QuTS hero h6.0 Beta Feedback

QuTS hero h6.0 Beta
1

Overall Performance & UI

  • Visual Refresh: The new UI looks great and feels noticeably snappier than 5.2. Even with a lean setup (primarily using the NAS for ZFS storage and Docker containers), the responsiveness of the web interface is a welcome improvement.
  • Login Stability: I am encountering a recurring bug where the login screen hangs on the loading animation roughly 1 out of 5 times. A hard refresh (Ctrl+Shift+R) resolves it, suggesting a potential timeout or handshake issue with the new authentication backend.

Core Storage & Snapshots

  • ZFS Reliability: The core ZFS implementation continues to be excellent and remains the primary reason I use this platform.
  • Sequential Snapshot Scheduling: Currently, there is no way to group datasets for sequential snapshots. To avoid resource contention, I have to manually stagger schedules (e.g., 4:00, 4:10, 4:20). If the backend is actually smart enough to stagger datasets automatically I have been given no indication of this.
    • Feedback: With the introduction of Immutable (WORM) snapshots, this is more critical than ever. If overlapping schedules cause a system hang, those snapshots cannot be easily cleared to troubleshoot. A “Group/Batch” schedule option would be a major usability win.

Networking & Container Station (cgroups)

  • IPv6 Support: IPv6 functionality still feels incomplete and occasionally inconsistent. I was hoping for more robust parity with IPv4 in this update.
  • Container Station & Kernel Mismatch: While the move to a 6.x kernel is great, the Docker engine version feels outdated by comparison.
  • cgroup v2 Friction: With the kernel update, the shift toward cgroup v2 seems to be causing issues with how Container Station 3.x recognizes storage. Specifically, CLI-created ZFS datasets are no longer recognized by the UI. Forcing all container storage through GUI-created “Shared Folders” suggests the middleware isn’t yet handling the new cgroup hierarchy or manual dataset mounts as flexibly as it did in 5.2.

UPS Management (NUT)

The current NUT (Network UPS Tools) implementation remains a significant pain point:

  • IP Restrictions: We are still limited to 10 manual IP entries for the UPS monitor.
  • Request: Please allow for CIDR notation (e.g., 192.168.0.0/24). Manually managing individual IPs for a home lab is tedious.
  • Modernization: The version of NUT and the reliance on custom scripts feel dated compared to the rest of the OS. I’ve actually offloaded my UPS management to a Raspberry Pi to bypass these restrictions, but I’d prefer to have it handled natively by QuTS hero.

Summary of Suggestions:

  1. Add Sequential Scheduling: Allow users to trigger a queue of snapshots for multiple datasets to prevent resource overlap.
  2. Improve NUT Client: Support subnets/CIDR for the UPS monitoring list and remove the 10-device cap.
  3. Refine Container Station: Update the Docker engine to better align with the 6.x kernel’s cgroup v2 implementation and restore visibility for CLI-created datasets.
1 Like