Fix ChatQnA manifest test issue on Xeon

.github/workflows/_manifest-e2e.yml

@@ -90,11 +90,16 @@ jobs:
               echo "Validate ${{ inputs.example }} successful!"
             else
               echo "Validate ${{ inputs.example }} failure!!!"
-              .github/workflows/scripts/k8s-utils.sh dump_all_pod_logs $NAMESPACE
+              echo "Check the logs in 'Dump logs when e2e test failed' step!!!"
               exit 1
             fi
           fi
 
+      - name: Dump logs when e2e test failed
+        if: failure()
+        run: |
+          .github/workflows/scripts/k8s-utils.sh dump_all_pod_logs $NAMESPACE
+
       - name: Kubectl uninstall
         if: always()
         run: |

.github/workflows/pr-manifest-e2e.yml

@@ -4,7 +4,7 @@
 name: E2E test with manifests
 
 on:
-  pull_request_target:
+  pull_request:
     branches: ["main", "*rc"]
     types: [opened, reopened, ready_for_review, synchronize] # added `ready_for_review` since draft is skipped
     paths:

ChatQnA/tests/test_manifest_on_gaudi.sh

@@ -111,29 +111,26 @@ function _cleanup_ns() {
 
 function install_and_validate_chatqna_guardrail() {
     echo "Testing manifests chatqna_guardrils"
-    local ns=${NAMESPACE}-gaurdrails
+    local ns=${NAMESPACE}
     _cleanup_ns $ns
     kubectl create namespace $ns
     # install guardrail
     kubectl apply -f chatqna-guardrails.yaml -n $ns
     # Sleep enough time for chatqna_guardrail to be ready
     sleep 60
     if kubectl rollout status deployment -n "$ns" --timeout "$ROLLOUT_TIMEOUT_SECONDS"; then
-        echo "Waiting for cahtqna_guardrail pod ready done!"
+        echo "Waiting for chatqna_guardrail pod ready done!"
     else
         echo "Timeout waiting for chatqna_guardrail pod ready!"
-        _cleanup_ns $ns
         exit 1
     fi
 
     # validate guardrail
     validate_chatqna $ns chatqna-guardrails
     local ret=$?
     if [ $ret -ne 0 ]; then
-        _cleanup_ns $ns
         exit 1
     fi
-    _cleanup_ns $ns
 }
 
 if [ $# -eq 0 ]; then
@@ -162,7 +159,6 @@ case "$1" in
             exit $ret
         fi
         pushd ChatQnA/kubernetes/intel/hpu/gaudi/manifest
-        set +e
         install_and_validate_chatqna_guardrail
         popd
         ;;

ChatQnA/tests/test_manifest_on_xeon.sh

@@ -40,7 +40,7 @@ function get_end_point() {
 function validate_chatqna() {
     local ns=$1
     local log=$2
-    max_retry=20
+    max_retry=10
     # make sure microservice retriever-usvc is ready
     # try to curl retriever-svc for max_retry times
     test_embedding=$(python3 -c "import random; embedding = [random.uniform(-1, 1) for _ in range(768)]; print(embedding)")
@@ -111,29 +111,26 @@ function _cleanup_ns() {
 
 function install_and_validate_chatqna_guardrail() {
     echo "Testing manifests chatqna_guardrils"
-    local ns=${NAMESPACE}-gaurdrails
+    local ns=${NAMESPACE}
     _cleanup_ns $ns
     kubectl create namespace $ns
     # install guardrail
     kubectl apply -f chatqna-guardrails.yaml -n $ns
     # Sleep enough time for chatqna_guardrail to be ready
     sleep 60
     if kubectl rollout status deployment -n "$ns" --timeout "$ROLLOUT_TIMEOUT_SECONDS"; then
-        echo "Waiting for cahtqna_guardrail pod ready done!"
+        echo "Waiting for chatqna_guardrail pod ready done!"
     else
         echo "Timeout waiting for chatqna_guardrail pod ready!"
-        _cleanup_ns $ns
         exit 1
     fi
 
     # validate guardrail
     validate_chatqna $ns chatqna-guardrails
     local ret=$?
     if [ $ret -ne 0 ]; then
-        _cleanup_ns $ns
         exit 1
     fi
-    _cleanup_ns $ns
 }
 
 if [ $# -eq 0 ]; then
@@ -162,7 +159,6 @@ case "$1" in
             exit $ret
         fi
         pushd ChatQnA/kubernetes/intel/cpu/xeon/manifest
-        set +e
         install_and_validate_chatqna_guardrail
         popd
         ;;