[solved] Nextcloud All-In-One: web access broken during upgrade from v26 to v27? No, but you need to access the server using its AIO port to resume the upgrade

ℹ️ Support 📦 Appliances (Docker, Snappy, VM, NCP, AIO)
1

Nextcloud version (eg, 20.0.5): from v26 to v27, running as Nextcloud All-in-One, installed from an image running on Linode. Cannot see whether v27 upgrade has suceeded (no web interface available).
Operating system and version (eg, Ubuntu 20.04): Ubuntu 22.04.2 LTS
Apache or nginx version (eg, Apache 2.4.25): don’t know. “httpd -v” returned “Command ‘httpd’ not found, did you mean…”
PHP version (eg, 7.4): don’t know. “php -v” returned “Command ‘php’ not found, but can be installed with…”

The issue you are facing:

Can log into the server using SSH from a desktop client.
Cannot access the server using a web browser from a desktop client. The client’s Chromium on Linux x64 says, “The site can’t be reached. [IP addr]|[hostname] refused to connect.” The client’s Firefox on Linux x64 says, “Unable to connect: an error occurred during a connection to [IP addr]|[hostname].” Firefox also seeks https connections, which creates a new message a HTTPS version of [hostname] is not available."

Is this the first time you’ve seen this error? Yes.

Steps to replicate it:

  1. Log into Nextcloud v26 via web interface as admin user.
  2. Log into All In One interface.
  3. See message that a new masterbase is available to upgrade. Click to stop all containers.
  4. Click on upgrade.
  5. Server does its thing, but, unlike the first upgrade to v26, services just become unavailable. Web connections to the server no longer possible.

The output of your Nextcloud log in Admin > Logging:
→ not available, needs web access.

The output of your config.php file in /path/to/nextcloud (make sure you remove any identifiable information!):
→ can’t see a path to nextcloud in this server. What should the path be in the All In One image available from the Linode marketplace?

The output of your Apache/nginx/system log in /var/log/____:
→ the path exists, but there is no file that is obviously related to Apache. What should the path and filename be in the All In One image available from the Linode marketplace?

Output errors in nextcloud.log in /var/www/ or as admin user in top right menu, filtering for errors. Use a pastebin service if necessary.
→ no such folder “/var/www”. What should the path be in the All In One image available from the Linode marketplace?

This issue looks similar to Broken UI after update to v27.0.1 , except that this ticket is definitely about an All In One installation where no files appear to exist on the server in the places that the admin user would expect to find them.

A re-boot of the server made no difference.

For both Nextcloud Android app 3.24.2 and 3.25.0, the app reports “Server not available.”

For dekstop Nextcloud desktop client sync app (appimage on Linux v64, v3.9) report “offline”.

A prior upgrade of the masterbase to Nextcloud v26 succeeded. A backup exists on the server for this upgrade. In the AIO interface, no backup option was available for the upgrade from v26 to v27. Is it going to be quicker just to delete this server and create a new one from the image in the Linode marketplace?

2

Hi, can you test if the server is reachable from mobile internet via the domain? Also is the dns config still correclty configured to point at your public ip-address and are the ports correctly forwarded in your router?

Also did you already verify if entering the domain into the client apps makes things work instead of trying the ip-address?

3

Thanks for your prompt reply. You have directed me to a solution.

A test on a mobile device using telco data connection results in the same thing:
using the hostname → “connection refused”
using the IP addr → “connection refused”

There have been no changes to the DNS config on the client side.

There have been no changed to the DNS config on the sever side (Linode). No changes to the IP addr. No changes to the A records.

On the server, ifconfig is not installed.

But on logging on via ssh (to find out whether ifconfig would tell me anything), I noticed a welcome message that suggested the IP addr with a specific port number. I’ve just pasted the IP:port number in a browser and landed got the AIO login (password) screen. After login, the containers needed starting. After two minutes, all of the containers were running, Nextcloud is at v27, the client apps have connected. All back to normal!

1 Like
4

Follow-up feature request.

In the AIO interface, it would be useful to have a paragraph on-screen to remind the user, “During this upgrade, this connection to the server will be lost and you will not be able to re-connect via https://. Instead, re-connect to the server via https://: to resume the update process and re-start the containers.”

Regular admins probably do this without thinking; for those of us who do admin work infrequently, we’re gonna forget.