Turris OS 3.11 is out!

This is a bug in the ddns-scripts package.
We or somebody else needs to dig this bug and look what’s happening inside the ddns-scripts and fix it. I think that we won’t update it unless we have to because of a security issue if it would be found. Hopefully not.
Updater guy (@cynerd ) did some improvements for Updater and prepared pull request for review. My rough estimate is that it can be included in Turris OS 3.12. You can see his response here about this issue.

Unfortunately, from this message, we don’t see what can be wrong. Can you check system logs or Nextcloud logs, if there is anything that will tell us more?

I’m keep getting notifications:

Updater selhal:

inconsistent: Requested package foris-tls-plugin that is not available.

I see we have removed foris-tls-plugin for Turris OS 3.11. It was a Foris plugin, which provides a frontend for NUCI TLS plugin, by small steps, we’re trying to remove nuci as much as we can, because it is just an old project, which we deprecated and thus we don’t provide support for it. NUCI (Networked UCI) is a daemon for accessing OpenWRT configuration through a network with the netconf protocol.

Is there any reason, why would you like to have it installed? I think you might have it installed by hand. So, you can remove it with the following command. Once you do that you can run Updater again.

opkg remove foris-tls-plugin

//EDIT: @freshdax: If it was uninstalled, then everything is fine. :slight_smile: No need to worry about it.

    root@turris:/tmp/log# grep -i err messages 
    2018-12-11 18:32:00 err kresd[2761]: [priming] cannot resolve '.' NS, next priming query in 10 seconds
    2018-12-11 18:32:00 err block[]: /dev/sda is already mounted
    2018-12-11 18:32:01 err block[]: /dev/sdb1 is already mounted
    2018-12-11 18:32:01 err block[]: /dev/sdc1 is already mounted
    2018-12-11 18:32:01 err block[]: /dev/sdd1 is already mounted
    2018-12-11 18:32:05 err ntpdate[4407]: name server cannot be used: Try again (-3)
    2018-12-11 18:32:47 err foris-controller[2965]: WARNING:turrishw:unsupported TOS version (on omnia): 3
    2018-12-11 18:32:54 err foris-controller[2723]: Last message 'WARNING:turrishw:uns' repeated 1 times, suppressed by syslog-ng on turris
    2018-12-11 18:33:12 err foris-controller[2965]: WARNING:turrishw:unsupported TOS version (on omnia): 3
    2018-12-11 18:33:20 err foris-controller[2723]: Last message 'WARNING:turrishw:uns' repeated 1 times, suppressed by syslog-ng on turris
    2018-12-11 18:36:18 info updater-supervisor[]: pkgupdate reported no errors
    2018-12-11 19:00:02 err server_uplink[]: Failed to download contract status
    2018-12-11 20:00:01 err server_uplink[]: Failed to download contract status
    2018-12-11 20:11:03 info updater-supervisor[]: pkgupdate reported no errors
    2018-12-11 21:00:01 err server_uplink[]: Failed to download contract status
    2018-12-11 22:00:01 err server_uplink[]: Failed to download contract status
    2018-12-11 22:01:01 err server_uplink[]: Failed to get registration code
    root@turris:/tmp/log# grep -i err /srv/www/nextcloud/data/nextcloud.log
    {"reqId":"AAn4o6h6QBYKIEco4lhu","level":3,"time":"2018-12-11T21:18:02+00:00","remoteAddr":"192.168.1.20","user":"--","app":"index","method":"GET","url":"\/nextcloud\/","message":{"Exception":"Error","Message":"Class 'OCP\\Config' not found"

Besides these error messages i haven’t found anything, yet. What should I look for?

This is another TERRIBLE update.

Thanks, dude. I can understand that in your case it didn’t end up well, but please, let’s avoid words like shit.
I know that it gave you some temporary relief, but in the end, it doesn’t help anyone.

Would not be better, if together we can focus and improve things, which can push us forward? So we can make everyone happy including you.
You as a part of the community can help us to catch bugs or issues, which appears. So, we can fix them. I would lie to myself if I said there would never be an issue or bug. There might be always something, so all of us can improve the next release as we can not test all setups, which are possible. That’s something that we’d like to deal with. Most users won’t notice it.

Unfortunately, Turris OS 3.11 brings a lot of changes as it is a major release. It was in RC for a month, you can check it here: Turris OS 3.11 in RC! It looked like that we could release it. However, not many users were in RC as you can see. I see it’s something, what can be improved. Should we at least give active users, who helped us T-shirts or any other swag?

Anyway, this release was a good one except for ddns-scripts issue even there is always what can be improved as most of the issues was caused by upgrading busybox because it provides ntpd. That’s what can be solved in a few seconds or minutes. I don’t know if it helps you, but this happens in the latest snapshot of OpenWRT.

If you would like to compare some numbers here or on Support (tech.support@turris.cz), it’s really small numbers comparing it against routers, which are running. Would you provide positive feedback about things or products, which you like to manufacturer or support? I doubt it. Most of the feedback you can see here or somewhere else it’s from users, who are not satisfied or they found some issue.

Let’s talk about your issues.
The issue, which you have mentioned is related to your internet connection and I doubt that’s what we can change if you look at the error message what it says - Operation timed out. If the restart solves it, then great.

ERROR: unreachable: https://repo.turris.cz/omnia/packages/base/kernel_4.4.165-1-4a7a81f8db0ad743e54c68e1845c60b6-1_mvebu.ipk: Operation timed out after 55209 milliseconds with 491250 out of 2807478 bytes received

About Updater and its hang, I have already answered it here:

Just my two cents, I don’t recommend to disable automatic updates completely as you can use updates with approval or delayed updates, because, in almost every version of Turris OS, we’re upgrading packages, which had security issues.

1 Like

I don’t remember installing it, so probably I don’t needed. I’ve uninstalled it.
Thank you!

1 Like

Did you find out out to update the ntp servers? Or enable ntp server in router?

The Wifi adapter 2.4 Ghz is slower than before. To connect 30/30 Mbps reach after the update speed between 7-15 Mbps. The 5 Ghz network is without speed reduction, previously both WiFi networks were at a speed of comparable or identical.

Both channel 2.4 and broadcast performance are the same as before update

We haven’t update ath9k, it might seem for you like it is caused by the update, but we didn’t do anything with it. There are many access points on 2.4 GHz wireless, so there is a lot of interference and not many spaces for another ap in the same frequency. I’d suggest changing the channel of the radio for 2.4 GHz and see if it helps.

As with the last update, i had to stop and disable /etc/init.d/ddns for the update to run through…

@alump I did figure out how to get DNS over TLS to Cloudflare (or other supported providers) working again. It seems that 3.11 added initial support for Foris configuration of DNS over TLS, but in a way that the previous manual configurations no longer work. It’s not clear why this wasn’t mentioned in the 3.11 release. Findings updated at gitlab here

2 Likes

Could you please edit community page https://doc.turris.cz/doc/en/public/dns_knot_misc to be up to date? It prevent future errors of potentionall new users…

@Jirka done, please check the page https://doc.turris.cz/doc/en/public/dns_knot_misc#using_dns_over_tls

1 Like

2 posts were merged into an existing topic: Mdadm 4.0 monitor mode doesn’t send email notifications

Great! I cant check new method, as I dont have 3.11 yet, but its important that other users will not be confused! :+1:

Sorry about the late reply for Pakon, yesterday I spoke with @mpetracek.
It should be just a temporary issue, rebooting of the router should fix it.

My updater also hangs forever. My updater hangs at
DEBUG:backend.lua:1009 (script_run):Running postinst of firewall
DEBUG:src/lib/interpreter.c:320 (lua_run_generic):Command: /usr/lib/opkg/info//firewall.postinst configure
TRACE:src/lib/events.c:542 (run_command_a):Running command /usr/lib/opkg/info//firewall.postinst

I have started support ticket and waiting for the reply.

Seems the updater got stuck. Turris already reports version 3.11, but no notification has been sent out.
Also the updater is apparently still running.

~# ps | grep update
18191 root     10396 S    python /usr/bin/updater-supervisor -d --rand-sleep
28296 root      302m S    pkgupdate --batch --state-log --task-log=/usr/share/updater/updater-log --reexec

When I’ve tried to run it again it died

~# pkgupdate 
INFO:Detected existing journal. Trying to recover it.
line not found
line not found
line not found
line not found
DIE:
[string "backend"]:1194: [string "backend"]:1185: Failed to lock the lock file //var/lock/opkg.lock: Resource temporarily unavailable
Aborted

Nothing special in the log

2018-12-10 16:07	TRANSACTION START
2018-12-10 16:07	install	20180409-2	ca-bundle
2018-12-10 16:07	install	7.62.0-2	libcurl
2018-12-10 16:07	install	60.4.6-3.6-2	updater-ng
2018-12-10 16:08	TRANSACTION START
2018-12-10 16:08	install	1.3-2	l10n_supported
...
2018-12-10 16:08	remove	-	python-jsonschema
2018-12-10 16:08	remove	-	python-functools32
2018-12-10 16:08	remove	-	updater-ng-py2-supervisor
2018-12-10 16:08	remove	-	python-uci
2018-12-10 16:08	remove	-	lcollect

Update: So I killed the process and run it again… It run for a while and got stuck again. This time with notification:

Error from 2018/12/12 17:11:48

Updater selhal:

[string “backend”]:1194: [string “backend”]:1185: Failed to lock the lock file //var/lock/opkg.lock: Resource temporarily unavailable