- Podman - Podman List Archives

image platform (linux/amd64) does not match expected platform (linux/linux/arm64)
by Mehdi Haghgoo 14 Jun '24

14 Jun '24

Hi all, I'm trying to run rabbitmq images on MacBook M1 and I see warnings like: WARNING: image platform (linux/amd64) does not match the expected platform (linux/linux/arm64) I run the command: podman run --rm -p 8000:8080 --arch linux/arm64 ghcr.io/code-with-mehdi/rabbitmq:3-management and it fails to boot with a lot of logs that I like below: Mozilla Community Pastebin/QTExU80S (Plain Text) Mozilla Community Pastebin/QTExU80S (Plain Text) Is there a special flag I need to pass to fetch arm64 images? I'm not sure the boot issue has anything to do with the image arch incompatibility, though. Some enlightening on this from someone who has experience and insight will be appreciated. Best regards, Mehdi

2 4

Podman v5.1.1 Released
by Matt Heon 06 Jun '24

06 Jun '24

Hi all, Podman v5.1.1 is now available. This is a small bugfix release with fixes for Windows and Mac including a configuration file parsing issue on Windows. You can find full details in the release notes: https://github.com/containers/podman/releases/tag/v5.1.1 Thanks, Matt Heon

1 0

Fly containers on the fly using podman YAML
by Mehdi Haghgoo 05 Jun '24

05 Jun '24

I have problem running docker-compose on a Rocky Linux server 9.4. Instead of Docker Compose, I want to use Podman. A service in my compose.yaml file creates a container on the fly using Containerfile and runs the service. When running an container equivalent to this service, how can I create the container on the fly by first building using the Containerfile and then bringing up the service? Is this possible? Or should I first create a custom image, push it to repo and pull it before running the container?

2 2

Re: podmansh and persistent container
by Felix Niederwanger 31 May '24

31 May '24

Hello, I managed to achieve my goal to get a persistent container for my non-root user with podmansh working. In the end it was surprisingly easy to achieve, once I figured out which pieces to put together. I wrote the steps down in https://www.feldspaten.org/2024/05/31/Persistent-container-podmansh/ There is a TL;DR and a longer step-by-step guide. I put the link to the article also in https://github.com/containers/podman/discussions/19620. My takeaway was the elegance of `podmansh`. It's really just a `podman exec -ti podmansh /bin/sh` (I love this simplicity!). Once I figured out that it's really this, the rest was relatively easy. Hope the guide is helpful for someone else as well. Cheers! On 02/05/24 14:55, Petr Lautrbach wrote: > Lokesh Mandvekar <lmandvek(a)redhat.com> writes: > >> Hello Felix, >> >> podmansh was created with the idea of the admin locking down user shell >> environments, so installing software (to /usr) by the user itself won't >> work. An alternative could be the user installing to a non-standard >> location (I think there are some dnf / package manager tricks for this) in >> a persistent volume that gets mounted. >> >> Maybe toolbx (toolbox) could fit your use case if you want user >> customizable persistent installation. >> >> Copying Dan, Petr and Rishi in case they have further ideas. > > It was discussed in > https://github.com/containers/podman/issues/19497 > https://github.com/containers/podman/discussions/19620 > > AFAIK it's quadlet which generates systemd unit for the container and > which adds "--rm" option to podman [1] so you would need to skip quadlet > and run the container using systemd service. > > e.g. > - use `/usr/lib/systemd/system-generators/podman-system-generator --user --dryrun` > to generate systemd unit configuration and save it to > `~user/.config/systemd/user/podmansh.service` > > - change it so it does not call `podman rm` and does not use `--rm` in > `podman run` > > - use a wrapper which would `run || start` on ExecStart > > Something like the snippet bellow: > > > [Service] > ExecStartPre=/usr/bin/mkdir -p %h/data > Environment=PODMAN_SYSTEMD_UNIT=%n > KillMode=mixed > # ExecStop=/usr/bin/podman rm -v -f -i --cidfile=%t/%N.cid > # ExecStopPost=-/usr/bin/podman rm -v -f -i --cidfile=%t/%N.cid > Delegate=yes > Type=notify > NotifyAccess=all > SyslogIdentifier=%N > # ExecStart=/usr/bin/podman run --name=podmansh --cidfile=%t/%N.cid --replace --rm --cgroups=split --init --sdnotify=conmon -d -w=%h -v %h/data:%h:Z registry.fedoraproject.org/fedora sleep infinity > ExecStart=/usr/bin/bash -c '/usr/bin/podman run --name=podmansh --cidfile=%t/%N.cid --cgroups=split --init --sdnotify=conmon -d -w=%h -v %h/data:%h:Z registry.fedoraproject.org/fedora sleep infinity || /usr/bin/podman start podmansh' > > > Disclaimer: I haven't really tested it, there might be several issues > with this. Persistent containers were not our main goal for `podmansh` > > [1] https://github.com/containers/podman/blob/c9644ebccf14309a77769cba00833cd13… > > > Petr > > >> On Fri, Apr 26, 2024 at 3:20 AM Felix Niederwanger via Podman < >> podman(a)lists.podman.io> wrote: >> >>> Hello, >>> >>> I'm trying to get podmansh to work in such a way, that every user has a >>> persistent container. >>> >>> I've followed the steps of >>> https://docs.podman.io/en/latest/markdown/podmansh.1.html, resulting in a >>> nice setup, where at user login every user get's its own container. >>> Unfortunately the container for the user session is ephemeral, meaning >>> after logging out or system reboot the container is destroyed. >>> >>> I would like to have a setup, where quadlet setups a base container, but >>> then every user can install their own software and environment, which lasts. >>> >>> Here is the quadlet file I'm currently using: >>> >>> ``` >>> # /etc/containers/systemd/users/podmansh.container >>> [Unit] >>> Description=podmansh container >>> After=local-fs.target >>> ExecStartPre=-/bin/mkdir -p %h/data >>> >>> [Container] >>> Image=registry.fedoraproject.org/fedora >>> ContainerName=podmansh >>> HostName=arctic-fox >>> RemapUsers=keep-id >>> RunInit=yes >>> User=0 >>> >>> Volume=%h/data:%h:Z >>> WorkingDir=%h >>> >>> Exec=sleep infinity >>> >>> [Service] >>> ExecStartPre=/usr/bin/mkdir -p %h/data >>> >>> [Install] >>> RequiredBy=default.target >>> ``` >>> >>> Anyone an idea how to achieve persistent podmansh containers? >>> >>> Greetings, >>> phoenix >>> _______________________________________________ >>> Podman mailing list -- podman(a)lists.podman.io >>> To unsubscribe send an email to podman-leave(a)lists.podman.io >>> >> >> >> -- >> Lokesh >> Libera, GitLab, GitHub, Fedora: lsm5 >> Matrix: @lsm5:matrix.org >> GPG: 9E33DD8704CC03E2DEB84D9A1C1EDD7CC7C3A0DD >> https://keybase.io/lsm5 > -- phoenix(a)feldspaten.org, gpg: 0x6E77A590E3F6D71C Consider using plain text | email is not SMS https://email.is-not-s.ms/

1 0

Podman v4.9.5 Released
by Matt Heon 30 May '24

30 May '24

Hi all, We've just released Podman v4.9.5, a security fix for Podman v4 including a fix to CVE-2024-3727, a containers/image vulnerability. We will continue to provide security fix releases to Podman v4.9 for the next several months to give folks time to migrate to v5.0 releases. Thanks, Matt Heon

1 0

Podman v5.1.0 Released!
by Matt Heon 29 May '24

29 May '24

Hi all, Podman v5.1.0 has just been released. This release includes a number of new features, including support for Rosetta on macOS, improvements to `podman update`, and improvements to the Windows installer. and over 20 bugfixes. We expect our next minor release, Podman 5.2.0, to arrive in the late July timeframe. Thanks, Matt Heon

1 0

[Rawhide Test Day] Podman 5.1 Test Day 2024-05-23
by Sumantro Mukherjee 22 May '24

22 May '24

Hey Folks! We will be testing Fedora 41 with the new update in Podman 5.1. To ensure a smooth transition, the Podman team and the Quality team of Fedora have decided to host a test day[1]. The idea is for users to test Podman 5.1 on a Fedora Rawhide machine and submit results in the Test Day App[2]. If you have spare cycles or use Podman as a daily driver, it will be great to have some folks try out and report bugs right away. We will want more Karmas on[3] and any following builds for Podman 5.1 [1] https://fedoraproject.org/wiki/Test_Day:2024-05-23_Podman_5.1 [2] https://testdays.fedoraproject.org/events/188 [3] https://bodhi.fedoraproject.org/updates/FEDORA-2024-7d9e4ee012 -- //sumantro Fedora QE TRIED AND PERSONALLY TESTED, ERGO TRUSTED

1 0

Next Cabal meeting on Tuesday, May 21, 2024 at 11:00 a.m. EDT (UTC-5)
by Tom Sweeney 17 May '24

17 May '24

Hey All! In just over 4 days from now, on Tuesday, May 21, 2024, at 11:00 a.m. EDT (UTC-5), we'll be in the Podman Cabal meeting. Currently only one topic about Fedora v5.0 Test Days, and room for more to be added! Agenda with video link: https://hackmd.io/gQCfskDuRLm7iOsWgH2yrg?both#podman#opensource Hope to see you there! t

1 0

Podman v5.0.3 Released
by Matt Heon 10 May '24

10 May '24

Hi all, We've just released Podman 5.0.3. This is a security update that addresses CVE-2024-3727, a vulnerability in the containers/image library which allows attackers to trigger unauthorized registry accesses on behalf of the victim. There are also several small bugfixes included, but most fixes and new features will wait until Podman 5.1, which we expect to release later this month. Thanks, Matt Heon

1 0

logging through up to host / systemd init container
by Leon Fauster 10 May '24

10 May '24

I wonder whats the current best approach for having the logs of one container accumulated at the host layer (journald). The container runs systemd/journald and is started via a quadlet config with LogDriver=passthrough. So, everything that the container spells out via stderr/stdout is captured at the host layer. How to pass everything of the container journald to the host journald? I played with LogTarget in /etc/systemd/system.conf without success. I'd really appreciate any ideas! -- Leon

2 4