You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: modules/ROOT/pages/kubernetes/openshift.adoc
+10-1Lines changed: 10 additions & 1 deletion
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -7,6 +7,10 @@ SDP operators are certified for the OpenShift platform and can be installed from
7
7
8
8
IMPORTANT: OpenShift installations with FIPS mode enabled are not supported. This is because neither the SDP operators, nor the supported Apache products are FIPS-compliant.
9
9
10
+
IMPORTANT: The operators deployed with OLM might have higher memory requests (and limits) compared to our Helm charts, as we have seen typical production OpenShift clusters being larger than typical non-Openshift clusters.
11
+
The reason is that memory usage of some of our operators scales with the number of objects in your Kubernetes cluster.
12
+
However, you can fine-tune the operator memory requirements using a custom Subscription object as described below, so this note is only about the default resources!
13
+
10
14
== Customizing operator installations
11
15
12
16
As described in the https://github.com/operator-framework/operator-lifecycle-manager/blob/master/doc/design/subscription-config.md[Openshift Subscription documentation{external-link-icon}^] you can configure the deployed operators.
@@ -53,7 +57,12 @@ spec:
53
57
value: kafka-namespace
54
58
----
55
59
56
-
IMPORTANT: The described configuration mechanism currently does not work for secret and listener-operator due to restrictions on deploying DaemonSets.
60
+
=== secret and listener-operator
61
+
62
+
The secret and listener operator are not deployed by OLM directly but rather by a helper tool called `olm-deployer`. This tool is needed to work around some OLM limitations that prevent it from directly installing these operators.
63
+
In this case, `olm-deployer` takes **it's own** resource requests and applies them to all containers on the `DaemonSets` of these operators..
64
+
65
+
This allows you to customize the resources of the DaemonSet in the same way as for the other operators - but doesn't allow individual resources for each container.
Failed Apache Spark applications are reported as successful.
627
+
The entrypoint of the Spark images does not propagate the exit code of the Spark process, so the driver Pod always terminates with exit code `0` and the SparkApplication ends up in the `Succeeded` state even if the application failed.
628
+
Until the fix is released, the driver logs must be checked to determine whether an application actually succeeded.
629
+
Fixed in https://github.com/stackabletech/docker-images/pull/1595[docker-images#1595].
0 commit comments