Misconfiguration through ssh - no more login page

NextBox
1

Hello, since one of the reasons I wanted to work with Nextcloud on a Nextbox was that of collaborative editing on data physically located at my home, I looked for a way to run Collabora.
Unfortunately the internal app for the CODE server had issues downloading, and even after

forcing the download and install of the CODE server

logged in with SSH through
sudo docker exec -it -u www-data nexbox-compose_app_l php -d memory_limit=512M /var/www/html/occ app=install richdocumentscode_arm64
which completed successfully,

the server showed up as installed in the Nextcloud app administration, but no configuration was possible, and the Collabora app was trying to autoconfigure the existing server installation but could not get through.

A web search suggested to install the CODE as a

separate docker image 
docker pull collabora/code@sha256:<hash>
docker run -t -d -p 127.0.0.1:9980:9980 -e 'domain=<doubly escaped domain name>' -e "extra_params=--o:ssl.enable=false --o:ssl.termination=true" --restart always collabora/code

and afterwards

install apache 
apt install apache2

and edit /etc/apache2/apache2.conf to include a Listen 8080

but after a reboot this made the nextcloud installation not accessible anymore.
Connecting to the nextbox.local url now only showed me the Apache2 Debian Default Page, removing any access to the Nextcloud instance. I had lost SSH access as well, since I uncarefully deactivated it from the inside of the Nextbox app before the last reboot. The green LED was fast-blinking for hours without any progress.

After retrieving a suitable monitor, video cable and keyboard to access the system directly, I noticed that the docker images kept reloading every 15-20s or so. I tried to revert the changes by first

deactivating the CODE docker container 
docker update --restart=no <CODE docker container name>
docker kill <CODE docker container name>

and after a reboot, I saw that at least the 4 nextcloud-compose docker containers stayed loaded, and the green LED in continuous-on, but no access from the outside to the nextcloud instance yet (still apache default page).
Afterwards I

removed the apache package 
apt remove --autoremove apache2

as well and commented the modified line to apache2.conf out,

but could not further revert the situation.

Any tips or suggestions on what to look for or what to do in order to recover the starting state? I had no data saved nor personal services configured yet (except for some dummy data for tryouts), but I would prefer to avoid physically opening the Nextbox itself.

2

Hey @ntkem

this is pretty much exactly the reason, why the FAQ states that CODE is not recommended by us and moreover there are multiple warnings that if you start changing things via ssh you are pretty much on your own. So please understand that I will not comment or walk you through what command to run when, as this is pretty much generic Linux/Docker stuff.

This being said, I would suggest you to get in touch with networking basics, because you need to understand the concept of ports. In a very few words: Your (tcp/udp) network works with ports to connect to, you have set up an apache on the host system which blocks port 80 (as you see the apache page), so you essentially blocked the port for docker that’s why docker is restarting as it tries to bind to port 80, but you have already blocked it.

best

3

Thank you for the clue with the ports! I already guessed that something in the installation/uninstallation of the webservers had broken the ex-works configuration, but could not find the problem.

By tweaking the apache configuration files with respect to ports forwarding I was able to have the instance accessible from the net again.

Do you think there could be any other configuration which I should restore/recover?

4

Hey,

nope apart from the blocked ports I don’t think there is something missing.

best