!SEjxdEhDPWoPxzjWCX:matrix.org

Podman

1081 Members
Podman (https://podman.io) is a tool for running PODS and OCI Containers on Linux. Mailing lists at lists.podman.io. Current version 5.0.0. Bridged with the #podman channel on Podman's Discord(https://discord.gg/x5GzFF6QH4). The bridge with #podman on Libera has been temporarily suspended, but we now have a `Podman Relay` in place.68 Servers

Load older messages


SenderMessageTime
18 Apr 2024
@syntolog:matrix.orgsyntolog joined the room.19:19:21
@urbenlegend:matrix.orgUrbenLegend joined the room.19:48:48
@dschadlich:fedora.imDavid Schadlich joined the room.23:38:38
19 Apr 2024
@xyhhx:matrix.orgxyhhx Ⓜ️ (they/them) joined the room.00:56:08
@znmeb:fedora.imM. Edward (Ed) Borasky joined the room.03:25:57
@bitblt:matrix.orgbitblt

Hello again! Has anybody noticed that the podman-next version from copr doesn't work?
On a vanilla almalinux 9.3 vm, installed podman using https://podman.io/docs/installation#installing-bleeding-edge-versions-of-podman, and even a simple version command gives:

[user@test ~]$ sudo podman version
Error:

The system logs aren't really helpful either, as don't show any errors
The weird thing is that if I run podman version as a regular user it works, only on root user it fails like above

08:29:34
@felix:0x23.me@felix:0x23.me joined the room.09:54:46
@felix:0x23.me@felix:0x23.me

I have two containers running on the same host.
One has traefik in it as a reverse proxy.

I try to access container2 (an app) from container 1(traefik) through the host ip and the respective bound port.
From any other machine this work fine but nut on the same machine from container to container.
Each container is in its own network. When I join them together in one network, it works.
So it seems the communication does not go through the host but directly from container to container.
Is there any way to solve that connectivity issue without joining them into one network?

10:09:49
@logo4poop:matrix.orglogo (he/him) joined the room.14:43:31
@potatodoggo:matrix.orgdoggo

trying to translate a docker-compose instruction to be more podman compatible, but idk if it works.

here is a line that defined a volume in the original docker-compose file im trynna convert:

  • /var/run/docker.sock:/var/run/docker.sock

for podman, it should be rootless, so idk how to translate this line... need a little help 😀️

20:48:06
@potatodoggo:matrix.orgdoggo *

trying to translate a docker-compose instruction to be more podman compatible, but idk if it works.

here is a line that defined a volume in the original docker-compose file im trynna convert:

volumes:
   - /var/run/docker.sock:/var/run/docker.sock

for podman, it should be rootless, so idk how to translate this line... need a little help 😀️

20:48:45
@potatodoggo:matrix.orgdoggo *

trying to translate a docker-compose instruction to be more podman compatible, but idk if it works.

here is a line that defined a volume in the original docker-compose file im trynna convert:

volumes:
   - /var/run/docker.sock:/var/run/docker.sock

for podman, it should be rootless, so idk how to translate this line to be rootless... need a little help 😀️

20:49:06
@kenrachynski:matrix.orgDoofusCanadensisThat just mounts a socket file23:47:09
@kenrachynski:matrix.orgDoofusCanadensisSo that the container mounting that can control the docker daemon.  23:47:29
20 Apr 2024
@gigidi5557:matrix.orggigidi5557 joined the room.12:57:51
@mreich1978:matrix.orgMatt joined the room.20:29:43
21 Apr 2024
@urbenlegend:matrix.orgUrbenLegendIs there a reason why the real IP of incoming requests isn't getting passed from Nginx Proxy Manager to my other containers when they're run in rootless mode? My configuration works in rootful mode, but in rootless, things like Nextcloud and Jellyfin only see requests as coming from the Nginx Proxy Manager IP address02:29:46
@urbenlegend:matrix.orgUrbenLegendSeems like its related to this bug: https://github.com/containers/podman/issues/819304:19:27
@urbenlegend:matrix.orgUrbenLegendHow are you guys handling reverse proxies in rootless mode?04:19:35
@eriksjolund:fedora.imErik SjölundThe source IP address is available to nginx if you configure nginx + rootless podman to use socket activation. I wrote some documentation about running nginx with rootless podman https://github.com/eriksjolund/podman-nginx-socket-activation/tree/main/examples/example4 On my todo-list is to see if X-Forwarded-For HTTP header works. Then the backend servers would also get the information about the source IP address.08:26:13
@eriksjolund:fedora.imErik Sjölund I also did some experimenting with nextcloud and rootless podman in November 2023. https://github.com/eriksjolund/nextcloud-podman I think I got so far that I could see the nextcloud web interface running but I didn't do any further testing. Socket activation is used https://github.com/eriksjolund/nextcloud-podman/blob/main/nginx.socket08:39:39
@felix:0x23.me@felix:0x23.me left the room.09:13:20
@baerentoeter:matrix.orgbaerentoeterhi guys, im - again - having issues with podman (5.0.0) and selinux: i have a container that starts an sshd on port 22022 using network_mode: host. it works perfeclty fine when selinux is in permissive mode, but i cannot connect to the server when selinux is in enforcing mode. i have followd the steps from https://access.redhat.com/documentation/de-de/red_hat_enterprise_linux/8/html/using_selinux/creating-selinux-policies-for-containers_using-selinux but installing the policy generated by udica fails with "Failed to resolve typeattributeset statement at /etc/selinux/targeted/tmp/modules/400/base_container/cil:6" any ideas?11:20:44
@baerentoeter:matrix.orgbaerentoeter
In reply to @baerentoeter:matrix.org
hi guys, im - again - having issues with podman (5.0.0) and selinux: i have a container that starts an sshd on port 22022 using network_mode: host. it works perfeclty fine when selinux is in permissive mode, but i cannot connect to the server when selinux is in enforcing mode. i have followd the steps from https://access.redhat.com/documentation/de-de/red_hat_enterprise_linux/8/html/using_selinux/creating-selinux-policies-for-containers_using-selinux but installing the policy generated by udica fails with "Failed to resolve typeattributeset statement at /etc/selinux/targeted/tmp/modules/400/base_container/cil:6" any ideas?
also, everything works fine when selinux is in enforcing and the container is ran in privileged mode. however, i doubt that this is the best practice, eh
11:23:41
@eriksjolund:fedora.imErik SjölundThe sshd problem might be related to this troubleshooting tip https://github.com/containers/podman/blob/main/troubleshooting.md#33-the-sshd-process-fails-to-run-inside-of-the-container11:37:58
@eriksjolund:fedora.imErik Sjölund(or maybe not, at least the tip was about sshd)11:40:13
@baerentoeter:matrix.orgbaerentoeter
In reply to @eriksjolund:fedora.im
The sshd problem might be related to this troubleshooting tip https://github.com/containers/podman/blob/main/troubleshooting.md#33-the-sshd-process-fails-to-run-inside-of-the-container
thanks Erik, but i think that is not the problem im facing - i dont have reason to believe that loginuid is an issue. actually my question aims for a more general answer to the question why the whole "podman-inspect->udica->semodule" toolchain fails
11:45:09
@baerentoeter:matrix.orgbaerentoeterSince my setup works fine when selinux is permissive, it is quite safe to say that it is an issue with the labels I guess11:52:57
@devurandom:matrix.orgDennis (7616) joined the room.13:41:41
@lucas_sr:matrix.org@lucas_sr:matrix.org joined the room.23:35:06

Show newer messages


Back to Room ListRoom Version: 6