Support graceful shutdown

CHANGELOG.md

@@ -7,6 +7,7 @@
 - Default stackableVersion to operator version ([#390]).
 - Support PodDisruptionBudgets ([#407]).
 - Added support for versions 2.1.1, 3.0.1 ([#415]).
+- Support graceful shutdown ([#422]).
 
 ### Changed
 
@@ -29,6 +30,7 @@
 [#396]: https://github.com/stackabletech/superset-operator/pull/396
 [#407]: https://github.com/stackabletech/superset-operator/pull/407
 [#415]: https://github.com/stackabletech/superset-operator/pull/415
+[#422]: https://github.com/stackabletech/superset-operator/pull/422
 
 ## [23.7.0] - 2023-07-14
 

Cargo.toml

@@ -15,12 +15,12 @@ clap = "4.3"
 fnv = "1.0"
 futures = { version = "0.3", features = ["compat"] }
 indoc = "2.0"
+product-config = { git = "https://github.com/stackabletech/product-config.git", tag = "0.6.0" }
 serde = { version = "1.0", features = ["derive"] }
 serde_json = "1.0"
 serde_yaml = "0.9"
 snafu = "0.7"
 stackable-operator = { git = "https://github.com/stackabletech/operator-rs.git", tag = "0.56.0" }
-product-config = { git = "https://github.com/stackabletech/product-config.git", tag = "0.6.0" }
 strum = { version = "0.25", features = ["derive"] }
 tokio = { version = "1.29", features = ["full"] }
 tracing = "0.1"

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

@@ -629,6 +629,10 @@ spec:
                                   type: array
                               type: object
                           type: object
+                        gracefulShutdownTimeout:
+                          description: Time period Pods have to gracefully shut down, e.g. `30m`, `1h` or `2d`. Consult the operator documentation for details.
+                          nullable: true
+                          type: string
                         logging:
                           default:
                             enableVectorAgent: null
@@ -4079,6 +4083,10 @@ spec:
                                         type: array
                                     type: object
                                 type: object
+                              gracefulShutdownTimeout:
+                                description: Time period Pods have to gracefully shut down, e.g. `30m`, `1h` or `2d`. Consult the operator documentation for details.
+                                nullable: true
+                                type: string
                               logging:
                                 default:
                                   enableVectorAgent: null

docs/modules/superset/pages/usage-guide/operations/graceful-shutdown.adoc

@@ -1,7 +1,20 @@
 = Graceful shutdown
 
-Graceful shutdown of Superset nodes is either not supported by the product itself
-or we have not implemented it yet.
+You can configure the graceful shutdown as described in xref:concepts:operations/graceful_shutdown.adoc[].
 
-Outstanding implementation work for the graceful shutdowns of all products where this functionality is relevant is tracked in
-https://github.com/stackabletech/issues/issues/357
+== Nodes
+
+As a default, Superset nodes have `2 minutes` to shut down gracefully.
+
+The Superset node process will receive a `SIGTERM` signal when Kubernetes wants to terminate the Pod.
+It will log the received signal as shown in the log below and initiate a graceful shutdown.
+After the graceful shutdown timeout runs out, and the process still didn't exit, Kubernetes will issue a `SIGKILL` signal.
+
+[source,text]
+----
+superset [2023-11-08 13:14:39 +0000] [206] [INFO] Handling signal: term
+metrics ts=2023-11-08T13:14:39.818Z caller=main.go:553 level=info msg="Received os signal, exiting" signal=terminated
+superset [2023-11-08 13:14:39 +0000] [207] [INFO] Worker exiting (pid: 207)
+superset Loaded your LOCAL configuration at [/stackable/app/pythonpath/superset_config.py]
+superset [2023-11-08 13:14:40 +0000] [206] [INFO] Shutting down: Master
+----

rust/crd/src/lib.rs

@@ -1,5 +1,3 @@
-use std::collections::BTreeMap;
-
 use product_config::flask_app_config_writer::{FlaskAppConfigOptions, PythonType};
 use serde::{Deserialize, Serialize};
 use snafu::{OptionExt, ResultExt, Snafu};
@@ -22,26 +20,30 @@ use stackable_operator::{
     role_utils::{GenericRoleConfig, Role, RoleGroup, RoleGroupRef},
     schemars::{self, JsonSchema},
     status::condition::{ClusterCondition, HasStatusCondition},
+    time::Duration,
 };
+use std::collections::BTreeMap;
 use strum::{Display, EnumIter, EnumString, IntoEnumIterator};
 
-use crate::{affinity::get_affinity, authentication::SupersetAuthentication};
-
 pub mod affinity;
 pub mod authentication;
 pub mod druidconnection;
 
+use crate::{affinity::get_affinity, authentication::SupersetAuthentication};
+
 pub const APP_NAME: &str = "superset";
-pub const CONFIG_DIR: &str = "/stackable/config";
-pub const LOG_CONFIG_DIR: &str = "/stackable/log_config";
-pub const LOG_DIR: &str = "/stackable/log";
+pub const STACKABLE_CONFIG_DIR: &str = "/stackable/config";
+pub const STACKABLE_LOG_CONFIG_DIR: &str = "/stackable/log_config";
+pub const STACKABLE_LOG_DIR: &str = "/stackable/log";
 pub const PYTHONPATH: &str = "/stackable/app/pythonpath";
 pub const SUPERSET_CONFIG_FILENAME: &str = "superset_config.py";
 pub const MAX_LOG_FILES_SIZE: MemoryQuantity = MemoryQuantity {
     value: 10.0,
     unit: BinaryMultiple::Mebi,
 };
 
+const DEFAULT_NODE_GRACEFUL_SHUTDOWN_TIMEOUT: Duration = Duration::from_minutes_unchecked(2);
+
 #[derive(Debug, Snafu)]
 pub enum Error {
     #[snafu(display("unknown Superset role found {role}. Should be one of {roles:?}"))]
@@ -313,6 +315,10 @@ pub struct SupersetConfig {
 
     #[fragment_attrs(serde(default))]
     pub affinity: StackableAffinity,
+
+    /// Time period Pods have to gracefully shut down, e.g. `30m`, `1h` or `2d`. Consult the operator documentation for details.
+    #[fragment_attrs(serde(default))]
+    pub graceful_shutdown_timeout: Option<Duration>,
 }
 
 impl SupersetConfig {
@@ -334,6 +340,7 @@ impl SupersetConfig {
             },
             logging: product_logging::spec::default_logging(),
             affinity: get_affinity(cluster_name, role),
+            graceful_shutdown_timeout: Some(DEFAULT_NODE_GRACEFUL_SHUTDOWN_TIMEOUT),
             ..Default::default()
         }
     }

rust/operator-binary/src/operations/graceful_shutdown.rs

@@ -0,0 +1,26 @@
+use snafu::{ResultExt, Snafu};
+use stackable_operator::builder::PodBuilder;
+use stackable_superset_crd::SupersetConfig;
+
+#[derive(Debug, Snafu)]
+pub enum Error {
+    #[snafu(display("Failed to set terminationGracePeriod"))]
+    SetTerminationGracePeriod {
+        source: stackable_operator::builder::pod::Error,
+    },
+}
+
+pub fn add_graceful_shutdown_config(
+    merged_config: &SupersetConfig,
+    pod_builder: &mut PodBuilder,
+) -> Result<(), Error> {
+    // This must be always set by the merge mechanism, as we provide a default value,
+    // users can not disable graceful shutdown.
+    if let Some(graceful_shutdown_timeout) = merged_config.graceful_shutdown_timeout {
+        pod_builder
+            .termination_grace_period(&graceful_shutdown_timeout)
+            .context(SetTerminationGracePeriodSnafu)?;
+    }
+
+    Ok(())
+}

rust/operator-binary/src/operations/mod.rs

@@ -1 +1,2 @@
+pub mod graceful_shutdown;
 pub mod pdb;

rust/operator-binary/src/product_logging.rs

@@ -14,7 +14,7 @@ use stackable_operator::{
     },
     role_utils::RoleGroupRef,
 };
-use stackable_superset_crd::LOG_DIR;
+use stackable_superset_crd::STACKABLE_LOG_DIR;
 
 #[derive(Snafu, Debug)]
 pub enum Error {
@@ -91,7 +91,7 @@ where
         choice: Some(ContainerLogConfigChoice::Automatic(log_config)),
     }) = logging.containers.get(main_container)
     {
-        let log_dir = format!("{LOG_DIR}/{main_container}");
+        let log_dir = format!("{STACKABLE_LOG_DIR}/{main_container}");
         cm_builder.add_data(
             LOG_CONFIG_FILE,
             create_superset_config(log_config, &log_dir),

rust/operator-binary/src/superset_controller.rs

@@ -1,10 +1,4 @@
 //! Ensures that `Pod`s are configured and running for each [`SupersetCluster`]
-use std::{
-    borrow::Cow,
-    collections::{BTreeMap, HashMap},
-    sync::Arc,
-};
-
 use indoc::formatdoc;
 use product_config::{
     flask_app_config_writer::{self, FlaskAppConfigWriterError},
@@ -34,21 +28,33 @@ use stackable_operator::{
     labels::{role_group_selector_labels, role_selector_labels},
     logging::controller::ReconcilerError,
     product_config_utils::{transform_all_roles_to_config, validate_all_roles_and_groups_config},
-    product_logging::{self, spec::Logging},
+    product_logging::{
+        self,
+        framework::{create_vector_shutdown_file_command, remove_vector_shutdown_file_command},
+        spec::Logging,
+    },
     role_utils::{GenericRoleConfig, RoleGroupRef},
     status::condition::{
         compute_conditions, operations::ClusterOperationsConditionBuilder,
         statefulset::StatefulSetConditionBuilder,
     },
     time::Duration,
+    utils::COMMON_BASH_TRAP_FUNCTIONS,
 };
 use stackable_superset_crd::{
     authentication::SuperSetAuthenticationConfigResolved, Container, SupersetCluster,
     SupersetClusterStatus, SupersetConfig, SupersetConfigOptions, SupersetRole, APP_NAME,
-    CONFIG_DIR, LOG_CONFIG_DIR, LOG_DIR, PYTHONPATH, SUPERSET_CONFIG_FILENAME,
+    PYTHONPATH, STACKABLE_CONFIG_DIR, STACKABLE_LOG_CONFIG_DIR, STACKABLE_LOG_DIR,
+    SUPERSET_CONFIG_FILENAME,
+};
+use std::{
+    borrow::Cow,
+    collections::{BTreeMap, HashMap},
+    sync::Arc,
 };
 use strum::{EnumDiscriminants, IntoStaticStr};
 
+use crate::operations::graceful_shutdown::add_graceful_shutdown_config;
 use crate::{
     config::{self, PYTHON_IMPORTS},
     controller_commons::{self, CONFIG_VOLUME_NAME, LOG_CONFIG_VOLUME_NAME, LOG_VOLUME_NAME},
@@ -202,6 +208,11 @@ pub enum Error {
     FailedToCreatePdb {
         source: crate::operations::pdb::Error,
     },
+
+    #[snafu(display("failed to configure graceful shutdown"))]
+    GracefulShutdown {
+        source: crate::operations::graceful_shutdown::Error,
+    },
 }
 
 type Result<T, E = Error> = std::result::Result<T, E>;
@@ -656,9 +667,9 @@ fn build_server_rolegroup_statefulset(
     superset_cb
         .image_from_product_image(resolved_product_image)
         .add_container_port("http", APP_PORT.into())
-        .add_volume_mount(CONFIG_VOLUME_NAME, CONFIG_DIR)
-        .add_volume_mount(LOG_CONFIG_VOLUME_NAME, LOG_CONFIG_DIR)
-        .add_volume_mount(LOG_VOLUME_NAME, LOG_DIR)
+        .add_volume_mount(CONFIG_VOLUME_NAME, STACKABLE_CONFIG_DIR)
+        .add_volume_mount(LOG_CONFIG_VOLUME_NAME, STACKABLE_LOG_CONFIG_DIR)
+        .add_volume_mount(LOG_VOLUME_NAME, STACKABLE_LOG_DIR)
         .add_env_var_from_secret("ADMIN_USERNAME", secret, "adminUser.username")
         .add_env_var_from_secret("ADMIN_FIRSTNAME", secret, "adminUser.firstname")
         .add_env_var_from_secret("ADMIN_LASTNAME", secret, "adminUser.lastname")
@@ -669,8 +680,8 @@ fn build_server_rolegroup_statefulset(
             "-c".to_string(),
             formatdoc! {"
                 mkdir --parents {PYTHONPATH} && \
-                cp {CONFIG_DIR}/* {PYTHONPATH} && \
-                cp {LOG_CONFIG_DIR}/{LOG_CONFIG_FILE} {PYTHONPATH} && \
+                cp {STACKABLE_CONFIG_DIR}/* {PYTHONPATH} && \
+                cp {STACKABLE_LOG_CONFIG_DIR}/{LOG_CONFIG_FILE} {PYTHONPATH} && \
                 superset db upgrade && \
                 superset fab create-admin \
                     --username \"$ADMIN_USERNAME\" \
@@ -679,18 +690,26 @@ fn build_server_rolegroup_statefulset(
                     --email \"$ADMIN_EMAIL\" \
                     --password \"$ADMIN_PASSWORD\" && \
                 superset init && \
+                {COMMON_BASH_TRAP_FUNCTIONS}
+                {remove_vector_shutdown_file_command}
+                prepare_signal_handlers
                 gunicorn \
                 --bind 0.0.0.0:${{SUPERSET_PORT}} \
                 --worker-class gthread \
                 --threads 20 \
                 --timeout {webserver_timeout} \
                 --limit-request-line 0 \
                 --limit-request-field_size 0 \
-                'superset.app:create_app()'
-            "},
+                'superset.app:create_app()' &
+                wait_for_termination $!
+                {create_vector_shutdown_file_command}",
+            remove_vector_shutdown_file_command =
+                remove_vector_shutdown_file_command(STACKABLE_LOG_DIR),
+            create_vector_shutdown_file_command =
+                create_vector_shutdown_file_command(STACKABLE_LOG_DIR),
+            },
         ])
         .resources(merged_config.resources.clone().into());
-
     let probe = Probe {
         http_get: Some(HTTPGetAction {
             port: IntOrString::Int(APP_PORT.into()),
@@ -708,12 +727,18 @@ fn build_server_rolegroup_statefulset(
     superset_cb.liveness_probe(probe);
 
     pb.add_container(superset_cb.build());
+    add_graceful_shutdown_config(merged_config, &mut pb).context(GracefulShutdownSnafu)?;
 
     let metrics_container = ContainerBuilder::new("metrics")
         .context(InvalidContainerNameSnafu)?
         .image_from_product_image(resolved_product_image)
         .command(vec!["/bin/bash".to_string(), "-c".to_string()])
-        .args(vec!["/stackable/statsd_exporter".to_string()])
+        .args(vec![formatdoc! {"
+            {COMMON_BASH_TRAP_FUNCTIONS}
+            prepare_signal_handlers
+            /stackable/statsd_exporter &
+            wait_for_termination $!
+        "}])
         .add_container_port(METRICS_PORT_NAME, METRICS_PORT)
         .resources(
             ResourceRequirementsBuilder::new()

tests/templates/kuttl/smoke/30-assert.yaml

@@ -11,6 +11,10 @@ apiVersion: apps/v1
 kind: StatefulSet
 metadata:
   name: superset-node-default
+spec:
+  template:
+    spec:
+      terminationGracePeriodSeconds: 120
 status:
   readyReplicas: 1
   replicas: 1