Release notes for 1.18.0

Container images

docker.io/scylladb/scylla-operator:1.18.0

Changes By Kind (since 1.17.1)

Feature

• Support v1alpha1.ScyllaDBDatacenter registration with global ScyllaDB Manager instance (#2590,@rzetelskik)
• Collect StorageClass objects into must-gather archive (#2634,@zimnx)
• Support backup and repair task scheduling with global ScyllaDB Manager for v1alpha1.ScyllaDBDatacenter (#2641,@rzetelskik)
• Migrate v1.ScyllaCluster to in-operator integration with global ScyllaDB Manager instance (#2681,@rzetelskik)
• Support IgnoreDownHosts parameter in ScyllaDB Manager's repair tasks (#2734,@rzetelskik)
• Support v1alpha1.ScyllaDBCluster registration with global ScyllaDB Manager instance (#2738,@rzetelskik)
• Add an identity service for ScyllaDBCluster (#2741,@rzetelskik)
• Delay ScyllaDBDatacenter creation until Service and EndpointSlice controllers are ready (#2762,@zimnx)
• Generate ScyllaDB Manager agent auth tokens for ScyllaDBClusters (#2764,@rzetelskik)
• Support backup and repair task scheduling for v1alpha1.ScyllaDBCluster (#2780,@rzetelskik)
• Mirror ScyllaDBCluster ConfigMaps referred from Volumes (#2789,@zimnx)
• Allow opting out of v1.ScyllaCluster global manager registration (#2793,@czeslavo)

Bug

• Fix potential aggregated conditions' msg and reason overflow (#2635,@czeslavo)
• Add missing Pod watch permission to scyllacluster-member ClusterRole (#2648,@jSherz)
• Fix JUnit and JSON report setup for Ginkgo in test runner (#2657,@rzetelskik)
• Fix NonRetriableError matching (#2672,@rzetelskik)
• Strip ScyllaDB Manager client errors propagated to objects' statuses from non-idemponent information and unify log levels (#2676,@rzetelskik)
• Fix ID propagation to ScyllaDBClusterRegistration's status (#2684,@rzetelskik)
• Fix ScyllaDB Manager task adoption in case of missing owner UID label (#2694,@rzetelskik)
• Revert "Migrate rate limiting queue to typed version" (#2702,@rzetelskik)
• Omit empty value for switch parameters (#2708,@zimnx)
• Use advertised Scylla addresses on cleanup Jobs (#2717,@zimnx)
• Fix kustomization in ci-deploy-release script after manager-controller obsoletion (#2733,@rzetelskik)
• Support ServiceClusterIP node broadcast type of ScyllaDBCluster in endpoint creation for remote DCs (#2748,@rzetelskik)
• Do not install prometheus-operator in OpenShift (#2750,@czeslavo)
• Avoid leading hyphens in truncated hashed names generation (#2763,@rzetelskik)
• Fix extraction of ScyllaDB Manager agent auth token from custom config secrets (#2773,@rzetelskik)
• Mirror ScyllaDBCluster Secrets referred from Volumes (#2782,@czeslavo)
• Await for previous DCs to finish progressing before creating next DC (#2786,@zimnx)
• Fix Grafana container image reference to use a multi-arch index (#2830,@rzetelskik)

Documentation

• Fix GitOps local-csi-driver installation command (#2625,@czeslavo)
• Build v1.17 documentation (#2628,@mflendrich)
• Add documentation and examples for automated multi-dc (#2636,@zimnx)
• Mark 1.17 docs as stable (#2638,@mflendrich)
• Add the 1.17 release and the 1.18 plans (#2640,@mflendrich)
• Rename generic NodeConfig to match docs expectations (#2643,@czeslavo)
• Document ScyllaCluster volume expansion procedure (#2646,@zimnx)
• Add czeslavo to reviewers and approvers (#2651,@czeslavo)
• Add redirections for pages that are moved or deleted (#2675,@ylebi)
• Fix redirect loop in docs (#2678,@czeslavo)
• Recommend using multi-rack clusters in documentation (#2740,@zimnx)
• Do not mention standalone Manager Controller (#2770,@czeslavo)
• Fix grammatical errors and improve wording in replace-node documentation (#2785,@copilot-swe-agent)
• Update deprecated kubectl drain flag from --delete-local-data to --delete-emptydir-data (#2804,@copilot-swe-agent)
• Pin tags to 1.18 (#2806,@czeslavo)
• Add dedicated upgrade guide (#2846,@czeslavo)

Flaky/Failing Test

• Move RemoteKubernetesCluster Secret to dedicated namespace in ScyllaDBCluster webhook e2e (#2667,@zimnx)
• Fix loop devices manipulations in Node Setup E2Es (#2743,@czeslavo)
• Change the default ScyllaDBCluster topology used in E2E test to 3x3x1 to survive rollouts (#2794,@rzetelskik)

Other

• Change the source of the ScyllaDBVersion parameter value in replace e2e (#2578,@grzywin)
• Refactor ScyllaDBCluster controller resources specs (#2622,@zimnx)
• Bump the gomod group with 3 updates (#2633,@dependabot)
• Deduplicate ScyllaDBDatacenter resource test cases for MakeManagedScyllaDBConfig function (#2645,@rzetelskik)
• Bump the gomod group across 1 directory with 13 updates (#2647,@dependabot)
• Add minimal golangci-lint config and fix found issues (#2652,@czeslavo)
• Make E2E suite timeout configurable through an env variable (#2658,@rzetelskik)
• Bump google.golang.org/grpc from 1.72.1 to 1.72.2 in the gomod group (#2659,@dependabot)
• Merge the Dependabot schedules (#2660,@mflendrich)
• Mark generated files in .gitattributes (#2666,@czeslavo)
• Extend ScyllaDBCluster Webhook E2E with update being declined (#2668,@zimnx)
• Enable importas linter (#2669,@czeslavo)
• Bump github.com/go-git/go-git/v5 from 5.16.0 to 5.16.1 in the gomod group (#2673,@dependabot)
• Ensure runtime namespaces are collected regardless of ScyllaDBCluster rollout success (#2679,@zimnx)
• Remove linter from Makefile's update target (#2680,@rzetelskik)
• Update OWNERS to the new structure (#2682,@mflendrich)
• Migrate rate limiting queue to typed version (#2689,@rishabkrish)
• Bump the gomod group with 9 updates (#2690,@dependabot)
• Remove required reviewers from OWNERS' dependency filter (#2691,@rzetelskik)
• Bump the gomod group with 7 updates (#2697,@dependabot)
• Fix time in dependabot's schedule (#2699,@rzetelskik)
• Migrate depracted RateLimitingInterface to TypedRateLimitingInterface (#2703,@rzetelskik)
• Bump the gomod group with 11 updates (#2711,@dependabot)
• Deduplicate existing ScyllaDB Manager integration tests (#2715,@rzetelskik)
• Distinguish the control plane cluster in machinery and E2E framework for multi-datacenter scenarios (#2725,@rzetelskik)
• Add runtime information collection to must-gather (#2728,@zimnx)
• Use ignition container to validate sysctls in e2e (#2729,@zimnx)
• Use ephemeralcontainer to validate prlimits in e2e (#2731,@zimnx)
• Bump the gomod group with 3 updates (#2736,@dependabot)
• Make ScyllaDB Manager deployment conditional in CI release deployment script (#2749,@rzetelskik)
• Use HostIDs instead of broadcasted addresses as node identifier for E2E quorum check (#2751,@zimnx)
• Make the Git references in docs point at the release branch instead of master (#2754,@mflendrich)
• Stop relying on KUBECONFIG_DIR env var in E2E machinery (#2756,@rzetelskik)
• Add per-worker object storage support to E2E framework (#2758,@czeslavo)
• Handle multi-DC object storage buckets in E2E setup scripts (#2759,@czeslavo)
• Fix order of namespace collection in multidc boostrap e2e (#2760,@zimnx)
• Apply e2e workarounds on all clusters (#2771,@zimnx)
• Deduplicate existing global ScyllaDB Manager registration tests for ScyllaDBDatacenter (#2774,@rzetelskik)
• Bump the gomod group with 14 updates (#2783,@dependabot)
• Update ScyllaDB Manager to 3.5.1 (#2790,@rzetelskik)
• Do not perform post-schema-restore redeployment in manual restore procedure E2Es for default ScyllaDB version (#2798,@rzetelskik)
• Bump scylla-monitoring submodule to 4.11.0 (#2799,@czeslavo)
• Update scylla-monitoring to 4.11.1 (#2832,@czeslavo)
• Update ScyllaDB to 2025.1.5 (#2845,@czeslavo)