Fix Alluxio worker/master/job-server communication to enable S3 access

charts/accumulo/templates/alluxio-master-deployment.yaml

@@ -65,13 +65,20 @@ spec:
 
           # Start master
           /opt/alluxio/bin/alluxio-start.sh master
+
+          # Start job master
+          /opt/alluxio/bin/alluxio-start.sh job_master
 
           # Keep container running and monitor process
           while true; do
             if ! pgrep -f "alluxio.master.AlluxioMaster" > /dev/null; then
               echo "Alluxio master process died, restarting..."
               /opt/alluxio/bin/alluxio-start.sh master
             fi
+            if ! pgrep -f "alluxio.master.AlluxioJobMaster" > /dev/null; then
+              echo "Alluxio job master process died, restarting..."
+              /opt/alluxio/bin/alluxio-start.sh job_master
+            fi
             sleep 30
           done
         ports:
@@ -81,6 +88,9 @@ spec:
         - name: web
           containerPort: 19999
           protocol: TCP
+        - name: job-rpc
+          containerPort: 20001
+          protocol: TCP
         env:
         - name: ALLUXIO_MASTER_HOSTNAME
           valueFrom:

charts/accumulo/templates/alluxio-master-service.yaml

@@ -40,6 +40,10 @@ spec:
     port: 19999
     targetPort: web
     protocol: TCP
+  - name: job-rpc
+    port: 20001
+    targetPort: job-rpc
+    protocol: TCP
   selector:
     {{- include "accumulo.selectorLabels" . | nindent 4 }}
     app.kubernetes.io/component: alluxio-master

charts/accumulo/templates/alluxio-worker-daemonset.yaml

@@ -56,7 +56,7 @@ spec:
         - |
           # Wait for master to be ready
           echo "Waiting for Alluxio master to be ready..."
-          until (echo > /dev/tcp/accumulo-alluxio-master/19998) >/dev/null 2>&1; do
+          until (echo > /dev/tcp/{{include "accumulo.fullname" .}}-alluxio-master/19998) >/dev/null 2>&1; do
             echo "Waiting for master..."
             sleep 5
           done
@@ -71,31 +71,57 @@ spec:
 
           # Start worker
           /opt/alluxio/bin/alluxio-start.sh worker
+
+          # Start job worker
+          /opt/alluxio/bin/alluxio-start.sh job_worker
 
           # Keep container running and monitor process
           while true; do
             if ! pgrep -f "alluxio.worker.AlluxioWorker" > /dev/null; then
               echo "Alluxio worker process died, restarting..."
               /opt/alluxio/bin/alluxio-start.sh worker
             fi
+            if ! pgrep -f "alluxio.worker.AlluxioJobWorker" > /dev/null; then
+              echo "Alluxio job worker process died, restarting..."
+              /opt/alluxio/bin/alluxio-start.sh job_worker
+            fi
             sleep 30
           done
         ports:
         - name: rpc
           containerPort: 29999
           protocol: TCP
         - name: data
-          containerPort: 29999
+          containerPort: 29998
           protocol: TCP
         - name: web
           containerPort: 30000
           protocol: TCP
+        - name: job-rpc
+          containerPort: 30002
+          protocol: TCP
+        - name: job-data
+          containerPort: 30003
+          protocol: TCP
+        - name: job-web
+          containerPort: 30004
+          protocol: TCP
         env:
         - name: POD_IP
           valueFrom:
             fieldRef:
               fieldPath: status.podIP
-        - name: WORKER_HOSTNAME
+        - name: POD_NAME
+          valueFrom:
+            fieldRef:
+              fieldPath: metadata.name
+        - name: POD_NAMESPACE
+          valueFrom:
+            fieldRef:
+              fieldPath: metadata.namespace
+        - name: ALLUXIO_WORKER_HOSTNAME
+          value: "$(POD_NAME).{{ include "accumulo.fullname" . }}-alluxio-worker.$(POD_NAMESPACE).svc.cluster.local"
+        - name: ALLUXIO_JOB_WORKER_HOSTNAME
           value: "$(POD_NAME).{{ include "accumulo.fullname" . }}-alluxio-worker.$(POD_NAMESPACE).svc.cluster.local"
         - name: ALLUXIO_MASTER_HOSTNAME
           value: {{include "accumulo.fullname" .}}-alluxio-master

charts/accumulo/templates/alluxio-worker-service.yaml

@@ -26,15 +26,32 @@ metadata:
     app.kubernetes.io/component: alluxio-worker
 spec:
   clusterIP: None  # headless
+  publishNotReadyAddresses: true  # Enable DNS for all pods including not ready
   ports:
     - name: rpc
       port: 29999
       targetPort: rpc
       protocol: TCP
+    - name: data
+      port: 29998
+      targetPort: data
+      protocol: TCP
     - name: web
       port: 30000
       targetPort: web
       protocol: TCP
+    - name: job-rpc
+      port: 30002
+      targetPort: job-rpc
+      protocol: TCP
+    - name: job-data
+      port: 30003
+      targetPort: job-data
+      protocol: TCP
+    - name: job-web
+      port: 30004
+      targetPort: job-web
+      protocol: TCP
   selector:
     {{- include "accumulo.selectorLabels" . | nindent 4 }}
     app.kubernetes.io/component: alluxio-worker

charts/accumulo/templates/configmap.yaml

@@ -289,14 +289,24 @@ data:
     alluxio.master.hostname={{ include "accumulo.fullname" . }}-alluxio-master
     alluxio.master.rpc.port=19998
     alluxio.master.web.port=19999
-    alluxio.worker.hostname={{ include "accumulo.fullname" . }}-alluxio-worker
+    # Worker hostname set via ALLUXIO_WORKER_HOSTNAME environment variable
+    alluxio.worker.bind.host=0.0.0.0
     alluxio.worker.port=29999
     alluxio.worker.web.port=30000
-    alluxio.worker.data.port=29999
+    alluxio.worker.data.port=29998
     alluxio.worker.rpc.port=29999
     alluxio.worker.rpc.bind.host=0.0.0.0
     alluxio.worker.data.bind.host=0.0.0.0
     alluxio.worker.web.bind.host=0.0.0.0
+
+    ## Job worker configuration
+    alluxio.job.master.hostname={{ include "accumulo.fullname" . }}-alluxio-master
+    alluxio.job.master.rpc.port=20001
+    # Job worker hostname set via ALLUXIO_JOB_WORKER_HOSTNAME environment variable
+    alluxio.job.worker.bind.host=0.0.0.0
+    alluxio.job.worker.rpc.port=30002
+    alluxio.job.worker.data.port=30003
+    alluxio.job.worker.web.port=30004
 
     ## Memory and storage configuration
     alluxio.worker.memory.size={{ .Values.alluxio.properties.alluxio.worker.memory.size }}