rebae on main

NASA-IMPACT · Mar 12, 2024 · 817c222 · 817c222
2 parents 91baa0b + 0ad5e3a
commit 817c222
Show file tree

Hide file tree

Showing 2 changed files with 21 additions and 13 deletions.
diff --git a/.github/workflows/config.py b/.github/workflows/config.py
@@ -76,10 +76,10 @@ def calc_task_manager_resources(task_manager_process_memory):
 
 resource_profile_choice = os.environ.get("RESOURCE_PROFILE")
 task_manager_process_memory_map = {
-    "small": 7168,
-    "medium": 10240,
-    "large": 15360,
-    "xlarge": 20480,
+    "small": 7824,
+    "medium": 9824,
+    "large": 11824,
+    "xlarge": 13824,
 }
 if resource_profile_choice not in list(task_manager_process_memory_map.keys()):
     raise ValueError(
@@ -98,9 +98,9 @@ def calc_task_manager_resources(task_manager_process_memory):
 c.Bake.feedstock_subdir = os.environ.get("FEEDSTOCK_SUBDIR")
 
 c.FlinkOperatorBakery.parallelism = int(os.environ.get("PARALLELISM_OPTION"))
-c.FlinkOperatorBakery.enable_job_archiving = True
+c.FlinkOperatorBakery.enable_job_archiving = False
 c.FlinkOperatorBakery.flink_version = "1.16"
-c.FlinkOperatorBakery.job_manager_resources = {"memory": "1536m", "cpu": 0.3}
+c.FlinkOperatorBakery.job_manager_resources = {"memory": "1280m", "cpu": 0.3}
 c.FlinkOperatorBakery.task_manager_resources = {
     "memory": f"{task_manager_process_memory_map[resource_profile_choice]}m",
     "cpu": 0.3
@@ -121,6 +121,6 @@ def calc_task_manager_resources(task_manager_process_memory):
     "client_kwargs": {"region_name": "us-west-2"},
 }
 
-c.InputCacheStorage.fsspec_class = c.TargetStorage.fsspec_class
-c.InputCacheStorage.fsspec_args = c.TargetStorage.fsspec_args
-c.InputCacheStorage.root_path = f"{BUCKET_PREFIX}/cache/"
+# c.InputCacheStorage.fsspec_class = c.TargetStorage.fsspec_class
+# c.InputCacheStorage.fsspec_args = c.TargetStorage.fsspec_args
+# c.InputCacheStorage.root_path = f"{BUCKET_PREFIX}/cache/"
diff --git a/.github/workflows/job-runner.yaml b/.github/workflows/job-runner.yaml
@@ -26,8 +26,11 @@ on:
         description: 'What auth mode (edl or iamrole) to use when accessing files.'
         required: false
         default: 'iamrole'
+      job_name:
+        description: 'A unique job name (no other existing filnk deployment can have it) so we can inspect metrics easier in Grafana.'
+        required: true
       resource_profile:
-        description: 'jobs have different memory requirements so choose (small[7168M], medium[10240M], large[15360M], xlarge[20480M])'
+        description: 'jobs have different memory requirements so choose (small[7824_MiB], medium[9824_MiB], large[11824_MiB], xlarge[13824_MiB])'
         required: false
         default: 'small'
 
@@ -61,10 +64,11 @@ jobs:
       uses: actions/checkout@v3
 
     - name: Configure AWS credentials
-      uses: aws-actions/configure-aws-credentials@v2
+      uses: aws-actions/configure-aws-credentials@v3
       with:
         role-to-assume: arn:aws:iam::444055461661:role/github-actions-role-eodc
         role-session-name: veda-pforge-run-job
+        role-duration-seconds: 3600
         aws-region: us-west-2
 
     - name: set up python 3.10
@@ -106,6 +110,7 @@ jobs:
           bake \
           --repo=${{ github.event.inputs.repo }} \
           --ref=${{ github.event.inputs.ref }} \
+          --Bake.job_name="${{ github.event.inputs.job_name }}" \
           -f .github/workflows/config.py
 
         # we just use the `inspect` branch to dump runner errors and fail fast
@@ -133,6 +138,7 @@ jobs:
         PARALLELISM_OPTION: ${{ github.event.inputs.parallelism }}
         OUTPUT_BUCKET: ${{ vars.OUTPUT_BUCKET }}
         AUTH_MODE: ${{ github.event.inputs.auth_mode }}
+        AWS_ROLE_ARN: ${{ vars.AWS_ROLE_ARN }}
         RESOURCE_PROFILE: ${{ github.event.inputs.resource_profile }}
 
     - name: cleanup if "pangeo-forge-runner bake" failed
@@ -195,13 +201,15 @@ jobs:
   monitor-job:
     runs-on: ubuntu-latest
     name: monitor job ${{ needs.name-job.outputs.repo_name }}@${{ github.event.inputs.ref }}
+    environment: veda-smce
     needs: [name-job, run-job]
     steps:
       - name: Configure AWS credentials
-        uses: aws-actions/configure-aws-credentials@v2
+        uses: aws-actions/configure-aws-credentials@v3
         with:
           role-to-assume: arn:aws:iam::444055461661:role/github-actions-role-eodc
           role-session-name: veda-pforge-monitor-job
+          role-duration-seconds: 43200  # note this has to match our timeout-minutes below for monitoring
           aws-region: us-west-2
 
       - name: install kubectl
@@ -219,7 +227,7 @@ jobs:
 #
       - name: monitor logs of job manager and report final status
         id: monitorjob
-        timeout-minutes: 240
+        timeout-minutes: 720
         continue-on-error: true
         run: |
           # TODO: this needs to not check the logs but the historyserver status