Latest release
0.1.3
Version 0.1.3 (2026-07-17)
Summary
Version 0.1.3 expands bare-metal diagnostics, improves PXE and remote console reliability, and makes switch automation suitable for batch operations. It also provides clearer hardware reporting and safer job lifecycle handling.
Features
- Added automatic PXE boot mode detection for servers whose mode is unknown, with UEFI-first fallback and persistence of the detected mode.
- Added BMC restart jobs that reset only the management controller without affecting host power.
- Added server diagnostic history and dedicated diagnostic report pages.
- Added memory and disk diagnostic workflows, including Memtest86+ 7.20 as a preboot diagnostic and rescue profile.
- Added force deletion support for active jobs when operational cleanup is required.
Improvements
- Added a per-switch pending queue so batch port actions are accepted and executed sequentially instead of being rejected while the switch is busy.
- Dispatches the next queued action immediately after job completion, failure or cancellation, with deterministic ordering and faster agent polling.
- Allows dashboard actions to be queued independently across different ports on the same switch.
- Improved CPU discovery by removing redundant vendor and trademark text and reporting physical cores and logical threads separately as
C/T. - Improved disk diagnostics with native ATA capacity reporting, post-scan kernel error detection and stricter health gates before performance tests.
- Improved memory test supervision, keepalive reporting, console status messages and result preparation.
- Improved Supermicro HTML5 console login and detection through Redfish, with more reliable console launch and status handling.
- Added clearer boot mode fallback warnings, broken server status indicators and improved server overview text contrast.
- Increased the internal API rate limit for long-running operational flows.
Bug Fixes
- Fixed discovery labelling logical CPU threads as physical cores.
- Fixed rapid switch suspension batches rejecting every action after the first active switch job.
- Fixed queued switch actions waiting for the normal polling interval after terminal job states.
- Fixed Debian automated installs running storage setup too early by moving it to
partman/early_command. - Fixed Proxmox automated installs racing network initialization by retrying DHCP before fetching the answer file.
- Fixed concurrent profile or ISO uninstall jobs cleaning up the same assets.
- Fixed disk tests completing successfully despite kernel-reported storage errors.