Merge pull request #222 from bcgov/chore/more-tuning

Chore/more tuning

Author: joshgamache

.github/workflows/release.yaml

@@ -52,7 +52,7 @@ jobs:
           context: .
           dockerfile_path: Dockerfile
           github_token: ${{ secrets.GITHUB_TOKEN }}
-          chart_version: 3.0.0
+          chart_version: 3.1.0
           is_chart_release: true
   cas-airflow-dag-trigger-build:
     needs: [release]

docs/deployment-troubleshooting.md

@@ -0,0 +1,35 @@
+# Airflow upgrade deployment troubleshooting
+
+If DAG tasks are failing for no apparent reason or are giving 401/403 or authentication errors, it may be due to the `cas-airflow-admin` user not updating correctly. You can attempt to fix this by cycling the Airflow pods, or by deleting the `cas-airflow-admin` user and recreating it.
+
+## Cycling the Airflow pods
+
+To cycle the Airflow pods, you can use the following command (replace `airflow-dev` with the namespace you are using):
+
+```bash
+oc -n airflow-dev rollout restart deployment cas-airflow-api-server
+oc -n airflow-dev rollout restart deployment cas-airflow-dag-processor
+oc -n airflow-dev rollout restart deployment cas-airflow-scheduler
+```
+
+This will restart the Airflow pods, which may resolve the issue. If not, delete and recreate the `cas-airflow-admin` user as below.
+
+## Recreating the `cas-airflow-admin` user
+
+### Delete the `cas-airflow-admin` user
+
+To delete the `cas-airflow-admin` user, connect to the terminal of one of the airflow pods' (ie. `cas-airflow-api-server-yyyyy-zzzzz`) and run the following command:
+
+```bash
+airflow users delete --username "cas-airflow-admin"
+```
+
+### Create a new `cas-airflow-admin` user
+
+To recreate the `cas-airflow-admin` user, you can use the following command:
+
+```bash
+airflow users create --username "cas-airflow-admin" --firstname admin --lastname user --role Admin --email admin@example.com
+```
+
+After this you will be prompted to enter a password for the `cas-airflow-admin` user. Use the password set in the `airflow-default-user-password` secret.

helm/cas-airflow-postgres-cluster/values.yaml

@@ -13,6 +13,7 @@ postgresCluster:
     requests:
       cpu: 100m
       memory: 256Mi
+  storageSize: 5Gi
 
   # The "users" value(s) is passed to the crunch postgres operator to create the database.
   # See http://access.crunchydata.com/documentation/postgres-operator/latest/tutorials/basic-setup/user-management

helm/cas-airflow/Chart.yaml

@@ -1,7 +1,7 @@
 apiVersion: v2
 name: cas-airflow
 type: application
-version: 3.0.0 # Changing this requires updating the image tag in .github/workflows/release.yaml
+version: 3.1.0 # Changing this requires updating the image tag in .github/workflows/release.yaml
 appVersion: 3.1.2 # The airflow version
 description: Helm chart to deploy cas' flavour of airflow, compatible with OpenShift 4. This chart uses the vanilla airflow chart and adds cas' own templates and values.
 icon: https://www.astronomer.io/static/airflowNewA.png

helm/cas-airflow/values.yaml

@@ -115,10 +115,7 @@ airflow:
     resources:
       requests:
         cpu: 50m
-        memory: 700Mi
-      limits:
-        cpu: 500m
-        memory: 2Gi
+        memory: 400Mi
     defaultUser:
       enabled: true
       role: Admin
@@ -146,9 +143,9 @@ airflow:
   # Airflow scheduler settings
   scheduler:
     resources:
-      limits:
-        cpu: 500m
-        memory: 1.3Gi
+      requests:
+        cpu: 50m
+        memory: 600Mi
     extraVolumes:
       - name: cas-airflow-dynamic-dags
         persistentVolumeClaim:
@@ -163,9 +160,9 @@ airflow:
   # Airflow dag processor settings
   dagProcessor:
     resources:
-      limits:
-        cpu: 500m
-        memory: 1Gi
+      requests:
+        cpu: 50m
+        memory: 800Mi
     extraVolumes:
       - name: cas-airflow-dynamic-dags
         persistentVolumeClaim: