[DPE-5679] Use jmx_exporter 0.15 to avoid conflict with spark-metrics (…

…#57)

.github/workflows/ci.yaml

@@ -73,12 +73,21 @@ jobs:
           - integration-tls
           - integration-azure
           - integration-logs
+        runner:
+          - ubuntu-22.04
+          - [self-hosted, linux, AMD64, X64, large, jammy]
+        exclude:
+          - tox-environments: integration-azure           # Don't test azure twice
+            runner: [self-hosted, linux, AMD64, X64, large, jammy]
+          - tox-environments: integration-charm           # Don't test charm twice
+            runner: [self-hosted, linux, AMD64, X64, large, jammy]
+
     name: ${{ matrix.tox-environments }}
     needs:
       - lint
       - unit-test
       - build
-    runs-on: ubuntu-22.04
+    runs-on: ${{ matrix.runner }}
     timeout-minutes: 120
     steps:
       - name: Checkout
@@ -135,4 +144,3 @@ jobs:
           AZURE_STORAGE_KEY: ${{ secrets.AZURE_STORAGE_KEY }}
           CI_PACKED_CHARMS: ${{ needs.build.outputs.charms }}
         run: tox run -e ${{ matrix.tox-environments }} -- -m '${{ steps.select-tests.outputs.mark_expression }}'
-

metadata.yaml

@@ -23,7 +23,7 @@ resources:
   spark-history-server-image:
     type: oci-image
     description: OCI image for Spark History Server
-    upstream-source: ghcr.io/canonical/charmed-spark@sha256:75a01dfca493b5a457fc7d3258daba3e9891f0408f0097f1fd189100d5de4891 # spark-version: 3.4.2 revision: 3a877ab9facaf0a648f529ec1628a3d38c37208e 
+    upstream-source: ghcr.io/canonical/charmed-spark@sha256:1d9949dc7266d814e6483f8d9ffafeff32f66bb9939e0ab29ccfd9d5003a583a # spark-version: 3.4.2 revision: c9d47a39f3c2a214dbe4bd0b88dfcc6c37b651b6 
 
 requires:
   s3-credentials:

src/core/workload.py

@@ -42,7 +42,7 @@ def jmx_prometheus_javaagent(self):
 
         Used for scraping and exposing mBeans of a JMX target.
         """
-        return self.lib_path / "jmx_prometheus_javaagent-0.20.0.jar"
+        return self.lib_path / "jmx_prometheus_javaagent-0.15.0.jar"
 
     @property
     def jmx_prometheus_config(self):

src/workload.py

@@ -45,15 +45,17 @@ def envs(self):
 
         self._envs = self.from_env(self.read(self.ENV_FILE)) if self.exists(self.ENV_FILE) else {}
 
-        self._envs["SPARK_DAEMON_JAVA_OPTS"] = (
-            f"-javaagent:{self.paths.jmx_prometheus_javaagent}={JMX_EXPORTER_PORT}:{self.paths.jmx_prometheus_config}"
-        )
-
         return self._envs
 
     @property
     def _spark_history_server_layer(self):
         """Return a dictionary representing a Pebble layer."""
+        self.set_environment(
+            {
+                "SPARK_DAEMON_JAVA_OPTS": f"-javaagent:{self.paths.jmx_prometheus_javaagent}={JMX_EXPORTER_PORT}:{self.paths.jmx_prometheus_config}"
+            }
+        )
+
         layer = {
             "summary": "spark history server layer",
             "description": "pebble config layer for spark history server",

tests/integration/setup/run_spark_job.sh

@@ -1,3 +1,7 @@
 #!/bin/bash
 
-spark-client.spark-submit --username hello --conf spark.kubernetes.executor.request.cores=0.1 --class org.apache.spark.examples.SparkPi local:///opt/spark/examples/jars/spark-examples_2.12-3.4.2.jar 100
+spark-client.spark-submit --username hello --conf spark.kubernetes.executor.request.cores=0.1 --class org.apache.spark.examples.SparkPi local:///opt/spark/examples/jars/spark-examples_2.12-3.4.2.jar 1000
+echo "Print logs"
+kubectl logs -l spark-version=3.4.2
+echo "Kubectl get pods -A"
+kubectl get pods -A

tests/integration/setup/setup_spark_azure.sh

@@ -18,6 +18,7 @@ spark-client.service-account-registry create --username hello \
     --conf spark.hadoop.fs.azure.account.key.$account_name.dfs.core.windows.net=$secret_key \
     --conf spark.eventLog.enabled=true \
     --conf spark.eventLog.dir=$folder \
-    --conf spark.history.fs.logDirectory=$folder
+    --conf spark.history.fs.logDirectory=$folder \
+    --conf spark.kubernetes.container.image=ghcr.io/canonical/charmed-spark@sha256:1d9949dc7266d814e6483f8d9ffafeff32f66bb9939e0ab29ccfd9d5003a583a
 
 spark-client.service-account-registry get-config --username hello

tests/integration/test_charm_azure.py

@@ -174,13 +174,14 @@ async def test_build_and_deploy(ops_test: OpsTest, charm_versions, azure_credent
             apps = json.loads(
                 urllib.request.urlopen(f"http://{address}:18080/api/v1/applications").read()
             )
-        except Exception:
+        except Exception as e:
+            logger.warning(f"Exception e: {e}")
             apps = []
 
         if len(apps) > 0:
             break
         else:
-            sleep(10)
+            sleep(30)
 
     assert len(apps) == 1