bump longhorn

Makefile

@@ -6,6 +6,7 @@ KUBECTLFLAGS =
 KUBECTLDIFFFLAGS =
 KUBECTLDIFFPRINTER = highlight --out-format xterm256 --syntax diff
 KUBECTLAPPLYFLAGS = --server-side --force-conflicts
+KUBECTLAPPLYPRUNEFLAGS = --server-side --force-conflicts --prune
 KUBECTLAPPLYPRINTER = sed -E -e 's/(.+configured)/\o033[32m\1\o033[0m/g' -e 's/(.+pruned)/\o033[31m\1\o033[0m/g'
 
 SRC := $(shell find kustomize/ -type f)
@@ -25,6 +26,10 @@ diff: $(PRODOUT)
 apply: $(PRODOUT)
 	$(KUBECTL) $(KUBECTLFLAGS) apply $(KUBECTLAPPLYFLAGS) -l app.kubernetes.io/managed-by=kustomize -f $(PRODOUT) | $(KUBECTLAPPLYPRINTER)
 
+.PHONY: apply-prune
+apply-prune: $(PRODOUT)
+	$(KUBECTL) $(KUBECTLFLAGS) apply $(KUBECTLAPPLYPRUNEFLAGS) -l app.kubernetes.io/managed-by=kustomize -f $(PRODOUT) | $(KUBECTLAPPLYPRINTER)
+
 $(PRODOUT): $(SRC)
 	mkdir -p $(OUT)
 	$(KUSTOMIZE) build $(KUSTOMIZEFLAGS) $(PRODSRC) >$(PRODOUT) || (rm $(PRODOUT); exit 1)

kustomize/bases/litellm/litellm-deployment.yaml

@@ -13,6 +13,16 @@ spec:
       labels:
         app.kubernetes.io/component: proxy
     spec:
+      affinity:
+        nodeAffinity:
+          requiredDuringSchedulingIgnoredDuringExecution:
+            nodeSelectorTerms:
+              - matchExpressions:
+                  # Image only supports amd64
+                  - key: kubernetes.io/arch
+                    operator: In
+                    values:
+                      - amd64
       containers:
         - name: litellm
           image: ghcr.io/berriai/litellm:main-latest

kustomize/bases/longhorn/longhorn-helmchart.yaml

@@ -6,7 +6,7 @@ metadata:
 spec:
   repo: https://charts.longhorn.io
   chart: longhorn
-  version: 1.5.5
+  version: 1.6.2
   targetNamespace: longhorn-system
   set:
     backupTargetCredentialSecret: s3-backupstore-credentials

kustomize/bases/openwebui/openwebui-deployment.yaml

@@ -26,10 +26,10 @@ spec:
               value: http://ollama-server.$(NAMESPACE).svc:11434
           resources:
             requests:
-              cpu: 500m
+              cpu: 250m
               memory: 1Gi
             limits:
-              cpu: 1000m
+              cpu: 500m
               memory: 1Gi
           ports:
             - containerPort: 8080

kustomize/env/prod/configurations/litellm/config.yaml

@@ -1,9 +1,13 @@
 model_list:
-  - model_name: llama3
+  - model_name: llama3-8b
     litellm_params:
       model: ollama_chat/llama3
       api_base: http://ollama-server:11434
-  - model_name: gemma2
+  - model_name: gemma2-9b
     litellm_params:
       model: ollama_chat/gemma2
+      api_base: http://ollama-server:11434
+  - model_name: phi3-3b
+    litellm_params:
+      model: ollama_chat/phi3:mini
       api_base: http://ollama-server:11434