Leantime upgrade to 3.4 is broken

[max]

Hi all,
Leantime is only partially functional after upgrade to 3.4. For example you get 'Server Error' when trying to add a milestone.
The logs show various errors after restarting the app.
E.g. Healtheck error: Error: connect ECONNREFUSED 172.18.16.53:8080
This is consistent across 2 installations on different boxes, clients and domains.
The normal restart, reboot etc do not fix the errors.
Restoring a backup to a fresh install on a different VM did not either.

[max]

I believe the update broke the app, probably when going from 3.4.2 to 3.4.3. (but it could have been earlier and not spotted)
Looking into the upgrade process I see there are three issues:

1. Changes in permissions - related to changes in Docker configuration updates - probably responsible for the bland 'Server error' issue
2. Internal port changed to 8080, probably responsible for the ECONNREFUSED 172.18.16.53:8080 issue
3. PHP 8.2 required, Laravel no longer supports 8.1 - no idea if this is an issue, are we on 8.2 already?

[max]

Leantime have posted:
Docker configuration updates
Due to the user/group updates on the docker image there is a small chance you run into file permission issues on existing volumes check https://github.com/Leantime/docker-leantime?tab=readme-ov-file#common-issues--solutions for setting the permissions correctly.

Additionally we updated the internal port to be 8080. Please make sure your docker compose files reflect the recent changes

Highlights
For a full list of 3.4 updates see: https://github.com/Leantime/leantime/releases/tag/v3.4.0
==> Fixed date issue in milestone, timesheet, sprint generation <==
Fixed cron handler for long running schedule:work job to be able to run cron jobs in docker
Misc other fixes

[max]

@marcelfolaron can you add your wisdom please
Could our issues be permissions or ports?

[max]

@nebulon Would you mind checking this?
I see you are working on PHP 8.2 for Leantime

[leantime]

Hey,
Yeah Leantime 3.4.x requires PHP 8.2

Some more info:
With 3.4 we updated and hardened our dockerfile and docker compose config. Not sure if you all run the docker containers as root or not but if not you can pass in pgid and uid now.

Inside the container Leantime is listening on port 8080 so if the docker compose file still maps to 80 internally that would need to change (external doesn’t matter). The user inside of the docker image is www-data:www-data

With that said old volumes may have the wrong permissions depending on the set up. We added a docker startup script to handle that for most cases but if permissions are off that may not work.

Lastly we tested the setup with various reverse proxy’s and it works in most environnents just make sure that:
A. LEAN_APP_URL is set correctly
B. The host server var is forwarded (ideally just forward any standard host variables)

More on that here https://github.com/Leantime/docker-leantime
And here
https://github.com/Leantime/leantime/releases/tag/v3.4.0

[girish]

It's updated to php 8.3 now - https://git.cloudron.io/packages/leantime-app/-/merge_requests/10

[max]

Thank you @girish and @marcelfolaron, I updated my test system with the fix and it worked fine, so I have updated production ones too.
That was resolved remarkably quickly and saved me a lot of strife, amazing, thanks again

[leantime]

Awesome, that was quick!

[leantime]

Oh while I see your package @girish
We fixed the cron issue and you can use the following command now (rather than calling that http endpoint:

php /var/www/html/bin/leantime schedule:work

This is the long running php process so if you have supervisor running that would be better.
Alternatively you can run the following command in a crontab which will execute and exit:

php /var/www/html/bin/leantime schedule:run

[girish]

@leantime thanks! i will fix the package . I found this bug in the meantime - https://github.com/Leantime/leantime/issues/2980 .

[leantime]

Thanks, on it!