diff --git a/.github/workflows/code-scans.yaml b/.github/workflows/code-scans.yaml
index d6aaecd2..14046c9d 100644
--- a/.github/workflows/code-scans.yaml
+++ b/.github/workflows/code-scans.yaml
@@ -52,7 +52,7 @@ jobs:
         continue-on-error: true
         with:
           scan-type: 'image'
-          image-ref: 'vllm/vllm-openai-cpu:v0.21.0'
+          image-ref: 'vllm/vllm-openai-cpu:v0.24.0'
           severity: 'HIGH,CRITICAL'
           exit-code: 0
           format: 'table'
diff --git a/core/helm-charts/vllm/values.yaml b/core/helm-charts/vllm/values.yaml
index cf384277..264806d7 100644
--- a/core/helm-charts/vllm/values.yaml
+++ b/core/helm-charts/vllm/values.yaml
@@ -25,7 +25,7 @@ image:
   # Uncomment the following line to set desired image pull policy if needed, as one of Always, IfNotPresent, Never.
   # pullPolicy: ""
   # Overrides the image tag whose default is the chart appVersion.
-  tag: "v0.21.0"
+  tag: "v0.24.0"
 
 imagePullSecrets: []
 nameOverride: ""
diff --git a/core/helm-charts/vllm/xeon-values.yaml b/core/helm-charts/vllm/xeon-values.yaml
index 2e7f2beb..5f5c23ca 100644
--- a/core/helm-charts/vllm/xeon-values.yaml
+++ b/core/helm-charts/vllm/xeon-values.yaml
@@ -27,7 +27,7 @@ max_model_len: 8192
 # vLLM CPU image configuration
 image:
   repository: vllm/vllm-openai-cpu
-  tag: "v0.21.0"
+  tag: "v0.24.0"
   pullPolicy: IfNotPresent
 
 # Node affinity for Xeon inference nodes
@@ -59,6 +59,7 @@ modelConfigs:
       HF_HUB_DISABLE_XET: "1"
       VLLM_CPU_SGL_KERNEL: "1"
       USER: "vllm"
+      TRITON_CACHE_DIR: "/tmp/.triton"
     extraCmdArgs:
       [
         "--block-size",
@@ -91,6 +92,7 @@ modelConfigs:
       VLLM_CPU_SGL_KERNEL: "1"
       HF_HUB_DISABLE_XET: "1"
       USER: "vllm"
+      TRITON_CACHE_DIR: "/tmp/.triton"
     extraCmdArgs:
       [
         "--block-size",
@@ -123,6 +125,7 @@ modelConfigs:
       VLLM_CPU_SGL_KERNEL: "1"
       HF_HUB_DISABLE_XET: "1"
       USER: "vllm"
+      TRITON_CACHE_DIR: "/tmp/.triton"
     extraCmdArgs:
       [
         "--block-size",
@@ -153,6 +156,7 @@ modelConfigs:
       VLLM_CPU_SGL_KERNEL: "1"
       HF_HUB_DISABLE_XET: "1"
       USER: "vllm"
+      TRITON_CACHE_DIR: "/tmp/.triton"
     extraCmdArgs:
       [
         "--block-size",
@@ -183,6 +187,7 @@ modelConfigs:
       VLLM_CPU_SGL_KERNEL: "1"
       HF_HUB_DISABLE_XET: "1"
       USER: "vllm"
+      TRITON_CACHE_DIR: "/tmp/.triton"
     extraCmdArgs:
       [
         "--block-size",
@@ -215,6 +220,7 @@ modelConfigs:
       VLLM_CPU_SGL_KERNEL: "1"
       HF_HUB_DISABLE_XET: "1"
       USER: "vllm"
+      TRITON_CACHE_DIR: "/tmp/.triton"
     extraCmdArgs:
       [
         "--block-size",
@@ -249,6 +255,7 @@ modelConfigs:
       HF_HUB_DISABLE_XET: "1"
       LOGNAME: "vllm"
       USER: "vllm"
+      TRITON_CACHE_DIR: "/tmp/.triton"
     extraCmdArgs:
       [
         "--block-size",
@@ -281,6 +288,7 @@ defaultModelConfigs:
     VLLM_CPU_SGL_KERNEL: "1"
     HF_HUB_DISABLE_XET: "1"
     USER: "vllm"
+    TRITON_CACHE_DIR: "/tmp/.triton"
   extraCmdArgs:
     [
       "--block-size",
diff --git a/core/roles/inference-tools/tasks/main.yml b/core/roles/inference-tools/tasks/main.yml
index 12801161..51ae0eec 100644
--- a/core/roles/inference-tools/tasks/main.yml
+++ b/core/roles/inference-tools/tasks/main.yml
@@ -19,7 +19,7 @@
   ansible.builtin.pip:
     name: kubernetes
     state: present
-    extra_args: "--break-system-packages"
+    extra_args: "--break-system-packages --ignore-installed"
   become: true
   when: pip_install_result is failed
   tags: always
diff --git a/core/scripts/vllm-quickstart/README.md b/core/scripts/vllm-quickstart/README.md
index 5b113814..3a0977c3 100644
--- a/core/scripts/vllm-quickstart/README.md
+++ b/core/scripts/vllm-quickstart/README.md
@@ -116,7 +116,7 @@ The `models.json` file contains all configuration:
 ```json
 {
   "docker": {
-    "image": "vllm/vllm-openai-cpu:v0.21.0",
+    "image": "vllm/vllm-openai-cpu:v0.24.0",
     "port": "8000:8000",
     "environment": { ... },
     "volumes": [ ... ]
diff --git a/core/scripts/vllm-quickstart/models.json b/core/scripts/vllm-quickstart/models.json
index fa0f3acb..330bbe3f 100644
--- a/core/scripts/vllm-quickstart/models.json
+++ b/core/scripts/vllm-quickstart/models.json
@@ -1,6 +1,6 @@
 {
   "docker": {
-    "image": "vllm/vllm-openai-cpu:v0.21.0",
+    "image": "vllm/vllm-openai-cpu:v0.24.0",
     "port": "8000:8000",
     "environment": {
       "VLLM_CPU_SGL_KERNEL": "1",