Add multiple Dockerfiles for various Gitea runner profiles and update host vars

ansible/host_vars/ci-01.lab.alexpires.me.yml

@@ -115,7 +115,7 @@ windows_gitea_runner_podman_runners:
     container_name: "gitea-runner-cicd-base"
     image: "local/gitea-runner:cicd-base"
     dockerfile_src: "Dockerfile.cicd-base"
-    labels: "cicd-base:host"
+    labels: "linux:host,cicd-base"
     config_file: "D:\\GiteaRunner\\cicd-base\\config.yaml"
     data_dir: "D:\\GiteaRunner\\cicd-base\\data"
     log_dir: "D:\\Logs\\GiteaRunner\\cicd-base"
@@ -124,7 +124,7 @@ windows_gitea_runner_podman_runners:
     container_name: "gitea-runner-infra-tools"
     image: "local/gitea-runner:infra-tools"
     dockerfile_src: "Dockerfile.infra-tools"
-    labels: "infra-tools:host"
+    labels: "linux:host,infra-tools"
     config_file: "D:\\GiteaRunner\\infra-tools\\config.yaml"
     data_dir: "D:\\GiteaRunner\\infra-tools\\data"
     log_dir: "D:\\Logs\\GiteaRunner\\infra-tools"
@@ -133,7 +133,7 @@ windows_gitea_runner_podman_runners:
     container_name: "gitea-runner-security-scan"
     image: "local/gitea-runner:security-scan"
     dockerfile_src: "Dockerfile.security-scan"
-    labels: "security-scan:host"
+    labels: "linux:host,security-scan"
     config_file: "D:\\GiteaRunner\\security-scan\\config.yaml"
     data_dir: "D:\\GiteaRunner\\security-scan\\data"
     log_dir: "D:\\Logs\\GiteaRunner\\security-scan"
@@ -142,7 +142,7 @@ windows_gitea_runner_podman_runners:
     container_name: "gitea-runner-cpp-build"
     image: "local/gitea-runner:cpp-build"
     dockerfile_src: "Dockerfile.cpp-build"
-    labels: "cpp-build:host"
+    labels: "linux:host,cpp-build"
     config_file: "D:\\GiteaRunner\\cpp-build\\config.yaml"
     data_dir: "D:\\GiteaRunner\\cpp-build\\data"
     log_dir: "D:\\Logs\\GiteaRunner\\cpp-build"
@@ -151,7 +151,7 @@ windows_gitea_runner_podman_runners:
     container_name: "gitea-runner-go-build"
     image: "local/gitea-runner:go-build"
     dockerfile_src: "Dockerfile.go-build"
-    labels: "go-build:host"
+    labels: "linux:host,go-build"
     config_file: "D:\\GiteaRunner\\go-build\\config.yaml"
     data_dir: "D:\\GiteaRunner\\go-build\\data"
     log_dir: "D:\\Logs\\GiteaRunner\\go-build"
@@ -160,7 +160,7 @@ windows_gitea_runner_podman_runners:
     container_name: "gitea-runner-image-build"
     image: "local/gitea-runner:image-build"
     dockerfile_src: "Dockerfile.image-build"
-    labels: "image-build:host"
+    labels: "linux:host,image-build"
     config_file: "D:\\GiteaRunner\\image-build\\config.yaml"
     data_dir: "D:\\GiteaRunner\\image-build\\data"
     log_dir: "D:\\Logs\\GiteaRunner\\image-build"

ansible/host_vars/gpu-01.lab.alexpires.me.yml

@@ -119,14 +119,6 @@ llama_server_devices:
 
 llama_server_models:
   # TODO: replace revision values with pinned commit SHAs for strict reproducibility.
-  - name: "gpt-oss-20b-q4-0"
-    model_id: "unsloth/gpt-oss-20b-GGUF"
-    quant: "Q4_0"
-    revision: "main"
-  - name: "gpt-oss-20b-q8-0"
-    model_id: "unsloth/gpt-oss-20b-GGUF"
-    quant: "Q8_0"
-    revision: "main"
   - name: "gemma-4-12b-q8-0"
     model_id: "unsloth/gemma-4-12b-it-GGUF"
     quant: "Q8_0"
@@ -145,27 +137,17 @@ llama_server_models:
     revision: "main"
     preset:
       ctx-size: 98304
-  - name: "qwen3-5-9b-iq4-xs"
-    model_id: "unsloth/Qwen3.5-9B-GGUF"
-    quant: "IQ4_XS"
+  - name: "gpt-oss-20b-q4-0"
+    model_id: "unsloth/gpt-oss-20b-GGUF"
+    quant: "Q4_0"
     revision: "main"
-    preset:
-      ctx-size: 98304
-  - name: "qwen3-5-9b-q8-0"
-    model_id: "unsloth/Qwen3.5-9B-GGUF"
+  - name: "gpt-oss-20b-q8-0"
+    model_id: "unsloth/gpt-oss-20b-GGUF"
     quant: "Q8_0"
     revision: "main"
-    preset:
-      ctx-size: 98304
-  - name: "qwen3-5-27b-mtp-iq4-xs"
-    model_id: "unsloth/Qwen3.5-27B-MTP-GGUF"
-    quant: "IQ4_XS"
-    revision: "main"
-    preset:
-      ctx-size: 98304
-  - name: "qwen3-5-35b-a3b-ud-iq4-xs"
-    model_id: "unsloth/Qwen3.5-35B-A3B-GGUF"
-    quant: "UD-IQ4_XS"
+  - name: "qwen3-5-27b-q4-k-m"
+    model_id: "unsloth/Qwen3.5-27B-GGUF"
+    quant: "Q4_K_M"
     revision: "main"
     preset:
       ctx-size: 98304
@@ -175,9 +157,15 @@ llama_server_models:
     revision: "main"
     preset:
       ctx-size: 98304
-  - name: "qwen3-6-27b-iq4-xs"
+  - name: "qwen3-5-9b-q8-0"
+    model_id: "unsloth/Qwen3.5-9B-GGUF"
+    quant: "Q8_0"
+    revision: "main"
+    preset:
+      ctx-size: 98304
+  - name: "qwen3-6-27b-q4-k-m"
     model_id: "unsloth/Qwen3.6-27B-GGUF"
-    quant: "IQ4_XS"
+    quant: "Q4_K_M"
     revision: "main"
     preset:
       ctx-size: 98304
@@ -187,24 +175,6 @@ llama_server_models:
     revision: "main"
     preset:
       ctx-size: 98304
-  - name: "qwen3-6-27b-mtp-ud-q4-0-k-xl"
-    model_id: "unsloth/Qwen3.6-27B-MTP-GGUF"
-    quant: "UD-Q4_K_XL"
-    revision: "main"
-    preset:
-      ctx-size: 98304
-  - name: "qwen3-6-27b-mtp-ud-q4-xl"
-    model_id: "unsloth/Qwen3.6-27B-MTP-GGUF"
-    quant: "UD-Q4_K_XL"
-    revision: "main"
-    preset:
-      ctx-size: 98304
-  - name: "qwen3-6-35b-a3b-mtp-ud-iq4-xs"
-    model_id: "unsloth/Qwen3.6-35B-A3B-MTP-GGUF"
-    quant: "UD-IQ4_XS"
-    revision: "main"
-    preset:
-      ctx-size: 98304
   - name: "qwen3-6-35b-a3b-mtp-ud-q4-m"
     model_id: "unsloth/Qwen3.6-35B-A3B-MTP-GGUF"
     quant: "UD-Q4_K_M"
@@ -217,28 +187,6 @@ llama_server_models:
     revision: "main"
     preset:
       ctx-size: 98304
-  - name: "qwen3-6-35b-a3b-claude47-distilled-iq4-xs"
-    model_id: "lordx64/Qwen3.6-35B-A3B-Claude-4.7-Opus-Reasoning-Distilled-IQ4_XS-GGUF"
-    quant: "IQ4_XS"
-    revision: "main"
-    preset:
-      ctx-size: 98304
-  - name: "ministral-3-14b-instruct-iq4-xs"
-    model_id: "unsloth/Ministral-3-14B-Instruct-2512-GGUF"
-    quant: "IQ4_XS"
-    revision: "main"
-    preset:
-      ctx-size: 98304
-  - name: "granite-4-1-30b-iq4-xs"
-    model_id: "unsloth/granite-4.1-30b-GGUF"
-    quant: "IQ4_XS"
-    revision: "main"
-    preset:
-      ctx-size: 98304
-  - name: "devstral-small-2-24b-instruct-iq4-xs"
-    model_id: "unsloth/Devstral-Small-2-24B-Instruct-2512-GGUF"
-    quant: "IQ4_XS"
-    revision: "main"
 
 llama_server_argv_extra:
   [
@@ -269,6 +217,10 @@ llama_server_argv_extra:
     "-mg",
     1,
     "--mmap",
+    "--spec-type",
+    "draft-mtp",
+    "--spec-draft-n-max",
+    2,
   ]
 
 auto_suspend_enabled: true

ansible/roles/windows_gitea_runner_podman/defaults/main.yml

@@ -9,7 +9,7 @@ windows_gitea_runner_podman_base_dockerfile_src: "Dockerfile.gitea-runner-base"
 windows_gitea_runner_podman_base_image_hash_label: "io.a13labs.gitea_runner.base_dockerfile_hash"
 windows_gitea_runner_podman_image_prefix: "local/gitea-runner"
 windows_gitea_runner_podman_image_hash_label: "io.a13labs.gitea_runner.dockerfile_hash"
-windows_gitea_runner_podman_dockerfile_source_root: "{{ playbook_dir }}/files/dockerfiles"
+windows_gitea_runner_podman_dockerfile_source_root: "{{ playbook_dir }}/files/dockerfiles/linux-runners"
 windows_gitea_runner_podman_build_root: "C:\\ProgramData\\GiteaRunner\\builds"
 windows_gitea_runner_podman_base_build_dir: "{{ windows_gitea_runner_podman_build_root }}\\base"