Release notes for 1.14.0

Container images

docker.io/scylladb/scylla-operator:1.14.0

Changes By Kind (since 1.13.0)

Feature

• Split startup logic out of ScyllaDB container (#1934,@tnozicka)
• Add an E2E test suite for multi-datacenter clusters (#1953,@rzetelskik)
• Add a NodeConfig example for generic environments (#1970,@rzetelskik)
• Add dedicated ServiceAccount for perftune jobs (#1971,@tnozicka)
• Add linux NodeSelector to generic NodeConfig example (#1977,@rzetelskik)
• Add support for forceRedeploymentReason in Helm chart (#2008,@gdubicki)
• Remove IP Address based node replace procedure (#2027,@zimnx)
• Add support for S3 object storage to e2e test runner, framework and set (#2050,@rzetelskik)
• Add scylla-manager-agent readiness probe (#2074,@tnozicka)
• Manage global configs using ScyllaOperatorConfig, add unsupported overrides and extract defaults to a common config (#2081,@tnozicka)
• Deprecate cpuset field in v1.ScyllaCluster (#2088,@zimnx)

Bug

• Skip CSI driver creation for empty SO_CSI_DRIVER_PATH in deploy release script (#1992,@rzetelskik)
• Fix NodeConfig controller handlers to react on related object updates (#2000,@tnozicka)
• Resolve symlinks in NodeConfig's mount paths (#2014,@tnozicka)
• Create a deploy directory per kubeconfig in CI release scripts (#2020,@rzetelskik)
• Update node setup image digest to latest multiplatform (#2022,@rzetelskik)
• Use identical file system for node tuning and keep symlinks working (#2033,@tnozicka)
• Fix infinite loop in tuning between sync timeout and cri client backoff (#2058,@tnozicka)
• Pin Scylla Manager version using sha reference (#2070,@zimnx)
• Ensure CRI calls are not retried in NodeConfig's sync loop and have a timeout (#2073,@tnozicka)
• Fix races in NodeConfig disk test (#2094,@tnozicka)
• Validate that artifacts dir exists (#2099,@tnozicka)
• Drop GinkgoRecover from ginkgo's container nodes (#2100,@rzetelskik)

Documentation

• Update release procedures (#1871,@tnozicka)
• Build v1.13 documentation (#1962,@zimnx)
• Bump minimum versions for v1.13 release (#1979,@zimnx)
• Mark v1.13 as stable version in documentation (#1993,@zimnx)
• Bump documentation dependencies (#1998,@zimnx)
• Make warning about using performance tuning more explicit (#2018,@gdubicki)
• Update supported releases in docs and stop building unsupported branches (#2024,@rzetelskik)
• Remove host networking mentions from documentation and examples (#2087,@mykaul)
• Add a note to documentation about exposeOptions being immutable (#2097,@zimnx)

Flaky/Failing Test

• Fix flakiness of cleanup E2E test (#1999,@zimnx)
• Increase operator cpu request (#2041,@rzetelskik)
• Fix NodeConfig e2e race check with node setup daemon (#2101,@tnozicka)

Other

• Move OperatorOptions flags to method (#1959,@zimnx)
• Add Execute() to OperatorOptions (#1960,@zimnx)
• Update docs dependencies (#1964,@zimnx)
• Extend Handlers' EnqueueAll with enqueing func parameter (#1974,@rzetelskik)
• Use broadcast_address for seeds in external seeds E2E (#1984,@rzetelskik)
• Bump default ScyllaDB version used in E2E's to 6.0.1 (#1991,@zimnx)
• Add proposal for new API v1alpha1.ScyllaDBDatacenter (#1994,@zimnx)
• Bump Scylla Manager to 3.3.0 (#1995,@rzetelskik)
• Add resourceapply helpers for Pods (#2002,@rzetelskik)
• Add resourceapply helpers for PersistentVolumeClaims (#2004,@rzetelskik)
• Bump certifi from 2024.6.2 to 2024.7.4 in /docs (#2006,@dependabot)
• Stop depending on default StorageClass' presence for ScyllaClusters in our manifests, examples, scripts and tests (#2009,@rzetelskik)
• Remove sphinx-scylladb-markdown dependency from docs (#2025,@rzetelskik)
• Bump local-csi-driver to 0.4.1-rc.0 in CI (#2031,@ylebi)
• Update stable version of scylladb/local-csi-driver (#2032,@ylebi)
• Remove pre-rollout check validating ScyllaDB version (#2035,@zimnx)
• Set an explicit key for gcs-service-account-credentials secret in e2e scripts (#2039,@rzetelskik)
• Add duration to certificate creation logs (#2040,@rzetelskik)
• Use canonical CRD locations in helm CRD symlinks (#2044,@tnozicka)
• Make parallelism configurable in e2e lib (#2047,@rzetelskik)
• Remove default annotation from scylladb-local-xfs StorageClass in CI manifests (#2048,@rzetelskik)
• Allow building images from git worktrees (#2053,@tnozicka)
• Add context on which condition failed in WaitForObjectState (#2055,@tnozicka)
• Adjust ci-deploy.sh rollout times to cover for graceful termination (#2059,@tnozicka)
• Substitute unbound SO_SCYLLACLUSTER_STORAGECLASS_NAME var in ci-deploy script (#2062,@rzetelskik)
• Set explicit storageClassName in monitoring examples and docs (#2064,@rzetelskik)
• Fix replacing images in ci-deploy.sh and waiting for rollouts (#2077,@tnozicka)
• Update ScyllaDB versions to 6.1.0 and 2024.1.7 (#2090,@rzetelskik)
• Collect, snapshot and gracefully restore global objects in e2e test (#2091,@tnozicka)
• Update ScyllaDB versions to 6.1.1 and 2024.1.8 (#2103,@rzetelskik)
• Update dependencies (#2106,@tnozicka)
• Synchronize ScyllaDB Manager version with config.yaml (#2107,@tnozicka)
• Pin tags to 1.14 (#2108,@tnozicka)