Nextcloud 20 new issues

[criky]

Hello,

I found some Nextcloud 20 issues on my box with the latest update. Please take a look to see what's happening.

1. sudo ubos-admin status problem
Emerging problem regarding Nextcloud Elasticseartch service failure

2. Nextcloud 20 group can not be unticked.
Once group is assigned to any user in "Users" menu, then it is not possible to untick the group anymore, and makes error message like below.
But previous version it can be changed.

Thanks.

[j12t]

Can you figure out why the elasticsearch daemon failed? Does restarting it work? What does the journal report?

[criky]

I didn't run anything special related to Nextcloud.

And restarting (booting the box)doesn't seem to fix it. If you mean the single service that needs to restart, then please let me know what to do.

If I invoke

Code: Select all

sudo journalctl

, it says;
(I picked some specifics presumably related to this matter, but not sure)

[j12t]

Here is more detail what to do:

If systemd service "foo" has failed, you can ask the journal for its messages (only), such as (if you wanted the last 5 minutes of messages only):

Code: Select all

sudo journalctl -u foo --since -5m

So in this case "foo" would be "nextcloud-fulltextsearch-elasticsearch@addbc..." because that's the service that has failed (as shown by "sudo ubos-admin status")

To try restarting it:

Code: Select all

sudo systemctl start foo

If that doesn't make it obvious what the problem is, or how to fix it, it would be best if you could send e-mail us the entire boot log.

[criky]

Here is the result of what journal says when I invoke

Code: Select all

sudo journalctl -u nextcloud-fulltextsearch-elasticsearch@addbc...

It seems "can't call method "vars" on an undefined value at ... " makes this matter.

Maybe this service has troubles to be compatible with the UBOS update.

Do you think reinstalling this service(delete this accessory and redeploy it) might help?

[j12t]

Hmm. Not sure what that is. Works fine here. Is it possible this is a leftover process from some installation changes you made and there is a second (and working) elasticsearch daemon running?

Try this:

Code: Select all

sudo ubos-admin listsites --with nextcloud-fulltextsearch-elasticsearch --detail

This should show a single site with your Nextcloud installation (unless you have more than one Nextcloud instance running on the same box) and the site id and appconfig id.

Then:

Code: Select all

sudo systemctl | grep nextcloud-fulltextsearch-elasticsearch

This should show a single line, like this:

Code: Select all

nextcloud-fulltextsearch-elasticsearch@addbc....service loaded active     running      Nextcloud full-text search using Elasticsearch

where the identifier after the "@" is the same as the appconfig id from above.

The best guess I have right now is that you have two lines there, one of which is active and the other one failed. If this guess is correct, disable the one that failed, it should not have been running in the first place:

Code: Select all

sudo systemctl disable --now nextcloud-fulltextsearch-elasticsearch@addbc...

If the guess isn't correct, then please send the output of

Code: Select all

sudo ubos-admin listsites --json
sudo systemctl | cat

[criky]

Is it possible this is a leftover process from some installation changes you made and there is a second (and working) elasticsearch daemon running?

Try this:

Code: Select all

sudo ubos-admin listsites --with nextcloud-fulltextsearch-elasticsearch --detail

This should show a single site with your Nextcloud installation (unless you have more than one Nextcloud instance running on the same box) and the site id and appconfig id.

Maybe. But this is my first installation of Nextcloud app & its accessories after UBOS is clean installed. (I repeated this whole process second time but just one installation per each tryout)

Since this is my first install, I asked you if I need to try uninstall and reinstall the problematic accessory.
Please also note I run only one Nextcloud instance on the box.

One suspicious thing is that after clean install I deployed the json file second time for there was a letsencrypt registration problem.

When I invoke

Code: Select all

sudo ubos-admin listsites --with nextcloud-fulltextsearch-elasticsearch --detail

It says like below and also proves that I've installed Nextcloud & its accessories only once at a time.

[criky]

Then:

Code: Select all

sudo systemctl | grep nextcloud-fulltextsearch-elasticsearch

This should show a single line, like this:

Code: Select all

nextcloud-fulltextsearch-elasticsearch@addbc....service loaded active     running      Nextcloud full-text search using Elasticsearch

where the identifier after the "@" is the same as the appconfig id from above.

The best guess I have right now is that you have two lines there, one of which is active and the other one failed. If this guess is correct, disable the one that failed, it should not have been running in the first place:

Code: Select all

sudo systemctl disable --now nextcloud-fulltextsearch-elasticsearch@addbc...

The best guess I have right now is that you have two lines there, one of which is active and the other one failed. If this guess is correct, disable the one that failed, it should not have been running in the first place:

Code: Select all

sudo systemctl disable --now nextcloud-fulltextsearch-elasticsearch@addbc...

Yes, there are two services that are running, but that is not like what we are expected.

As you can see below, the first one seems to be problematic.
It has a name(appconfig id) that is partly left out with three periods.

And it repeats continuously between "auto-restart" and "running", seemingly makes the system "degraded".

So after disable it and reboot, the problem is fixed.

[criky]

But some doubts from me left.

I wonder where the "nextcloud-fulltextsearch-elasticsearch@addbc....service" came from.

Is that the same service with what has a full name of "nextcloud-fulltextsearch-elasticsearch@addbc69ed3d3f0fb4df7d9697a54a851579ddc2af.service"?
I can see two sesrvices loaded at the same time by the UBOS.

Apparently there was a problem before I invoke this code; (I used three periods rather than full name)

Code: Select all

sudo systemctl start nextcloud-fulltextsearch-elasticsearch@addbc...

Starting this service as "nextcloud-fulltextsearch-elasticsearch@addbc..." makes the previously failed service starting again and again or causes new service to start?

Now I see you mentioned

where the identifier after the "@" is the same as the appconfig id from above.

Probably you mean "where the identifier after the "@" [should be exactly] the same as the appconfig id from above".
But previously I thought three periods can substitute the full name instead.

So complicated for me now, but anyway thanks for your advice, it really worked.

[j12t]

The three periods is a simplification we introduced for ubos-admin. The rest of Linux (that includes systemctl etc) knows nothing about it.

I filed an issue to print a better error message if somebody runs into this in the future, although it seems unlikely.

Glad you have it working.