From 1b4fa966be923cfb90b835b50fb0016a612df362 Mon Sep 17 00:00:00 2001
From: sauagarwa <sauagarw@redhat.com>
Date: Tue, 3 Dec 2024 14:16:56 -0500
Subject: [PATCH 1/2] Reverting back to using jobs for fixing race condition.
 The issue has been fixed in rhods 2.16

---
 .../templates/accelerator-profile.yaml          |  2 +-
 .../templates/serving-service-setup.yaml        | 17 +++++++++++++++++
 2 files changed, 18 insertions(+), 1 deletion(-)

diff --git a/charts/all/llm-serving-service/templates/accelerator-profile.yaml b/charts/all/llm-serving-service/templates/accelerator-profile.yaml
index 7a6be47c..972f03e3 100644
--- a/charts/all/llm-serving-service/templates/accelerator-profile.yaml
+++ b/charts/all/llm-serving-service/templates/accelerator-profile.yaml
@@ -12,4 +12,4 @@ spec:
   tolerations:
   - effect: NoSchedule
     key: odh-notebook
-    operator: Exists
\ No newline at end of file
+    operator: Exists
diff --git a/charts/all/llm-serving-service/templates/serving-service-setup.yaml b/charts/all/llm-serving-service/templates/serving-service-setup.yaml
index e5f115bd..68546cf9 100644
--- a/charts/all/llm-serving-service/templates/serving-service-setup.yaml
+++ b/charts/all/llm-serving-service/templates/serving-service-setup.yaml
@@ -1,4 +1,21 @@
 ---
+<<<<<<< HEAD
+=======
+apiVersion: dashboard.opendatahub.io/v1
+kind: AcceleratorProfile
+metadata:
+  name: nvidia-gpu
+  namespace: redhat-ods-applications
+spec:
+  displayName: NVIDIA GPU
+  enabled: true
+  identifier: nvidia.com/gpu
+  tolerations:
+  - effect: NoSchedule
+    key: odh-notebook
+    operator: Exists
+---
+>>>>>>> 37d2514 (Reverting back to using jobs for fixing race condition. The issue has been fixed in rhods 2.16)
 apiVersion: batch/v1
 kind: Job
 metadata:

From eef58184e8f5350c330f691733747d41cc24c191 Mon Sep 17 00:00:00 2001
From: sauagarwa <sauagarw@redhat.com>
Date: Tue, 3 Dec 2024 17:41:19 -0500
Subject: [PATCH 2/2] Remove minio dependency

---
 .../templates/model-pvc.yaml                  | 11 +++++++
 .../templates/serving-service-setup.yaml      | 32 +++++--------------
 2 files changed, 19 insertions(+), 24 deletions(-)
 create mode 100644 charts/all/llm-serving-service/templates/model-pvc.yaml

diff --git a/charts/all/llm-serving-service/templates/model-pvc.yaml b/charts/all/llm-serving-service/templates/model-pvc.yaml
new file mode 100644
index 00000000..34c31812
--- /dev/null
+++ b/charts/all/llm-serving-service/templates/model-pvc.yaml
@@ -0,0 +1,11 @@
+kind: PersistentVolumeClaim
+apiVersion: v1
+metadata:
+  name: model-pvc
+spec:
+  accessModes:
+    - ReadWriteOnce
+  resources:
+    requests:
+      storage: 50Gi
+  volumeMode: Filesystem
\ No newline at end of file
diff --git a/charts/all/llm-serving-service/templates/serving-service-setup.yaml b/charts/all/llm-serving-service/templates/serving-service-setup.yaml
index 68546cf9..75589764 100644
--- a/charts/all/llm-serving-service/templates/serving-service-setup.yaml
+++ b/charts/all/llm-serving-service/templates/serving-service-setup.yaml
@@ -1,21 +1,3 @@
----
-<<<<<<< HEAD
-=======
-apiVersion: dashboard.opendatahub.io/v1
-kind: AcceleratorProfile
-metadata:
-  name: nvidia-gpu
-  namespace: redhat-ods-applications
-spec:
-  displayName: NVIDIA GPU
-  enabled: true
-  identifier: nvidia.com/gpu
-  tolerations:
-  - effect: NoSchedule
-    key: odh-notebook
-    operator: Exists
----
->>>>>>> 37d2514 (Reverting back to using jobs for fixing race condition. The issue has been fixed in rhods 2.16)
 apiVersion: batch/v1
 kind: Job
 metadata:
@@ -63,7 +45,7 @@ spec:
                   - vllm.entrypoints.openai.api_server
                 env:
                   - name: HF_HOME
-                    value: /tmp/hf_home
+                    value: /cache
                   - name: HF_TOKEN
                     valueFrom:
                       secretKeyRef:
@@ -91,12 +73,13 @@ spec:
               - autoSelect: true
                 name: vLLM
             volumes:
-              - emptyDir:
+              - name: shm
+                emptyDir:
                   medium: Memory
                   sizeLimit: 2Gi
-                name: shm
-              - emptyDir: {}
-                name: cache
+              - name: cache
+                persistentVolumeClaim:
+                  claimName: model-pvc
           EOF
           cat << EOF | oc apply -f-
           apiVersion: serving.kserve.io/v1beta1
@@ -157,9 +140,10 @@ spec:
               echo -n 'openshift-ai initialized';echo
               echo -n 'Waiting for dscinitialization/default-dsci to initialize'
               echo
-              oc wait --for=jsonpath='{.status.phase}'=Ready --timeout=600s -n redhat-ods-operator dscinitialization/default-dsci
+              oc wait --for=jsonpath='{.status.phase}'=Ready --timeout=900s -n redhat-ods-operator dscinitialization/default-dsci
               sleep 10
               echo -n 'dscinitialization/default-dsci initialized';echo
+              sleep 120
           command:
             - /bin/bash
           image: image-registry.openshift-image-registry.svc:5000/openshift/tools:latest