Active Topics

 


Reply
Thread Tools
Fellfrosch's Avatar
Posts: 1,092 | Thanked: 4,995 times | Joined on Dec 2009 @ beautiful cave
#141
Well, I have that as well. I see, i have to make further tries on that side.
 

The Following User Says Thank You to Fellfrosch For This Useful Post:
Posts: 202 | Thanked: 385 times | Joined on Jul 2018
#142
Originally Posted by Fellfrosch View Post
So you have made no adpations in the scripts to the new username defaultuser???
I have updated ota since 3.2, so still on nemo... (defaultuser gets written in on clean install)
 

The Following 2 Users Say Thank You to levone1 For This Useful Post:
Fellfrosch's Avatar
Posts: 1,092 | Thanked: 4,995 times | Joined on Dec 2009 @ beautiful cave
#143
Thanx for clarification!
 

The Following User Says Thank You to Fellfrosch For This Useful Post:
Posts: 202 | Thanked: 385 times | Joined on Jul 2018
#144
Originally Posted by Fellfrosch View Post
So you have made no adpations in the scripts to the new username defaultuser???
So I did a clean flash of 4.0, and decided to tinker with Containers for defaultuser, and it's actually pretty simple:
1 - download source
2 - make change to daemon.py file
3 - add %setup line to spec

Then just use rpmbuild to create rpm. Installs and runs normally ...
 

The Following 3 Users Say Thank You to levone1 For This Useful Post:
Posts: 1,335 | Thanked: 3,931 times | Joined on Jul 2010 @ Brittany, France
#145
We should raise a PR for that, but I have not found how to write it so that daemon.py would work with any user name, not just defaultuser. Maybe using `import os` and then `os.environ.get('USER')` would work?

Anyone willing to try with this commit/fork? I don't have a device ready to try and I'm not prepared to compromise my LXC install right now as I really depend on it, but if that works, I can submit the PR.

I am sorry I didn't read TMO lately but I see this is being discussed since March, while others had found the same the same temporary workaround at some earlier point on either the Telegram or Discord groups.

Last edited by Kabouik; 2021-05-21 at 08:42.
 

The Following 3 Users Say Thank You to Kabouik For This Useful Post:
Posts: 39 | Thanked: 81 times | Joined on Jul 2011
#146
Originally Posted by Kabouik View Post
We should raise a PR for that, but I have not found how to write it so that daemon.py would work with any user name, not just defaultuser. Maybe using `import os` and then `os.environ.get('USER')` would work?

Anyone willing to try with this commit/fork? I don't have a device ready to try and I'm not prepared to compromise my LXC install right now as I really depend on it, but if that works, I can submit the PR.

I am sorry I didn't read TMO lately but I see this is being discussed since March, while others had found the same the same temporary workaround at some earlier point on either the Telegram or Discord groups.
I do not have any Sailfish devices currently, so I can't test it, but since you are on the quest of fixing the problem by removing coded in constants which are actually environment variables (ok for early test, no-go for release), you could try to also remove the fixed UID by using os.getuid().
 

The Following 5 Users Say Thank You to txus For This Useful Post:
Posts: 202 | Thanked: 385 times | Joined on Jul 2018
#147
Originally Posted by Kabouik View Post
We should raise a PR for that, but I have not found how to write it so that daemon.py would work with any user name, not just defaultuser. Maybe using `import os` and then `os.environ.get('USER')` would work?

Anyone willing to try with this commit/fork? I don't have a device ready to try and I'm not prepared to compromise my LXC install right now as I really depend on it, but if that works, I can submit the PR.

I am sorry I didn't read TMO lately but I see this is being discussed since March, while others had found the same the same temporary workaround at some earlier point on either the Telegram or Discord groups.
I built rpm using your daemon.py, and it is working fine. I already had Containers installed, and had a working Debian container, and I just uninstalled my Containers app, and installed the new build, so maybe not the best test... I can upload/link rpm if appropriate ....
 

The Following 5 Users Say Thank You to levone1 For This Useful Post:
Posts: 1,335 | Thanked: 3,931 times | Joined on Jul 2010 @ Brittany, France
#148
Thanks to both of you! I added `os.getuid()` too as you recommended txus. I'll try to get someone to test it on a device with no current installation of LXC and no existing container, in case that matters. I'll submit a PR if that works.
 

The Following 3 Users Say Thank You to Kabouik For This Useful Post:
Fellfrosch's Avatar
Posts: 1,092 | Thanked: 4,995 times | Joined on Dec 2009 @ beautiful cave
#149
I have currently no running containers. So I can test for you. So what do I have to do?
At first uninstall LXC if I got you right... And then? What rpms should I install?
 

The Following 3 Users Say Thank You to Fellfrosch For This Useful Post:
Posts: 1,335 | Thanked: 3,931 times | Joined on Jul 2010 @ Brittany, France
#150
You could install the rpm that is already compiled on the Github repository, which you may already have.

Then kill the daemon: run "ps aux | grep daemon.py", then "kill -9 <daemon.py PID>".

And then replace your /usr/share/harbour-containers/service/daemon.py with the one in the fork above. Make sure you download the raw daemon.py and not the HTML view. And reboot.

One thing to keep in mind is you may have leftover configs in /var/lib/lxc. So if that doesn't work, it might be worth checking those and/or creating a new container, to make sure the configs are not still trying to use nemo.
 

The Following 6 Users Say Thank You to Kabouik For This Useful Post:
Reply

Tags
application, development, linux


 
Forum Jump


All times are GMT. The time now is 23:04.