You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In case of rollbacks or non-default boots, Zincati will immediately try to re-upgrade to a (possibly bad) locally available deployment. This is because Zincati sees that an update is available in the update-graph, but it doesn't know that it had already tried it and for some reason the system booted back into an earlier deployment.
To avoid this, Zincati should look for deployments already available locally and filter them out from possible targets in the received update-graph. That will avoid immediate auto-updates both in auto-rollback and manual-boot cases.
This will only cover bad deployments as long as they are still locally available on the system. In case of garbage collection of deployments, that piece of history will be gone and Zincati may try again to auto-update into them again. This is fine for the short-term, and in the longer term it would be maybe covered in conjunction with coreos/rpm-ostree#1813.
The text was updated successfully, but these errors were encountered:
This is a followup to coreos/fedora-coreos-tracker#247.
In case of rollbacks or non-default boots, Zincati will immediately try to re-upgrade to a (possibly bad) locally available deployment. This is because Zincati sees that an update is available in the update-graph, but it doesn't know that it had already tried it and for some reason the system booted back into an earlier deployment.
To avoid this, Zincati should look for deployments already available locally and filter them out from possible targets in the received update-graph. That will avoid immediate auto-updates both in auto-rollback and manual-boot cases.
This will only cover bad deployments as long as they are still locally available on the system. In case of garbage collection of deployments, that piece of history will be gone and Zincati may try again to auto-update into them again. This is fine for the short-term, and in the longer term it would be maybe covered in conjunction with coreos/rpm-ostree#1813.
The text was updated successfully, but these errors were encountered: