feat: HBase Listener integration (#639)

* output listener refs

* add listener refs to config map

* changelog

* changed logging statement

* integration test

* changed listener class designation

* update listener class docs

* working tests

* wip: experimental command

* cleanup start command

* added hostname/port to hbase-env.sh

* wip: will move startup logic to hbase-entrypoint.sh

* cleaned up/reverted previous wip

* cleaned up/reverted previous wip II

* correct callout

* use pvcs for externally-reachable endpoints

* added comment

* docs

* resolve listener-class across role-groups/role

* regenerate nix

* updated tokio and nix packages

* added listener-relevant config settings

* Try a different tack at reconfiguring hbase

* formatting and partial clean-up

* formatting and partial clean-up

* replaced listener-class enum with string

* tweak test cases

* add generic properties for listener endpoints

* write endpoints to hbase-site instead of dedicated config map

* remove unused error defs

* remove unused error def/function

* fix some tests inline with 2.6.1, add UI port type

* use listener service resolved in entrypoint script rather than passing through as as argument

* updated env-vars based on patch changes

* fix smoke test

* fixed orphaned-resources test

* fixed opa test

* Update CHANGELOG.md

Co-authored-by: Malte Sander <[email protected]>

* Update docs/modules/hbase/pages/usage-guide/listenerclass.adoc

Co-authored-by: Malte Sander <[email protected]>

* review feedback

* added constant

* replace expects with snafu error

* correct property usage to match latest image changes

* provide bound address to match defaults used in patch

* Update CHANGELOG.md

Co-authored-by: Nick <[email protected]>

* clarified error

* removed unnecessary configs and added trouble-shooting doc page

* renamed headless service

* use ephemeral volumes for master/rs

* fixed test

* added missing pull policy directive

---------

Co-authored-by: Natalie Klestrup Röijezon <[email protected]>
Co-authored-by: Malte Sander <[email protected]>
Co-authored-by: Nick <[email protected]>

Author: 4 people

CHANGELOG.md

@@ -4,6 +4,7 @@
 
 ### Added
 
+- BREAKING: Add listener support for HBase ([#639]).
 - Adds new telemetry CLI arguments and environment variables ([#652]).
   - Use `--file-log-max-files` (or `FILE_LOG_MAX_FILES`) to limit the number of log files kept.
   - Use `--file-log-rotation-period` (or `FILE_LOG_ROTATION_PERIOD`) to configure the frequency of rotation.
@@ -39,6 +40,7 @@
 - test: HBase 2.4.18 removed ([#659]):
 - Remove operator support for HBase 2.4 including the JMX exporter ([#672]).
 
+[#639]: https://github.com/stackabletech/hbase-operator/pull/639
 [#640]: https://github.com/stackabletech/hbase-operator/pull/640
 [#645]: https://github.com/stackabletech/hbase-operator/pull/645
 [#647]: https://github.com/stackabletech/hbase-operator/pull/647

deploy/helm/hbase-operator/crds/crds.yaml

@@ -73,20 +73,6 @@ spec:
                     hdfsConfigMapName:
                       description: Name of the [discovery ConfigMap](https://docs.stackable.tech/home/nightly/concepts/service_discovery) for an HDFS cluster.
                       type: string
-                    listenerClass:
-                      default: cluster-internal
-                      description: |-
-                        This field controls which type of Service the Operator creates for this HbaseCluster:
-
-                        * cluster-internal: Use a ClusterIP service
-
-                        * external-unstable: Use a NodePort service
-
-                        This is a temporary solution with the goal to keep yaml manifests forward compatible. In the future, this setting will control which [ListenerClass](https://docs.stackable.tech/home/nightly/listener-operator/listenerclass.html) will be used to expose the service, and ListenerClass names will stay the same, allowing for a non-breaking change.
-                      enum:
-                        - cluster-internal
-                        - external-unstable
-                      type: string
                     vectorAggregatorConfigMapName:
                       description: Name of the Vector aggregator [discovery ConfigMap](https://docs.stackable.tech/home/nightly/concepts/service_discovery). It must contain the key `ADDRESS` with the address of the Vector aggregator. Follow the [logging tutorial](https://docs.stackable.tech/home/nightly/tutorials/logging-vector-aggregator) to learn how to configure log aggregation with Vector.
                       nullable: true
@@ -210,6 +196,10 @@ spec:
                         hbaseRootdir:
                           nullable: true
                           type: string
+                        listenerClass:
+                          description: This field controls which [ListenerClass](https://docs.stackable.tech/home/nightly/listener-operator/listenerclass.html) is used to expose this rolegroup.
+                          nullable: true
+                          type: string
                         logging:
                           default:
                             containers: {}
@@ -460,6 +450,10 @@ spec:
                               hbaseRootdir:
                                 nullable: true
                                 type: string
+                              listenerClass:
+                                description: This field controls which [ListenerClass](https://docs.stackable.tech/home/nightly/listener-operator/listenerclass.html) is used to expose this rolegroup.
+                                nullable: true
+                                type: string
                               logging:
                                 default:
                                   containers: {}
@@ -691,6 +685,10 @@ spec:
                         hbaseRootdir:
                           nullable: true
                           type: string
+                        listenerClass:
+                          description: This field controls which [ListenerClass](https://docs.stackable.tech/home/nightly/listener-operator/listenerclass.html) is used to expose this rolegroup.
+                          nullable: true
+                          type: string
                         logging:
                           default:
                             containers: {}
@@ -969,6 +967,10 @@ spec:
                               hbaseRootdir:
                                 nullable: true
                                 type: string
+                              listenerClass:
+                                description: This field controls which [ListenerClass](https://docs.stackable.tech/home/nightly/listener-operator/listenerclass.html) is used to expose this rolegroup.
+                                nullable: true
+                                type: string
                               logging:
                                 default:
                                   containers: {}
@@ -1228,6 +1230,10 @@ spec:
                         hbaseRootdir:
                           nullable: true
                           type: string
+                        listenerClass:
+                          description: This field controls which [ListenerClass](https://docs.stackable.tech/home/nightly/listener-operator/listenerclass.html) is used to expose this rolegroup.
+                          nullable: true
+                          type: string
                         logging:
                           default:
                             containers: {}
@@ -1478,6 +1484,10 @@ spec:
                               hbaseRootdir:
                                 nullable: true
                                 type: string
+                              listenerClass:
+                                description: This field controls which [ListenerClass](https://docs.stackable.tech/home/nightly/listener-operator/listenerclass.html) is used to expose this rolegroup.
+                                nullable: true
+                                type: string
                               logging:
                                 default:
                                   containers: {}

deploy/helm/hbase-operator/templates/roles.yaml

@@ -77,6 +77,12 @@ rules:
     verbs:
       - create
       - patch
+  - apiGroups:
+      - listeners.stackable.tech
+    resources:
+      - listeners
+    verbs:
+      - get
   - apiGroups:
       - {{ include "operator.name" . }}.stackable.tech
     resources:

docs/modules/hbase/pages/usage-guide/listenerclass.adoc

@@ -1,18 +1,31 @@
 = Service exposition with ListenerClasses
+:description: Configure HBase service exposure using ListenerClasses to control internal and external access for all roles.
 
-Apache HBase offers an API.
-The operator deploys a service called `<name>` (where `<name>` is the name of the HbaseCluster) through which HBase can be reached.
-
-This service can have either the `cluster-internal` or `external-unstable` type.
-`external-stable` is not supported for HBase at the moment.
-Read more about the types in the xref:concepts:service-exposition.adoc[service exposition] documentation at platform level.
-
-This is how the listener class is configured:
+The operator deploys a xref:listener-operator:listener.adoc[Listener] for each Master, Regionserver and Restserver pod.
+They all default to only being accessible from within the Kubernetes cluster, but this can be changed by setting `.spec.{masters,regionServers,restServers}.config.listenerClass`:
 
 [source,yaml]
 ----
 spec:
-  clusterConfig:
-    listenerClass: cluster-internal  # <1>
+  masters:
+    config:
+      listenerClass: external-unstable  # <1>
+  regionServers:
+    config:
+      listenerClass: external-unstable
+  restServers:
+    config:
+      listenerClass: external-unstable
+----
+<1> Specify one of `external-stable`, `external-unstable`, `cluster-internal` (the default setting is `cluster-internal`).
+This can be set separately for all roles.
+
+Listener endpoints are written to `hbase-site.xml` like this:
+
+[source,xml]
+----
+  <property>
+    <name>hbase.listener.endpoint</name>
+    <value>172.19.0.3:32445</value>
+  </property>
 ----
-<1> The default `cluster-internal` setting.

docs/modules/hbase/pages/usage-guide/troubleshooting.adoc

@@ -0,0 +1,20 @@
+= Trouble-shooting
+:description: Tips and guidelines to help trouble-shoot problems with running HBase.
+
+To add more useful information to stack-traces caused by RPC issues, xref:operators:overrides.adoc[podOverrides] can be used to replace the default Netty implementation:
+
+[source,yaml]
+----
+masters:
+  roleGroups:
+    default:
+      configOverrides:
+        hbase-site.xml:
+          hbase.rpc.client.impl: "org.apache.hadoop.hbase.ipc.BlockingRpcClient"
+regionServers:
+  roleGroups:
+    default:
+      configOverrides:
+        hbase-site.xml:
+          hbase.rpc.client.impl: "org.apache.hadoop.hbase.ipc.BlockingRpcClient"
+----

docs/modules/hbase/partials/nav.adoc

@@ -13,6 +13,7 @@
 ** xref:hbase:usage-guide/hbck2.adoc[]
 ** xref:hbase:usage-guide/snapshot-export.adoc[]
 ** xref:hbase:usage-guide/adls.adoc[]
+** xref:hbase:usage-guide/troubleshooting.adoc[]
 ** xref:hbase:usage-guide/operations/index.adoc[]
 *** xref:hbase:usage-guide/operations/cluster-operations.adoc[]
 *** xref:hbase:usage-guide/operations/pod-placement.adoc[]