Search results

[BUG] Upstream :server-vulkan floating image is broken (undefined libggml-vulkan.so symbol -> CPU fallback); pin a known-good tag for the AMD runtime

bugSomething isn't workingSomething isn't working

Status: Open.

#725 In defilantech/LLMKube;

· Defilan opened on Jun 16, 2026

[FEATURE] Pin a stable NodePort (endpoint.nodePort + llmkube deploy --node-port)

enhancementNew feature or requestNew feature or request

Status: Open.

#723 In defilantech/LLMKube;

· Defilan opened on Jun 16, 2026

[FEATURE] Explicit model-cache control on deploy/delete (--from-cache/--skip-cache, --save-to-cache)

enhancementNew feature or requestNew feature or request

Status: Open.

#722 In defilantech/LLMKube;

· Defilan opened on Jun 16, 2026

[FEATURE] llmkube scale <name> --replicas N (start/stop a deployment in place)

enhancementNew feature or requestNew feature or request

Status: Open.

#721 In defilantech/LLMKube;

· Defilan opened on Jun 16, 2026

[BUG] InferenceService reconcileService is create-only: endpoint.type/port changes ignored on existing Service

bugSomething isn't workingSomething isn't working

Status: Open.

#720 In defilantech/LLMKube;

· Defilan opened on Jun 15, 2026

[FEATURE] llmkube deploy: show model download progress (CLI appears to hang on large GGUFs)

enhancementNew feature or requestNew feature or request

Status: Open.

#719 In defilantech/LLMKube;

· Defilan opened on Jun 15, 2026

[BUG] checkAcceleratorAvailability ignores gpu.resourceName override, reporting inaccurate AcceleratorReady

bugSomething isn't workingSomething isn't working

Status: Open.

#710 In defilantech/LLMKube;

· Defilan opened on Jun 15, 2026

·  

M2: Sovereign enterprise (B200-ready)

[FEATURE] ROCm 7.2 HIP runtime tier for AMD nodes (per-model opt-in, follow-on)

enhancementNew feature or requestNew feature or request

Status: Open.

#701 In defilantech/LLMKube;

· Defilan opened on Jun 15, 2026

·  

M2: Sovereign enterprise (B200-ready)

[FEATURE] AMD GPU observability: rocm-smi exporter + llama.cpp /metrics + Grafana panel

enhancementNew feature or requestNew feature or request

Status: Open.

#700 In defilantech/LLMKube;

· Defilan opened on Jun 15, 2026

·  

M2: Sovereign enterprise (B200-ready)

[FEATURE] Validated AMD (Vulkan) example InferenceService + benchmark

enhancementNew feature or requestNew feature or request

Status: Open.

#699 In defilantech/LLMKube;

· Defilan opened on Jun 15, 2026

·  

M2: Sovereign enterprise (B200-ready)

[FEATURE] AMD/Strix Halo Kubernetes node enablement runbook (kernel, GTT, /dev/dri device plugin)

documentationImprovements or additions to documentationImprovements or additions to documentation

enhancementNew feature or requestNew feature or request

Status: Open.

#698 In defilantech/LLMKube;

· Defilan opened on Jun 15, 2026

·  

M2: Sovereign enterprise (B200-ready)

[FEATURE] AMD llama.cpp runtime image (Vulkan/RADV)

enhancementNew feature or requestNew feature or request

Status: Open.

#697 In defilantech/LLMKube;

· Defilan opened on Jun 15, 2026

·  

M2: Sovereign enterprise (B200-ready)