My Turris Omnis died today during operation.
Total network failure.
It stopped routing and does not boot any more. All lights remain blue.
All cables etc. removed, for no avail.
What I found:
Email, today, 16:45:
News announcements
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
…
I have experienced a device-bricking update in HBL branch about a week ago (it was just in the middle of me being overseas, so, I didn’t really have a chance to analyze what exactly have happened thus didn’t report it anywhere).
Nevertheless, do you have a USB/serial cable Serial connection - Turris Documentation ? I was eventually able to use it to reflash the firmware - Omnia - Turris Documentation - I have reflashed both “uboot” and “rescue” images, using the option “transfer via USB flash disk” (as the “Transfer via serial console” wasn’t for some strange reason cooperating with me). Next, I have recovered the software via the latest medkit - https://repo.turris.cz/hbs/medkit/omnia-medkit-latest.tar.gz - this one needs to use a USB flash disk anyways, so retrospectively I should not have spent my time/energy with the “Transfer via serial console”
It removed the switch driver from kernel so nothing will work. I was lucky I didnt aprove it. Waiting with pkgupdate to something that will not break the system.
Thank you very much chaps!
You are really great: Help where help is needed, fast. I never regretted to buy my first Turris Omnia and I never regretted to buy a second Turris Omnia for a friends company.
And I would do it again.
Your analysis was right, by the way.
And yes, I do use HBT – Here Be Turtles on my device for quite a while (the production device with a friends company uses HBS – Here Be Snails)
What I did now:
Tried to roll back into the last working state by pressing the button until 2 lamps are active. It did something. And it at least tried to reboot.
Interesting: Eventually ALL lamps were shining blue, all lamps essentially dysfunctional.
But the device was reachable even by WLAN! Working, at least all basic functionality including LuCi.
Easy.
Then
I asked schnapps for help: “schnapps rollback 449” was supposed to roll back to a much older state known to work properly. And it did!
Some people say schnapps is a little special but absolutely great: True! On other systems it is no swiss knife but it perfectly interacts with Turris Omnia.
So in this moment Turris Omnia seems to work as usual.
At best until the next update.
Understood:
This did not address the underlying issue (which is not understood yet) and can not heal corrupted files or file system areas.
I strongly assume there are plenty of corrupted files that did not matter much before the upgrade (Educated guess … OK, things like nextcloud or container and some other things stopped working quite some years ago …)
I intend to
back up almost all existing (config) files now,
then do a total factory reset and
manually reconfigure the device into the current “service status”. Like having an almost perfect prefix delegation mechanism in IPv6.
I have all the templates in the drawer (/etc backup). Configuring everything during the weekend sounds like a lot of stupid and error prone work ahead …
I want to have the same config like before but avoid all corrupted files (binaries as well as malconfigurations)
questioN:
Is there an idea how to re-configure an (updated) vanilla device into (my) latest “custom” config without copying files back?
However: Many thanks to the kind and helpful people in the Turris forum, appreciate it, will always continue to recommend Turris devices!
Breaking Changes
