strelok89 commented on code in PR #3854:
URL: https://github.com/apache/celeborn/pull/3854#discussion_r4249205945


##########
charts/celeborn/values.yaml:
##########
@@ -126,6 +131,36 @@ master:
   - name: celeborn-ratis
     mountPath: /mnt/celeborn_ratis
 
+  # -- Startup probe for Celeborn master containers. It holds off the liveness 
probe while the
+  # master waits for the DNS records of its peers, which can take a while if a 
replica is slow to
+  # be scheduled. Set to `null` to disable.
+  startupProbe:

Review Comment:
   Agreed. A startup probe gates readiness the same way a readiness probe does, 
so it hit the same cycle the README described for readiness. The chart now 
refuses `master.startupProbe` or `master.readinessProbe` with more than one 
master unless `master.podManagementPolicy: Parallel`. The default stays unset, 
so existing StatefulSets still upgrade, since the field is immutable. A fresh 
default install renders three masters with no probes, and both failure cases 
have tests.
   



##########
charts/celeborn/values.yaml:
##########
@@ -263,6 +304,33 @@ worker:
   - name: disk4
     mountPath: /mnt/disk4
 
+  # -- Startup probe for Celeborn worker containers. Disabled by default, 
since no liveness probe
+  # is configured by default either.
+  startupProbe:
+    # httpGet:
+    #   path: /healthz
+    #   port: metrics
+
+  # -- Liveness probe for Celeborn worker containers. Disabled by default: 
worker `/healthz`
+  # fails while the worker is not registered with the master, so a liveness 
probe on it would
+  # restart every worker during a master outage. See README.md.
+  livenessProbe:
+    # httpGet:
+    #   path: /healthz
+    #   port: metrics
+
+  # -- Readiness probe for Celeborn worker containers. `/healthz` reports 
healthy only once the
+  # worker is initialized and registered with the master, which keeps a 
rolling update from
+  # advancing to the next worker too early. Set to `null` to disable.
+  readinessProbe:
+    httpGet:
+      path: /healthz

Review Comment:
   Right, 0.7.0 doesn't serve `/healthz`. All probes are now opt-in. The README 
explains why and gives a working configuration for images that serve the 
endpoint. Turning them on by default can follow once the chart's `appVersion` 
points at a release that serves `/healthz`.
   



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to