The VM list stays quick with hundreds of VMs
Same list, less waiting: with 489 VMs, sorting by State went from 911 to 350 ms and the page froze 6.7% of the time during live updates instead of 23.9%.
Rancher UI team
Harvester and Longhorn UI changes
Each change has its own page with before and after images of the user flow, and an optional video. Open a card to see the flow; the links go to the pull request and the issue.
Harvester · reviewed
PRs from the Harvester team that we review and test.
Saving an unchanged VM no longer asks for a restart
With the PCI Devices add-on on, saving a VM without changes asked to restart it (10 of 10 times in our test); with the fix it saves quietly (0 of 10), and adding or removing devices still works.
Longhorn · developed by us
Changes we built in the Longhorn UI.
Choose how many NVMe-TCP I/O queues a v2 volume uses
Create Volume gets a per-volume I/O queue count for v2 Block Device volumes (0 = global setting), and volume details show it.
Schedule snapshot groups with a recurring job
A new Snapshot Group recurring job task with no retention (Retain locked, a warning, "-" in lists); it waits on the backend change.