblock, otherwise standard setup. Now I am facing two small issues and one bigger one:

  1. Foris Storage shows message:Your setup is currently broken and you are probably loosing data, set a new storage device as soon as you can! however my SSD disk is correctly mounted at least I can see it in the terminal:
$ df
Filesystem           1K-blocks      Used Available Use% Mounted on
/dev/mmcblk0p1         7633920   2313552   5331636  30% /
devtmpfs                   512         0       512   0% /dev
tmpfs                      512         0       512   0% /sys/fs/cgroup
tmpfs                  1033996     26164   1007832   3% /tmp
tmpfs                      512         0       512   0% /dev
/dev/sda1             26214400   2475288  21736936  10% /srv
/dev/sda3            946352836 479417180 464312644  51% /mnt/nas
  1. The netmetr does not work, or to be more precise, it measures the speed but could not report it, I am showing only last part of the netmetr --debug command:
Traceback (most recent call last):
  File "/usr/bin/netmetr", line 11, in <module>
    load_entry_point('netmetr==1.5.3', 'console_scripts', 'netmetr')()
  File "/usr/lib/python3.7/site-packages/netmetr/__main__.py", line 622, in main
  File "/usr/lib/python3.7/site-packages/netmetr/__main__.py", line 327, in upload_result
  File "/usr/lib/python3.7/site-packages/netmetr/__main__.py", line 94, in send_request
  File "/usr/lib/python3.7/urllib/request.py", line 222, in urlopen
  File "/usr/lib/python3.7/urllib/request.py", line 531, in open
  File "/usr/lib/python3.7/urllib/request.py", line 641, in http_response
  File "/usr/lib/python3.7/urllib/request.py", line 569, in error
  File "/usr/lib/python3.7/urllib/request.py", line 503, in _call_chain
  File "/usr/lib/python3.7/urllib/request.py", line 649, in http_error_default
urllib.error.HTTPError: HTTP Error 400: Bad Request
  1. Bigger one: I tried to pair (using netboot) my MOX classic. The pairing is successful. I can see the MOX in Foris or in CLI:
$ netboot-manager list
Accepted:
 * xxxxxxxxxxx

However in the subtabs in Foris (setup, wifi) I only see exclamation mark and do not see any devices nor channels in wifi subtab. Only spinning wheels.

Thanks for suggestions, Robert

I have one of the original Turris Omnias. Last I tried to get HBK on it I was getting i2c errors. No idea whatā€™s wrong.

One more question, crontab seems to be broken now. the cron jobs are not executed. I can see scheduled jobs in the LUCI but could access it via CLI. What is the way now?

$ crontab -e
no crontab for root - using an empty one
/bin/sh: /usr/bin/vi: not found
crontab: "/usr/bin/vi" exited with status 127

Thanks for help. Robert

Hello,

Thank you for trying our Here Be Kittens (hbk) :cat: branch!

Regarding the first error in Foris about Storage tab:

We have been able to reproduce it. Can you please share output of ls -lh /bin/stat?
Would you please try to run update_alternatives.sh on your router to see if it helps to you? It should be installed on your router.

Unfortunately, I am not able to reproduce your issue on Turris 1.1 and Turris Omnia in HBK and HBL branch. For example, Turris 1.1 HBK:
obrazek.
Can you please try it again to see if it happens to you now?

Can you please tell us more details about your configuration and send us diagnostics, which provide us insights to logs and provide us ways to help you.

Sorry for any trouble caused by this. The fix is on review, but it requires additional tweaks, which will be done in upcoming days. You can see the status of it in the merge request, which was shared here or it will be fixed once you see updated package cronie.

1 Like

Hi @neheb,

Nice to see you here! Would you please share more details? This will help us to know what is happening in your case. I am not aware of any errors about i2c protocol.

Itā€™s in production right now, so I canā€™t do much testing. However I do remember having a hell of a time getting this flashed back to stock (had a custom build of OpenWrt that ended up bricking it), It involved multiple medkit files as well as restoring U-Boot vars back to stock (env default -a).

In any case, I stopped using TOS because of https://gitlab.labs.nic.cz/turris/turris-os-packages/issues/398 (should not be a problem with 19.07). I donā€™t think Iā€™ll be using nextcloud this time around though.

Dear @Pepe,

here is the output of

$ ls -lh /bin/stat
lrwxrwxrwx    1 root     root           7 Mar 19 08:15 /bin/stat -> busybox

After running update_alternatives.sh the problem seems to be fixed. and the output of the command above changed:

$ ls -lh /bin/stat
lrwxrwxrwx    1 root     root          17 Mar 19 09:03 /bin/stat -> /usr/bin/gnu-stat

Regarding the netmetr I have got an information that the netmetr server itself has some issues with IPv6 which I am actually using, so I am waiting for the fix.

For the netboot of MOX, I am going to send you the diagnostics report to tech.support @ turris.cz.
Thanks for your help,
Best
Robert

The (in my opinion) easiest way is to flash the new image to the router with a usb stick. @lucenera has written a good step-by-step guide to complete the upgrade. Help upgrading to turris 4.0.3

Donā€™t forget to make a backup of your configuration first if you donā€™t want to start from scratch. After the upgrade, the switch-branch command is available.

Two days ago I accidentally upgraded from 3.x to 5.x (hbk) and then tried to downgrade with switch-branch hbt to 4.x but the command failed. Unfortunately, since hbk is a little bit too risky for me at the moment with my home office working situation, I flashed 4.x back.

Anyways, maybe if one of you have time to look into this issue, that a downgrade from 5.0.0 is not possible at the moment.

Hello,

May I ask you to execute these commands again and could you send us the output in the ticket? The e-mail address is: tech.support@turris.cz. We will look at this with my colleagues.

Thanks in advance.

Hello @Hashlog,

We have the guide for updating the router in our documentation. It is detailed and official, so you can use this one.

https://docs.turris.cz/hw/omnia/rescue_modes/
https://docs.turris.cz/geek/testing/

INFO:Queue removal of luci-i18n-mjpg-streamer-de
INFO:Queue removal of luci-i18n-tinyproxy-de
Press return to continue, CTRL+C to abort

INFO:Downloading packages
ERROR:(https://repo.turris.cz/hbk/turris1x/packages/packages/collectd-mod-memory_5.10.0-3_powerpc_8540.ipk) Opening output file ā€œ/usr/share/updater/download/collectd-mod-memory-5.10.0-3.2.ipkā€ failed: No file descriptors available
line not found
line not found
line not found
WARN:Crash report of stack trace dump failed.
DIE:
[string ā€œupdaterā€]:114: Error while registering for download: https://repo.turris.cz/hbk/turris1x/packages/packages/collectd-mod-memory_5.10.0-3_powerpc_8540.ipk: Unable to open output file for writing
Aborted
root@turris:~#

In our documentation How to try future releases of Turris OS and also in switch-branch utility, you can see that you are using branch Here be Kittens. It is suggested for experienced users as from to time, there might be occasional bugs, but you should be able to recover from it.

If anything happens, we need detailed bug reports, which is reported by proper way.

There are no environment details in your post and there is just few outputs. We would appreciate more details about it. How we can reproduce it, how it happened?

@viktor I am fully aware of all that. If you read my previous post carefully, it says, that I had to switch back to a more stable branch (hbt), because I need my internet connectivity at the moment. I still wanted to inform you, if maybe somebody else could look into that matter. If I will have time and an acceptable downtime slot, I will try to reproduce the error.

One more think related to update: when updating ddns there is an error with pre-installation scripts:

DIE:Failed to exec /usr/lib/opkg/info//ddns-scripts.preinst: Exec format error
DIE:Failed to exec /usr/lib/opkg/info//ddns-scripts_route53-v1.preinst: Exec format error

Thanks for reply. It was respond for @Twinkie in response to:

There are not much steps to reproduce as I had been on HBL with a just bunch of packages that I had on TOS3 installed via opkg install, Then as this news that HBK is finally TOS5 came out I just give ita try and switch-branch to hbk and run pkgupdate as it was stated here. I will try to install HBK medkid and then do factory rolllback with HBK directly as that would probably work. Both branches are highly experimental and even my experiments to migrate was what it is ā€¦ eperiments. So sorry that I did not posted much details. As i experimented with TOS5 i run into the situation when I had no other option than refresh SD card of TOS3 from image but it was my fault as I tried install hundreads packages together then I came with more clever approach to install first missing kmods then libs a and after that ifnally packages. What is now bigger problem for me that i had compiled openwrt toolchain for powepcspe and compiled some my packages that are not available on openwrt by default only to realize that I can not install them on TOS5 due to different versions of libraries as my toolchain was compiled against master so different versions of openssl is to blame and also I see much bigger problem that I was unable to start LXC container of pihole and debian that I used to run on TOS3 even after config migration under branches of TOS5. My post related to this had been so far unanswered so I donā€™t know if it is possible to run LXC under TOS5 of blue Turris 1 or not or what has to be done in order to make it running. I have all LXC packages installed but no luck. Not sure if create new container will solve the situation. If so I am able to migrate debian packages and config from old container to new one even it will be much more work than just copy whole root fs of LXC to new TOS5 and start it. I would like to persuade more people to help with testing of TOS5 as from my point of view if more and more people try to go on it it could help to test TOS5 to team.

Hi, i buy omnia couple of weeks ago and switch to hbk, it works flawlessly.
One think i would like to know where i can find changelogs to all updated packages ? They are uptaded daily, I tried to google it but not found it anywhere.