feat: Support setting TLS certificate lifetimes (#722)

* feat: Support setting TLS certificate lifetimes * Update CHANGELOG.md * chore: bump op-rs * cargo update -p rustls
stackabletech · Dec 4, 2024 · 95def6b · 95def6b
1 parent 7e78aa2
commit 95def6b
Show file tree

Hide file tree

Showing 9 changed files with 62 additions and 30 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -4,6 +4,11 @@ All notable changes to this project will be documented in this file.
 
 ## [Unreleased]
 
+### Added
+
+- The lifetime of auto generated TLS certificates is now configurable with the role and roleGroup
+  config property `requestedSecretLifetime`. This helps reducing frequent Pod restarts ([#722]).
+
 ### Fixed
 
 - Fix OIDC endpoint construction in case the `rootPath` does not have a trailing slash ([#718]).
@@ -13,6 +18,7 @@ All notable changes to this project will be documented in this file.
 
 [#717]: https://github.com/stackabletech/nifi-operator/pull/717
 [#718]: https://github.com/stackabletech/nifi-operator/pull/718
+[#722]: https://github.com/stackabletech/nifi-operator/pull/722
 
 ## [24.11.0] - 2024-11-18
 

diff --git a/Cargo.lock b/Cargo.lock
diff --git a/Cargo.toml b/Cargo.toml
@@ -25,13 +25,13 @@ serde = { version = "1.0", features = ["derive"] }
 serde_json = "1.0"
 serde_yaml = "0.9"
 snafu = "0.8"
-stackable-operator = { git = "https://github.com/stackabletech/operator-rs.git", tag = "stackable-operator-0.82.0" }
+stackable-operator = { git = "https://github.com/stackabletech/operator-rs.git", tag = "stackable-operator-0.83.0" }
 strum = { version = "0.26", features = ["derive"] }
 tokio = { version = "1.40", features = ["full"] }
 tracing = "0.1"
 url = { version = "2.5.2" }
 xml-rs = "0.8"
 
-# [patch."https://github.com/stackabletech/operator-rs.git"]
-# stackable-operator = { git = "https://github.com/stackabletech//operator-rs.git", branch = "main" }
+#[patch."https://github.com/stackabletech/operator-rs.git"]
+#stackable-operator = { git = "https://github.com/stackabletech//operator-rs.git", branch = "main" }
 # stackable-operator = { path = "../operator-rs/crates/stackable-operator" }
diff --git a/deploy/helm/nifi-operator/crds/crds.yaml b/deploy/helm/nifi-operator/crds/crds.yaml
@@ -358,6 +358,10 @@ spec:
                               nullable: true
                               type: boolean
                           type: object
+                        requestedSecretLifetime:
+                          description: Request secret (currently only autoTls certificates) lifetime from the secret operator, e.g. `7d`, or `30d`. Please note that this can be shortened by the `maxCertificateLifetime` setting on the SecretClass issuing the TLS certificate.
+                          nullable: true
+                          type: string
                         resources:
                           default:
                             cpu:
@@ -815,6 +819,10 @@ spec:
                                     nullable: true
                                     type: boolean
                                 type: object
+                              requestedSecretLifetime:
+                                description: Request secret (currently only autoTls certificates) lifetime from the secret operator, e.g. `7d`, or `30d`. Please note that this can be shortened by the `maxCertificateLifetime` setting on the SecretClass issuing the TLS certificate.
+                                nullable: true
+                                type: string
                               resources:
                                 default:
                                   cpu:

diff --git a/rust/crd/src/lib.rs b/rust/crd/src/lib.rs
@@ -411,10 +411,16 @@ pub struct NifiConfig {
     /// 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>,
+
+    /// Request secret (currently only autoTls certificates) lifetime from the secret operator, e.g. `7d`, or `30d`.
+    /// Please note that this can be shortened by the `maxCertificateLifetime` setting on the SecretClass issuing the TLS certificate.
+    #[fragment_attrs(serde(default))]
+    pub requested_secret_lifetime: Option<Duration>,
 }
 
 impl NifiConfig {
-    pub const NIFI_SENSITIVE_PROPS_KEY: &'static str = "NIFI_SENSITIVE_PROPS_KEY";
+    // Auto TLS certificate lifetime
+    const DEFAULT_NODE_SECRET_LIFETIME: Duration = Duration::from_days_unchecked(7);
 
     pub fn default_config(cluster_name: &str, role: &NifiRole) -> NifiConfigFragment {
         NifiConfigFragment {
@@ -458,6 +464,7 @@ impl NifiConfig {
             },
             affinity: get_affinity(cluster_name, role),
             graceful_shutdown_timeout: Some(DEFAULT_NODE_GRACEFUL_SHUTDOWN_TIMEOUT),
+            requested_secret_lifetime: Some(Self::DEFAULT_NODE_SECRET_LIFETIME),
         }
     }
 }

diff --git a/rust/operator-binary/src/controller.rs b/rust/operator-binary/src/controller.rs
@@ -112,6 +112,9 @@ pub struct Ctx {
 #[strum_discriminants(derive(IntoStaticStr))]
 #[allow(clippy::enum_variant_names)]
 pub enum Error {
+    #[snafu(display("missing secret lifetime"))]
+    MissingSecretLifetime,
+
     #[snafu(display("NifiCluster object is invalid"))]
     InvalidNifiCluster {
         source: error_boundary::InvalidObject,
@@ -1253,6 +1256,9 @@ async fn build_node_rolegroup_statefulset(
         .context(MetadataBuildSnafu)?
         .build();
 
+    let requested_secret_lifetime = merged_config
+        .requested_secret_lifetime
+        .context(MissingSecretLifetimeSnafu)?;
     let nifi_cluster_name = nifi.name_any();
     pod_builder
         .metadata(metadata)
@@ -1301,6 +1307,7 @@ async fn build_node_rolegroup_statefulset(
                     &build_reporting_task_service_name(&nifi_cluster_name),
                 ],
                 SecretFormat::TlsPkcs12,
+                &requested_secret_lifetime,
             )
             .context(SecuritySnafu)?,
         )

diff --git a/rust/operator-binary/src/reporting_task/mod.rs b/rust/operator-binary/src/reporting_task/mod.rs
@@ -24,10 +24,15 @@
 //!
 use std::collections::BTreeMap;
 
+use crate::security::{
+    authentication::{NifiAuthenticationConfig, STACKABLE_ADMIN_USERNAME},
+    build_tls_volume,
+};
 use snafu::{OptionExt, ResultExt, Snafu};
 use stackable_nifi_crd::{
     NifiCluster, NifiRole, APP_NAME, HTTPS_PORT, HTTPS_PORT_NAME, METRICS_PORT,
 };
+use stackable_operator::time::Duration;
 use stackable_operator::{
     builder::{
         self,
@@ -50,11 +55,6 @@ use stackable_operator::{
     utils::cluster_info::KubernetesClusterInfo,
 };
 
-use crate::security::{
-    authentication::{NifiAuthenticationConfig, STACKABLE_ADMIN_USERNAME},
-    build_tls_volume,
-};
-
 use super::controller::{build_recommended_labels, NIFI_UID};
 
 const REPORTING_TASK_CERT_VOLUME_NAME: &str = "tls";
@@ -359,6 +359,10 @@ fn build_reporting_task_job(
                 REPORTING_TASK_CERT_VOLUME_NAME,
                 vec![],
                 SecretFormat::TlsPem,
+                // The certificate is only used for the REST API call, so a short lifetime is sufficient.
+                // There is no correct way to configure this job since it's an implementation detail.
+                // Also it will be dropped when support for 1.x is removed.
+                &Duration::from_days_unchecked(1),
             )
             .context(SecretVolumeBuildFailureSnafu)?,
         )

diff --git a/rust/operator-binary/src/security/mod.rs b/rust/operator-binary/src/security/mod.rs
@@ -1,6 +1,7 @@
 use snafu::{ResultExt, Snafu};
 use stackable_nifi_crd::NifiCluster;
 use stackable_operator::client::Client;
+use stackable_operator::time::Duration;
 use stackable_operator::{builder::pod::volume::SecretFormat, k8s_openapi::api::core::v1::Volume};
 
 pub mod authentication;
@@ -42,6 +43,14 @@ pub fn build_tls_volume(
     volume_name: &str,
     service_scopes: Vec<&str>,
     secret_format: SecretFormat,
+    requested_secret_lifetime: &Duration,
 ) -> Result<Volume> {
-    tls::build_tls_volume(nifi, volume_name, service_scopes, secret_format).context(TlsSnafu)
+    tls::build_tls_volume(
+        nifi,
+        volume_name,
+        service_scopes,
+        secret_format,
+        requested_secret_lifetime,
+    )
+    .context(TlsSnafu)
 }
diff --git a/rust/operator-binary/src/security/tls.rs b/rust/operator-binary/src/security/tls.rs
@@ -1,12 +1,12 @@
+use crate::security::authentication::STACKABLE_TLS_STORE_PASSWORD;
 use snafu::{ResultExt, Snafu};
 use stackable_nifi_crd::NifiCluster;
+use stackable_operator::time::Duration;
 use stackable_operator::{
     builder::pod::volume::{SecretFormat, SecretOperatorVolumeSourceBuilder, VolumeBuilder},
     k8s_openapi::api::core::v1::Volume,
 };
 
-use crate::security::authentication::STACKABLE_TLS_STORE_PASSWORD;
-
 pub const KEYSTORE_VOLUME_NAME: &str = "keystore";
 pub const KEYSTORE_NIFI_CONTAINER_MOUNT: &str = "/stackable/keystore";
 pub const TRUSTSTORE_VOLUME_NAME: &str = "truststore";
@@ -26,6 +26,7 @@ pub(crate) fn build_tls_volume(
     volume_name: &str,
     service_scopes: Vec<&str>,
     secret_format: SecretFormat,
+    requested_secret_lifetime: &Duration,
 ) -> Result<Volume> {
     let mut secret_volume_source_builder =
         SecretOperatorVolumeSourceBuilder::new(nifi.server_tls_secret_class());
@@ -44,6 +45,7 @@ pub(crate) fn build_tls_volume(
                 .with_node_scope()
                 .with_pod_scope()
                 .with_format(secret_format)
+                .with_auto_tls_cert_lifetime(*requested_secret_lifetime)
                 .build()
                 .context(TlsCertSecretClassVolumeBuildSnafu)?,
         )