feat!: Support configuring JVM arguments

CHANGELOG.md

@@ -8,6 +8,11 @@
   config property `requestedSecretLifetime`. This helps reducing frequent Pod restarts ([#598]).
 - Run a `containerdebug` process in the background of each HBase container to collect debugging information ([#605]).
 - Aggregate emitted Kubernetes events on the CustomResources ([#612]).
+- Support configuring JVM arguments ([#620]).
+
+### Removed
+
+- BREAKING: The field `config.hbaseOpts` has been removed. Use JVM argument overrides instead to configure additional JVM arguments ([#620]).
 
 ### Changed
 
@@ -17,6 +22,7 @@
 [#605]: https://github.com/stackabletech/hbase-operator/pull/605
 [#611]: https://github.com/stackabletech/hbase-operator/pull/611
 [#612]: https://github.com/stackabletech/hbase-operator/pull/612
+[#620]: https://github.com/stackabletech/hbase-operator/pull/620
 
 ## [24.11.1] - 2025-01-09
 

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

@@ -95,3 +95,21 @@ The HBaseCluster Stacklet does not support environment variable overrides with t
 
 The HBase Stacklet and operator also support Pod overrides, allowing you to override any property that you can set on a Kubernetes Pod.
 Read the xref:concepts:overrides.adoc#pod-overrides[Pod overrides documentation] to learn more about this feature.
+
+== JVM argument overrides
+
+Stackable operators automatically determine the set of needed JVM arguments, such as memory settings or trust- and keystores.
+Using JVM argument overrides you can configure the JVM arguments xref:concepts:overrides.adoc#jvm-argument-overrides[according to the concepts page].
+
+One thing that is different for Kafka, is that all heap-related arguments will be passed in via the env variable `HBASE_HEAPSIZE`, all the other ones via `HBASE_OPTS`, `HBASE_MASTER_OPTS`, `HBASE_REGIONSERVER_OPTS` and `HBASE_REST_OPTS`.
+The `HBASE_HEAPSIZE` variable is documented as follows in the https://cwiki.apache.org/confluence/display/HADOOP2/Hbase+FAQ+Operations[HBase FAQs]:
+
+> Set the HBASE_HEAPSIZE environment variable in ${HBASE_HOME}/conf/hbase-env.sh if your install needs to run with a larger heap.
+> HBASE_HEAPSIZE is like HADOOP_HEAPSIZE in that its value is the desired heap size in MB.
+> The surrounding '-Xmx' and 'm' needed to make up the maximum heap size java option are added by the hbase start script
+> (See how HBASE_HEAPSIZE is used in the ${HBASE_HOME}/bin/hbase script for clarification).
+
+Looking at `bin/hbase`, you can actually add the `m` suffix to make the unit more clear, the script will detect this https://github.com/apache/hbase/blob/777010361abb203b8b17673d84acf4f7f1d0283a/bin/hbase#L165[here] and work correctly.
+
+Because of this, it is not possible to change `-XmS` and `-XmX` via JVM argument overrides.
+You need to envOverride `HBASE_HEAPSIZE` instead.

rust/crd/src/lib.rs

@@ -21,9 +21,7 @@ use stackable_operator::{
     kube::{runtime::reflector::ObjectRef, CustomResource, ResourceExt},
     product_config_utils::Configuration,
     product_logging::{self, spec::Logging},
-    role_utils::{
-        GenericProductSpecificCommonConfig, GenericRoleConfig, Role, RoleGroup, RoleGroupRef,
-    },
+    role_utils::{GenericRoleConfig, JavaCommonConfig, Role, RoleGroup, RoleGroupRef},
     schemars::{self, JsonSchema},
     status::condition::{ClusterCondition, HasStatusCondition},
     time::Duration,
@@ -51,16 +49,10 @@ pub const HBASE_SITE_XML: &str = "hbase-site.xml";
 pub const SSL_SERVER_XML: &str = "ssl-server.xml";
 pub const SSL_CLIENT_XML: &str = "ssl-client.xml";
 
-pub const HBASE_MANAGES_ZK: &str = "HBASE_MANAGES_ZK";
-pub const HBASE_MASTER_OPTS: &str = "HBASE_MASTER_OPTS";
-pub const HBASE_REGIONSERVER_OPTS: &str = "HBASE_REGIONSERVER_OPTS";
-pub const HBASE_REST_OPTS: &str = "HBASE_REST_OPTS";
-
 pub const HBASE_CLUSTER_DISTRIBUTED: &str = "hbase.cluster.distributed";
 pub const HBASE_ROOTDIR: &str = "hbase.rootdir";
 pub const HBASE_UNSAFE_REGIONSERVER_HOSTNAME_DISABLE_MASTER_REVERSEDNS: &str =
     "hbase.unsafe.regionserver.hostname.disable.master.reversedns";
-pub const HBASE_HEAPSIZE: &str = "HBASE_HEAPSIZE";
 
 pub const HBASE_UI_PORT_NAME_HTTP: &str = "ui-http";
 pub const HBASE_UI_PORT_NAME_HTTPS: &str = "ui-https";
@@ -81,8 +73,6 @@ pub const HBASE_REST_UI_PORT: u16 = 8085;
 // Newer versions use the same port as the UI because Hbase provides it's own metrics API
 pub const METRICS_PORT: u16 = 9100;
 
-pub const JVM_HEAP_FACTOR: f32 = 0.8;
-
 #[derive(Snafu, Debug)]
 pub enum Error {
     #[snafu(display("the role [{role}] is invalid and does not exist in HBase"))]
@@ -137,15 +127,15 @@ pub struct HbaseClusterSpec {
     /// The HBase master process is responsible for assigning regions to region servers and
     /// manages the cluster.
     #[serde(default, skip_serializing_if = "Option::is_none")]
-    pub masters: Option<Role<HbaseConfigFragment>>,
+    pub masters: Option<Role<HbaseConfigFragment, GenericRoleConfig, JavaCommonConfig>>,
 
     /// Region servers hold the data and handle requests from clients for their region.
     #[serde(default, skip_serializing_if = "Option::is_none")]
-    pub region_servers: Option<Role<HbaseConfigFragment>>,
+    pub region_servers: Option<Role<HbaseConfigFragment, GenericRoleConfig, JavaCommonConfig>>,
 
     /// Rest servers provide a REST API to interact with.
     #[serde(default, skip_serializing_if = "Option::is_none")]
-    pub rest_servers: Option<Role<HbaseConfigFragment>>,
+    pub rest_servers: Option<Role<HbaseConfigFragment, GenericRoleConfig, JavaCommonConfig>>,
 }
 
 #[derive(Clone, Debug, Default, Deserialize, Eq, JsonSchema, PartialEq, Serialize)]
@@ -325,7 +315,6 @@ impl HbaseRole {
 
         HbaseConfigFragment {
             hbase_rootdir: None,
-            hbase_opts: None,
             resources,
             logging: product_logging::spec::default_logging(),
             affinity: get_affinity(cluster_name, self, hdfs_discovery_cm_name),
@@ -413,12 +402,13 @@ pub enum Container {
 pub struct HbaseConfig {
     #[serde(default, skip_serializing_if = "Option::is_none")]
     pub hbase_rootdir: Option<String>,
-    #[serde(default, skip_serializing_if = "Option::is_none")]
-    pub hbase_opts: Option<String>,
+
     #[fragment_attrs(serde(default))]
     pub resources: Resources<HbaseStorageConfig, NoRuntimeLimits>,
+
     #[fragment_attrs(serde(default))]
     pub logging: Logging<Container>,
+
     #[fragment_attrs(serde(default))]
     pub affinity: StackableAffinity,
 
@@ -538,7 +528,10 @@ impl HbaseCluster {
         }
     }
 
-    pub fn get_role(&self, role: &HbaseRole) -> Option<&Role<HbaseConfigFragment>> {
+    pub fn get_role(
+        &self,
+        role: &HbaseRole,
+    ) -> Option<&Role<HbaseConfigFragment, GenericRoleConfig, JavaCommonConfig>> {
         match role {
             HbaseRole::Master => self.spec.masters.as_ref(),
             HbaseRole::RegionServer => self.spec.region_servers.as_ref(),
@@ -550,7 +543,7 @@ impl HbaseCluster {
     pub fn get_role_group(
         &self,
         rolegroup_ref: &RoleGroupRef<HbaseCluster>,
-    ) -> Result<&RoleGroup<HbaseConfigFragment, GenericProductSpecificCommonConfig>, Error> {
+    ) -> Result<&RoleGroup<HbaseConfigFragment, JavaCommonConfig>, Error> {
         let role_variant =
             HbaseRole::from_str(&rolegroup_ref.role).with_context(|_| InvalidRoleSnafu {
                 role: rolegroup_ref.role.to_owned(),