coreos-status February 2026

coreos-status@lists.fedoraproject.org

1 participants
2 discussions

[CoreOS] testing Stream: Zincati Upgrade Stall and Workaround
by Michael Armijo 26 Feb '26

26 Feb '26

A recent issue was identified in the Fedora CoreOS *testing* stream where nodes attempting to upgrade may become stuck. This is due to a missing fedora-coreos.stream annotation in the update metadata, which causes Zincati (versions < 0.0.31) to fail the upgrade process. This issue *did not* reach *stable*. Full details: https://github.com/coreos/fedora-coreos-tracker/issues/2112 While the rollout was paused, nodes on the testing stream that had already attempted the upgrade may currently be in a "stalled" state. *What’s Happening* Nodes that attempted to upgrade to *43.20260217.2.1*, and failed during staging due to this issue are now unable to upgrade to *43.20260217.2.2*. In other words, once a node entered the failed staging loop on *43.20260217.2.1*, it remains stuck and will not automatically progress to newer releases. *Why This Happened* The root cause was dropping the fedora-coreos.stream annotation. Older Zincati versions (< 0.0.31) still depend on that annotation during upgrade logic. When attempting to stage *43.20260217.2.1* that no longer includes it, Zincati fails and retries indefinitely. This only impacts older Zincati clients upgrading across that boundary. *How to Identify if You Are Affected* If you are on the testing stream, check your rpm-ostree status: rpm-ostree status If you see multiple failed deployment attempts in Zincati, your system is likely affected. *Workaround* The workaround described here resolves the issue: https://github.com/coreos/fedora-coreos-tracker/issues/2112#issuecomment-39… sudo systemctl stop zincati rpm-ostreed sudo rpm-ostree cleanup -p sudo systemctl start zincati # or run sudo /usr/bin/zincati-update-now for immediate upgrade+reboot You can also reboot the node, this action has been shown to clear the stalled state and allow the upgrade to proceed. *If you still see issues* If you believe you’re affected and the workaround does not resolve the issue, please comment on the tracker issue above with: - rpm-ostree status - rpm -q zincati - Relevant journalctl -u zincati logs Thanks to everyone running *testing* and helping us catch issues before they reach *stable*! Michael Armijo - CoreOS

1 0

[CoreOS] Podman 5.8.0 shipping in next stream for database migration
by Michael Armijo 23 Feb '26

23 Feb '26

Hi Everyone! We are shipping Podman 5.8.0 to the Fedora CoreOS next stream release *43.20260218.1.2*. Podman 5.8.0 introduces an automatic migration of the container database from BoltDB to SQLite [1]. Since BoltDB support will be fully removed in Podman 6.0, we are prioritizing this transition now to ensure a smooth experience for all users. We are temporarily "unlocking" next from the testing and stable streams. This allows for soak-testing the migration logic without affecting testing or stable users. *How to help* If you have an older FCOS install, check your current backend with: `podman info --format {{.Host.DatabaseBackend}}` If it says `boltdb`, update to the next stream and reboot (documentation for switching streams <https://docs.fedoraproject.org/en-US/fedora-coreos/update-streams/#_switchi…>). After the update, verify that it now shows sqlite and that your containers are running as expected. *Reporting* Please report successful migrations to the official Test Week ticket <https://forge.fedoraproject.org/quality/tickets/issues/877>. If the migration fails or you encounter issues, please file an upstream bug at github.com/containers/podman or reach out on the FCOS tracker <https://github.com/coreos/fedora-coreos-tracker/>. [1]: https://github.com/coreos/fedora-coreos-tracker/issues/2110 Thanks, Michael Armijo - CoreOS

1 0

2026

2025

2024

2023

2022

2021

2020

2019

coreos-status February 2026