It’s closed because the OP never replied (and it’s an older post).
Running the Updater from the command-line is not merely a workaround. It’s the most reliable way to use the Updater. It avoids problems due to web infrastructure timeouts, etc.
I’m sorry for not responding earlier. This problem still occurs with me, pretty much invariably. Now I simply wait a sufficient time to ensure the file has downloaded, go back and start again. That has always worked so far but it’s clunky.
I’m happy to try the updater from the command line but I assumed this only started the updating process after the file had been downloaded and verified. What does the updater from the command line actually do? Why is it different than doing the same thing from within the admin page on Nextcloud?
The command line mode of the Updater does the same thing (it’s the same code underneath, but with fewer variables such as web server end to end timeouts that can occur during long running processes). It is also easier to troubleshoot since the error messages are displayed (the web version tries to show what it can, but the nature of some errors means they aren’t always caught for display there).
P.S. In the upcoming v32 (note: not the about to be released v31), the Updater download step now has resume support and also real-time logging of the download progress. The latter in particular should help in environments where the download is interrupted or taking awhile for whatever reason.
The Step-by-Step Manual Upgrade in the manual involves a lot of steps before the OCC updater command is ready to be issued. The web-based updater seems to handle all these things (although maybe not stopping/starting the crontab jobs and definitely not the backup). So I will continue with my ‘workaround’ method as it seems fine (unless the risks are too great, are they? Maybe I have just been lucky with the numerous upgrades I’ve done so far)