YunoHost 3.8 testing / Call for feedback

Hello everyone!

We just released a new testing version for YunoHost and would be happy to receive feedback before releasing it as a stable version :yum:

For now, this release essentially includes the following major changes:

  • :stethoscope: New diagnosis system : The configuration of a server is complex and prone to a lot of small errors, that’s why we have set up a brand new diagnostic system that analyzes different parameters such as: the possibility to access the server from outside, the DNS records, port forwarding, system resources… We hope that this feature will help you and that’s why we are looking forward to your feedback during this test version. (Disclaimer: this feature is still experimental and being polished)

  • :bookmark: App categories : As the list of packaged applications for YunoHost is getting wider and wider, a better search and visualization system was needed. Applications are now grouped by area of interest.

  • :card_file_box: Support for XMPP http upload : You can now send files through XMPP using you YunoHost server! Metronome also gets upgraded to 3.13.4.

  • :jigsaw: Easier PHP version and apt repo management (for app packagers) : We include a new way of managing PHP versions and to simplify version migrations. This should also be useful with regard to the transition towards Buster and next Debian versions. See here for more details.

  • :rocket: And a lot more as you’ll find out in the changelog!

Thanks to all contributors :heart: ! Abdulkadir F. Ş., advocatux, Aeris One, Aleks, Allan N., Alvaro, amirale qt, Armando F., Arthur L., Augustin T., Bram, ButterflyOfFire, Damien P., decentral1se, Elie G., Eric G., Filip B., frju365, Giovanni G., Gustavo M., Hem S., Jeroen F., Jimmy M., Josué, Juan, Kay0u, ljf, Maniack Crudelis, Mario, Matthew D., Mélanie C., Patrick B., pitchum, ppr, Romain R., Quentí, tituspijean, xaloc33, yalh76, Yasss Gurl, Yasin S. T., Yifei D., Zeik0s

:hammer_and_wrench: Detailed changelog

Somewhat major stuff

Misc technical fixes and improvements


Tests, cleaning, refactoring

How to participate to the beta-testing :construction_worker_woman: :construction_worker_man:

:warning: DO NOT do this on a critical production server!

From the command line, you can launch the following command to switch to testing:

curl | bash

(If you are familiar with bash scripting, you might want to read what this does before blindly running the command)

After this command, you should be running YunoHost 3.8.0. Before using the webadmin, please make sure to force-refresh the cache of your browser.

What to test? :space_invader: :telescope:

Here are some specific items which are important to check to validate the current work:

  • upgrade an existing instance and check that everything is working
  • try out the new diagnostic system and let us know if you find a problem in the detection/non-detection of potential problems
  • test the new app categories to find what you are looking for in our catalogue

My feedback:

My YunoHost server

Hardware: Raspberry Pi at home
YunoHost version: 3.7.1
I have access to my server : Through SSH through the webadmin
Are you in a special context or did you perform some particular tweaking on your YunoHost instance ? : yes
If yes, please explain: Local network access only, otherwise everything goes through a ZeroTier network. Internal fake tld used as main domain. DNS server to handle it on another YNH server.


Stops at apt dist-upgrade because it requires a Y to continue.
Continuing by manually doing apt dist-upgrade -y.

Re-diagnosing server health...
Error: Diagnosis failed for category 'basesystem': Command 'systemd-detect-virt' returned non-zero exit status 1
Warning: Found 2 item(s) that could be improved for Internet connectivity.
Error: Found 4 significant issue(s) (and 1 warning(s)) related to DNS records!
Error: Found 8 significant issue(s) related to Ports exposure!
Error: Found 2 significant issue(s) related to Web!
Success! Everything looks good for Email!
Success! Everything looks good for Services status check!
Error: Found 2 significant issue(s) (and 1 warning(s)) related to System resources!
Warning: Found 3 item(s) that could be improved for System configurations.
Success! Everything looks good for Security checks!
  • I find the errors a bit scary, but I understand the Success/Warning/Error gradation.
  • 'systemd-detect-virt failing may need to be looked into though?
Diagnostics panel

  • Back in my day… :older_man:
    This item fails again with Échec du diagnostic pour la catégorie 'basesystem': Command 'systemd-detect-virt' returned non-zero exit status 1
  • The remainder of my page is full of horrors errors, but that’s due to my weird networking setup.
  • A great idea would be to link warnings and errors to their corresponding documentation page: how to assess if they are bad or not, how to fix them.
Groups and permissions panel

Partial translation to French here.

Apps catalog

  • Had a real hope when I saw the screenshot in the first post, but it was only the one used for the mockup. I already wrote that elsewhere, and I understand there are limitations, but having the search bar directly on the first catalog page would be great.
  • Loving the responsiveness!
  • Slight visual bug though, with the warning icon being cut out.
  • Should we display the apps level in addition to “high quality” and “does not work”? Maybe not in the catalog, but in the apps information panel when starting to install them?


  • Weird error while try to update all apps at once. Individual update buttons work though.
  • Oh, and the “powered by YunoHost” text is missing at the bottom right of the admin page. I did not notice it before. Is it specific to this testing release, or is it removed any time we switch to testing?

[enh] Add subcategories management in bash completion (YunoHost#839)

  • How do I test it? Either in bash or zsh, I get no autocompletion on yunohost commands.


I just upgrade to the 3.8 testing without any trouble. I did not checked everything yet, but the Diagnosis tools is super cool !
I have a lot of warnings I need to look after, but no major flow it seems.

Thanks for your hard work guys !


hi here

  1. i just updated to 3.8 testing , without any issue -all good
  2. the Diagnosis tool works well - except sometimes lazy when i re run some parts of diag --> then all good.
    just a point , i’m connected to IPV4 and IPV6 network and i checked via command line it works i can see IPV4 and IPV6 sites, but the diag says ‘yes’ for IPV4 and ‘no’ for IPV6, so perhaps a point to be checked
  3. a question - out of this is the 3.8 compatible with Debian 10? if so can i upgrade my debian from 9 to 10? any advise here?
    thanks for all your efforts.

How did you exactly ? Can you confirm that if you run curl on your server you see your ipv6 ?

Nope it’s not. But the plan is to, after the 3.8, provide a way to migrate as easily as possible to 4.0. That’ll be shipped on top of 3.8 (or call it 3.9, doesn’t really matter)

Anyway, in parallel I’m doing some fixes / improvements to the diagnosis system to address several issues (as reported by @tituspijean and others)

hi @Aleks
i m sure my server is connected IPV6 as i can ssh it via its IPV6 adress
but the curl doesn t work from the server
seems i need to fix something in debian to make it full IPV6, or i yunohost config?
when i ping6 it works from the server
any advice here

Alright, but there’s a difference between having a working local IPv6 and being able to contact other server on the internet via ipv6 …

With that said, it looks weird that you’re able to ping6 but not curl

Maybe as a complement : if you go to this website (from your laptop/desktop) does it show you have a working ipv6 ?

yes it says ok in IP V6 and shows the IPV6 of my mac - all good

i can also ping6 with suceess this site ping6 from my server where yunohost is located

Imokay so maybe the issue is on our infra lol … Are you able to ping6 ?

when i do PING i get that below with nothing moving fwd, like waiting forever
'PING (2001:910:1410::1)) 56 data bytes

when i do
root@xour:~# ping6 2001:4860:4860::8888
PING 2001:4860:4860::8888(2001:4860:4860::8888) 56 data bytes
64 bytes from 2001:4860:4860::8888: icmp_seq=1 ttl=50 time=4.78 ms
64 bytes from 2001:4860:4860::8888: icmp_seq=2 ttl=50 time=4.79 ms
64 bytes from 2001:4860:4860::8888: icmp_seq=3 ttl=50 time=4.81 ms
64 bytes from 2001:4860:4860::8888: icmp_seq=4 ttl=50 time=4.80 ms
64 bytes from 2001:4860:4860::8888: icmp_seq=5 ttl=50 time=4.80 ms
— 2001:4860:4860::8888 ping statistics —
5 packets transmitted, 5 received, 0% packet loss, time 4005ms

Yea so I suspect we have an issue with ipv6 on our infra, wouldn’t be the first time … Investigating … Thanks for the tests

@pascalv : Infrastructure should be fixed. Can you re-run the diagnosis for Internet connectivity from the webadmin ?

@Aleks it works now ipV4 and IPV6 - both are green and ok - cool fixed

1 Like

hi here
@Aleks. seems diagnostic is moving forward
had 2 new issues poping up this morning
1/ on DNS that i fixed by adding my IPV6 adress in TXT DNS - all good
2/ this new message on Web diagnosis
’ Domain xxxx appears unreachable through HTTP from outside the local network’ - An error happened while trying to reach your domain, it’s very likely unreachable.’
how can i fix this ?
i have no idea, is it DNS related or something else?
please advise - pour info je parle aussi francais :wink: pascal

Indeed there was another issue on our infra in the diagnosis system when in IPv6 :wink: Sounds like those tests are pretty useful :stuck_out_tongue:

Can you try to force re-run this piece of diagnosis and see if the results is better ?

all fixed. all my diagnosis is green - thanks a lot

another comment
HI here
for the diagnostic, i received at 7am an email with errors in the diag and each team i re run it manually --> all good. Could you implement a mechanism that run the diag twice or three time with few min intervalls and only send an email with alarming if the 3 results have issues. Seems something can be improve here
just my 2 cts

1 Like

To all people on testing right now: I’m gonna play a bit tonight with a new version of the diagnosis server which includes breaking changes. So if you happen to run diagnosis for http or ports, it’s gonna fail which is expected


6 posts were split to a new topic: Weird high load