Integrity check fails after update to 20.0

ℹ️ Support
1

Hello there,

after an update of Nextloud from 19 to 20, I have a big list of files that did not pass the integrity check.

Additional info: As I had some problems with the installation itself, I just copied the whole zip-folder (except the config-dir) in the running installation. But that did not lead to such problems in the past.

List of files that did not pass integrity test can be viewed here:

How can I solve this?

Nextcloud version (eg, 18.0.2): 20.0
Operating system and version (eg, Ubuntu 20.04): Ubuntu 20.04
Apache or nginx version (eg, Apache 2.4.25): 2.4.41-4ubuntu3.1
PHP version (eg, 7.1): php7.3

2

Copying over all files isn’t the optimal way to go. Better is to rename the existing Nextcloud document root, extract all files to a new directory with the original name, copy over the app directories, create symbolic links to the config and data directory (if necessary) and set the ownership and access rights as required.

This prevents you from getting a list of files, which might not be required any more and are not part of the original archive file.

3

Thank you for your reply. I will take this as basis steps for further updates.

But this does not help to solve my problem now. Is there any way I can resolve the mentioned files in integrity check?

4

Manually remove all files which don’t have a valid checksum.

5

Hi there,

thanks for your reply.
I managed to delete all files (of course, after having created a backup).
Now error is gone.

What I came across: The list of files that did not pass the integrity test is surely readable. But if you need to delete such a high amount of files, it would be good if there is a possibility to create an output usable directly in shell. Did I miss something here and this is already possible or should I create a feature request?

6

I have in mind that the upgrade process description definitely says that you shouldn’t extract a new release over an existing one to prevent such problems. Instead you should extract it to a new directory. So there is normally no requirement to to pipe this information to a file, because of only a handful number of files which might being reported on an upgrade.