Releases: scylladb/scylla-operator
Releases · scylladb/scylla-operator
v1.10.0-beta.0
Release notes for 1.10.0-beta.0
Container images
docker.io/scylladb/scylla-operator:1.10.0-beta.0
Changes By Kind (since 1.9.0)
Feature
- Run node cleanup after scaling the ScyllaCluster (#1294,@zimnx)
- Identify nodes by their Host ID instead of IP address in readiness probes (#1297,@zimnx)
- Replace ScyllaCluster nodes using Host ID (#1312,@zimnx)
Bug
- Fix NodeConfig DaemonSet prunning (#1278,@zimnx)
- Fix panic in Operator occuring during node replacement (#1288,@zimnx)
Documentation
- Replace xfs-formatter with NodeConfig in GKE examples (#1269,@rzetelskik)
- Use Dynamic Local Volume Provisioner in EKS and GKE examples (#1272,@rzetelskik)
- Build v1.9 documentation (#1274,@zimnx)
- Set v1.9 documentation as stable (#1295,@zimnx)
- Update docs theme version to 1.5 (#1308,@rzetelskik)
- Bump ScyllaDB to 5.1.15 and ScyllaDB Manager to 3.1.2 in examples, Helm charts and e2e suites (#1322,@rzetelskik)
- Adjust docs development workflow and add developers' README (#1323,@rzetelskik)
- Update
Deploying Scylla on GKEandDeploying Scylla on EKSdocumentation sections to reflect changes in storage setup (#1324,@rzetelskik) - Update docs dependencies (#1326,@rzetelskik)
Other
- Fix release notes generation (#1275,@zimnx)
- Replace
github.com/ghodss/yamllibrary withsigs.k8s.io/yaml(#1298,@Juneezee) - Enhancement proposal for running cleanup after scaling (#1300,@zimnx)
- Cleanup of
scyllaclientpackage (#1309,@zimnx) - Enhancement proposal for replacing nodes using their HostIDs (#1313,@zimnx)
- Pin tags to 1.10 (#1328,@zimnx)
v1.9.0
Release notes for 1.9.0
Container images
docker.io/scylladb/scylla-operator:1.9.0
Changes By Kind (since 1.8.1)
Feature
- Automated Local Disk Setup (#1107,@zimnx)
- Add ScyllaDBMonitoring (#1111,@tnozicka)
- Update Grafana version to 9.3.4 (#1154,@tnozicka)
- Update and pin Ubuntu to 22.04 (#1160,@tnozicka)
- Add support for tuning EKS 1.24 using containerd CRI (#1165,@zimnx)
- Pregenerate RSA keys and speed up bootstrap times (#1202,@tnozicka)
- Bump ScyllaDB and ScyllaDB Manager versions in Helm chart, examples and e2es (#1238,@zimnx)
- Update dependencies (#1243,@tnozicka)
- Revert "Bump ScyllaDB and ScyllaDB Manager versions in Helm chart, examples and e2es " (#1246,@zimnx)
- Use ScyllaDB 5.1.9 and ScyllaDB Manager 3.1.0 as default versions in Helm chart, examples and e2e (#1247,@zimnx)
Bug
- Fix NodeConfigPod controller to exclusively enqueue and sync Scylla Pods (#1131,@rzetelskik)
- Fix leader election error handling and context cancelation (#1142,@tnozicka)
- Fix ApplyRoleBinding updates for immutable roleRef (#1143,@tnozicka)
- Fix NodeConfigDaemonController working on objects from a different shard (#1146,@tnozicka)
- Fix support for immutable fields in Job Apply (#1148,@zimnx)
- Add missing symlink to ScyllaDBMonitoring CRD in helm CRDs (#1197,@rzetelskik)
- Fix "job" labels in Grafana dashboards' queries (#1199,@rzetelskik)
- Install Prometheus ClusterRoles in Helm deployments (#1205,@rzetelskik)
- Fix node setup image in EKS example (#1221,@zimnx)
- Extend test runner with a report verifier checking if every Spec was ran exactly once (#1253,@rzetelskik)
- Fix seed propagation in test runner (#1256,@rzetelskik)
- Fix ScyllaDB version upgrade being stuck on Pod staying in maintenance mode (#1267,@zimnx)
- Fix NodeConfig DaemonSet prunning (#1279,@zimnx)
- Fix panic in Operator occuring during node replacement (#1290,@zimnx)
Documentation
- Build v1.8 documentation (#1134,@zimnx)
- Bump docs stable version to 1.8 (#1153,@zimnx)
- Add (optional) extended dashboards in ScyllaDBMonitoring and drop the unmanaged setup (#1161,@rzetelskik)
- Update docs version to 1.3 (#1185,@rzetelskik)
- Fix command go-template in monitoring docs (#1230,@rzetelskik)
- Update HAProxy ingress example (#1235,@tnozicka)
- Fix EKS examples (#1268,@rzetelskik)
Flaky/Failing Test
- Fix flakyness of replace node e2e (#1126,@zimnx)
- Await schema agreement before verifying CQL data (#1128,@zimnx)
- Fix eventual waiting in monitoring e2e (#1162,@tnozicka)
- Fix crypto unit test not to capture loop variable (#1163,@tnozicka)
- Wait for haproxy-ingress to be ready when setting up the example (#1177,@tnozicka)
- Fix flaky kubelet boot in CI (#1266,@zimnx)
Other
- Use explicitly created SA token in E2E test suite (#1099,@rzetelskik)
- Bump Kubernetes version to 1.24.9 in CI (#1106,@zimnx)
- Use generic handlers (#1125,@tnozicka)
- Use generic controller ref manager (#1127,@tnozicka)
- Bump Go version in release pipeline (#1136,@zimnx)
- Condense existing apply methods into a generic one (#1138,@tnozicka)
- Switch cgroup manager from systemd to cgroupfs (#1158,@zimnx)
- Make releases even for alpha version to be easy to consume (#1159,@tnozicka)
- Fix serial job artifact names (#1164,@zimnx)
- Remove CC from Slack failure report (#1166,@rzetelskik)
- Fix kubelet cgroup slices in CI setup (#1168,@tnozicka)
- Tune kubelet config in CI (#1169,@tnozicka)
- Set explicit timeouts for our main GH actions workflow (#1170,@tnozicka)
- Fix GH action setting up Kubernetes (#1173,@zimnx)
- Increase maxPods value in kubelet config in CI (#1174,@rzetelskik)
- Force default resources in CI (#1178,@tnozicka)
- Limit CI parallelism (#1179,@tnozicka)
- Tune CI and adjust timeouts (#1182,@tnozicka)
- Update dependencies (#1183,@tnozicka)
- Update docs dependencies (#1184,@tnozicka)
- Make the hack gomod valid (#1187,@tnozicka)
- Fix sysctl settings in kubernetes setup Action (#1195,@zimnx)
- Update Go to 1.20 (#1196,@rzetelskik)
- Fix PrometheusExposeOptions struct name in API definitions (#1198,@tnozicka)
- Workaround mkfs.xfs min size limits in newer versions (#1200,@tnozicka)
- Replace deprecated GCP Auth with a dedicated GitHub action (#1211,@zimnx)
- Cleanup tune2fs_test.sh (#1214,@mykaul)
- Use Dynamic Local Volume Provisioner (#1227,[@zimnx]...
v1.9.0-rc.2
v1.9.0-rc.1
v1.9.0-rc.0
Release notes for 1.9.0-rc.0
Container images
docker.io/scylladb/scylla-operator:1.9.0-rc.0
Changes By Kind (since 1.8.1)
Feature
- Automated Local Disk Setup (#1107,@zimnx)
- Add ScyllaDBMonitoring (#1111,@tnozicka)
- Update Grafana version to 9.3.4 (#1154,@tnozicka)
- Update and pin Ubuntu to 22.04 (#1160,@tnozicka)
- Add support for tuning EKS 1.24 using containerd CRI (#1165,@zimnx)
- Pregenerate RSA keys and speed up bootstrap times (#1202,@tnozicka)
- Bump ScyllaDB and ScyllaDB Manager versions in Helm chart, examples and e2es (#1238,@zimnx)
- Update dependencies (#1243,@tnozicka)
- Revert "Bump ScyllaDB and ScyllaDB Manager versions in Helm chart, examples and e2es " (#1246,@zimnx)
- Use ScyllaDB 5.1.9 and ScyllaDB Manager 3.1.0 as default versions in Helm chart, examples and e2e (#1247,@zimnx)
Bug
- Fix NodeConfigPod controller to exclusively enqueue and sync Scylla Pods (#1131,@rzetelskik)
- Fix leader election error handling and context cancelation (#1142,@tnozicka)
- Fix ApplyRoleBinding updates for immutable roleRef (#1143,@tnozicka)
- Fix NodeConfigDaemonController working on objects from a different shard (#1146,@tnozicka)
- Fix support for immutable fields in Job Apply (#1148,@zimnx)
- Add missing symlink to ScyllaDBMonitoring CRD in helm CRDs (#1197,@rzetelskik)
- Fix "job" labels in Grafana dashboards' queries (#1199,@rzetelskik)
- Install Prometheus ClusterRoles in Helm deployments (#1205,@rzetelskik)
- Fix node setup image in EKS example (#1221,@zimnx)
- Extend test runner with a report verifier checking if every Spec was ran exactly once (#1253,@rzetelskik)
- Fix seed propagation in test runner (#1256,@rzetelskik)
- Fix ScyllaDB version upgrade being stuck on Pod staying in maintenance mode (#1267,@zimnx)
Documentation
- Build v1.8 documentation (#1134,@zimnx)
- Bump docs stable version to 1.8 (#1153,@zimnx)
- Add (optional) extended dashboards in ScyllaDBMonitoring and drop the unmanaged setup (#1161,@rzetelskik)
- Update docs version to 1.3 (#1185,@rzetelskik)
- Fix command go-template in monitoring docs (#1230,@rzetelskik)
- Update HAProxy ingress example (#1235,@tnozicka)
- Fix EKS examples (#1268,@rzetelskik)
Flaky/Failing Test
- Fix flakyness of replace node e2e (#1126,@zimnx)
- Await schema agreement before verifying CQL data (#1128,@zimnx)
- Fix eventual waiting in monitoring e2e (#1162,@tnozicka)
- Fix crypto unit test not to capture loop variable (#1163,@tnozicka)
- Wait for haproxy-ingress to be ready when setting up the example (#1177,@tnozicka)
- Fix flaky kubelet boot in CI (#1266,@zimnx)
Other
- Use explicitly created SA token in E2E test suite (#1099,@rzetelskik)
- Bump Kubernetes version to 1.24.9 in CI (#1106,@zimnx)
- Use generic handlers (#1125,@tnozicka)
- Use generic controller ref manager (#1127,@tnozicka)
- Bump Go version in release pipeline (#1136,@zimnx)
- Condense existing apply methods into a generic one (#1138,@tnozicka)
- Switch cgroup manager from systemd to cgroupfs (#1158,@zimnx)
- Make releases even for alpha version to be easy to consume (#1159,@tnozicka)
- Fix serial job artifact names (#1164,@zimnx)
- Remove CC from Slack failure report (#1166,@rzetelskik)
- Fix kubelet cgroup slices in CI setup (#1168,@tnozicka)
- Tune kubelet config in CI (#1169,@tnozicka)
- Set explicit timeouts for our main GH actions workflow (#1170,@tnozicka)
- Fix GH action setting up Kubernetes (#1173,@zimnx)
- Increase maxPods value in kubelet config in CI (#1174,@rzetelskik)
- Force default resources in CI (#1178,@tnozicka)
- Limit CI parallelism (#1179,@tnozicka)
- Tune CI and adjust timeouts (#1182,@tnozicka)
- Update dependencies (#1183,@tnozicka)
- Update docs dependencies (#1184,@tnozicka)
- Make the hack gomod valid (#1187,@tnozicka)
- Fix sysctl settings in kubernetes setup Action (#1195,@zimnx)
- Update Go to 1.20 (#1196,@rzetelskik)
- Fix PrometheusExposeOptions struct name in API definitions (#1198,@tnozicka)
- Workaround mkfs.xfs min size limits in newer versions (#1200,@tnozicka)
- Replace deprecated GCP Auth with a dedicated GitHub action (#1211,@zimnx)
- Cleanup tune2fs_test.sh (#1214,@mykaul)
- Use Dynamic Local Volume Provisioner (#1227,@zimnx)
- Refactor release notes publishing into a reusable GitHub Action (#1234,@zimnx)
- Update kubernetes to v1.27.1 in CI setup ([#1244](https://github.com/scylladb/scylla-op...
v1.9.0-beta.1
v1.8.1
v1.8.1-rc.0
v1.8.0
Release notes for 1.8.0
Container images
docker.io/scylladb/scylla-operator:1.8.0
Changes By Kind (since 1.7.4)
Feature
- Expose node-exporter metrics in Scylla Pods (#952,@zimnx)
- Add host_id annotation sync to sidecar controller (#955,@rzetelskik)
- Reconcile metadata explicitly with apply (#960,@tnozicka)
- Allow directly selecting member services for building CQL connections (#969,@rzetelskik)
- Bump scylla to 4.6.3 and manager to 2.6.3 (#977,@tnozicka)
- Reconcile Ingress objects for ScyllaCluster (#993,@zimnx)
- Support tuning for minikube on cri-o (#995,@zimnx)
- Bump scylla-manager and scylla-manager-agent to 3.0.0 (#1000,@rzetelskik)
- Promote PodDisruptionBudgets from V1Beta1 to V1 (#1005,@Choraden)
- Manage TLS certificates (#1021,@tnozicka)
- Use more workers by default to make the controller more responsive (#1039,@tnozicka)
- Add Status Conditions (#1070,@tnozicka)
- Create scylla cql connection bundle (#1074,@tnozicka)
- Update golang libraries (#1088,@tnozicka)
- Bump scylla to 5.0.5 and manager to 3.0.1 (#1105,@tnozicka)
- Allow to apply our manifests using server-side apply (#1109,@zimnx)
Bug
- Include NodeConfigs with disabled optimizations as matching ones in tuning ConfigMap (#920,@zimnx)
- Tune nodes immediately when tuning DaemonSet is scheduled (#926,@zimnx)
- Keep custom annotations on ServiceAccounts (#933,@zimnx)
- Bring back ServiceAccount template in Scylla Helm chart (#937,@zimnx)
- Fix decommissioning of drained nodes (#967,@tnozicka)
- Stabilize tuning Job parameters and status of NodeConfig (#996,@zimnx)
- Close idle connections to scylla API when done using them (#1017,@zimnx)
- Avoid watch cache when using UntilWithSync (#1049,@tnozicka)
- Don't use "any" prefix for LB CQL service (#1085,@tnozicka)
- Make sure manager controller reacts to SA token changes (#1087,@Choraden)
- Fix repository paths so it can be vendored (#1098,@tnozicka)
Documentation
- Enable building documentation from v1.7 (#925,@zimnx)
- Mark 1.7 documentation as stable (#929,@zimnx)
- Bump release schedule (#930,@zimnx)
- Add link to Scylla University lesson (#949,@guy9)
- Fix "local ssd disks" link in EKS deployment docs (#970,@rzetelskik)
- Use local NVMe storage class in EKS example (#984,@fee-mendes)
- Fix Scylla Monitoring support (#1001,@rzetelskik)
- Drop unsupported releases from docs (#1008,@rzetelskik)
- Add 'Download Scylla Operator' section to 'Deploying Scylla Operator' in docs (#1014,@rzetelskik)
Flaky/Failing Test
- Add conntrack-cleaner DaemonSet to CI deployment (#973,@rzetelskik)
- Fix NodeConfig e2e flake and correct resourcequotas resource name in test machinery (#1029,@tnozicka)
- Disable shadawareness E2E giving false positives (#1048,@zimnx)
- Wait for scylla nodes to have a full quorum before testing CQL in e2e tests (#1054,@tnozicka)
- Fix e2e wait timeout for scylla quorum (#1065,@tnozicka)
- Bump gocql to scylladb/[email protected] (#1102,@zimnx)
Other
- Use shared images (#859,@tnozicka)
- Update Scylla docs theme to 1.1.0 (#942,@tnozicka)
- Bump docomentation dependencies (#945,@tnozicka)
- Pin google cloud SDK (#957,@tnozicka)
- Update Golang to 1.18 (#961,@tnozicka)
- Bump docs theme from 1.1 to 1.2 (#965,@tnozicka)
- Gather logs from previous iterations of restarted containers of non-e2e namespaced resources (#976,@rzetelskik)
- Fallback image builds to cgroupfs to workaround ubuntu systemd issues (#999,@tnozicka)
- Fix webhook service name in helm template (#1004,@petercable)
- Fix loglevel type and change image tag in example helm values for Scylla Operator Chart (#1011,@rzetelskik)
- Bump dependencies (#1022,@tnozicka)
- Update Golang to 1.19 (#1026,@tnozicka)
- Unify WaitForState implementations (#1032,@tnozicka)
- Raise loglevel for manager controller in CI (#1040,@Choraden)
- Drop migration workaround for manual ServiceAccounts (#1042,@Choraden)
- Split serial and parallel e2e suites into dedicated jobs (#1046,@tnozicka)
- Add resource version to e2e messages waiting for ScyllaCluster rollout (#1060,@tnozicka)
- Raise loglevel for manager controller in CI (#1062,@Choraden)
- Tee test logs of e2e test runner's parallel processes to files (#1064,@rzetelskik)
- Update actions in CI (#1067,@Choraden)
- Adjust rate limits and timeouts for e2e tests (#1069,@tnozicka)
- Fix running parallel processes in test runner (#1084,[@rzetelskik](https://github.co...
v1.8.0-rc.0
Release notes for 1.8.0-rc.0
Container images
docker.io/scylladb/scylla-operator:1.8.0-rc.0
Changes By Kind (since 1.7.4)
Feature
- Expose node-exporter metrics in Scylla Pods (#952,@zimnx)
- Add host_id annotation sync to sidecar controller (#955,@rzetelskik)
- Reconcile metadata explicitly with apply (#960,@tnozicka)
- Allow directly selecting member services for building CQL connections (#969,@rzetelskik)
- Bump scylla to 4.6.3 and manager to 2.6.3 (#977,@tnozicka)
- Reconcile Ingress objects for ScyllaCluster (#993,@zimnx)
- Support tuning for minikube on cri-o (#995,@zimnx)
- Bump scylla-manager and scylla-manager-agent to 3.0.0 (#1000,@rzetelskik)
- Promote PodDisruptionBudgets from V1Beta1 to V1 (#1005,@Choraden)
- Manage TLS certificates (#1021,@tnozicka)
- Use more workers by default to make the controller more responsive (#1039,@tnozicka)
- Add Status Conditions (#1070,@tnozicka)
- Create scylla cql connection bundle (#1074,@tnozicka)
- Update golang libraries (#1088,@tnozicka)
- Bump scylla to 5.0.5 and manager to 3.0.1 (#1105,@tnozicka)
- Allow to apply our manifests using server-side apply (#1109,@zimnx)
Bug
- Include NodeConfigs with disabled optimizations as matching ones in tuning ConfigMap (#920,@zimnx)
- Tune nodes immediately when tuning DaemonSet is scheduled (#926,@zimnx)
- Keep custom annotations on ServiceAccounts (#933,@zimnx)
- Bring back ServiceAccount template in Scylla Helm chart (#937,@zimnx)
- Fix decommissioning of drained nodes (#967,@tnozicka)
- Stabilize tuning Job parameters and status of NodeConfig (#996,@zimnx)
- Close idle connections to scylla API when done using them (#1017,@zimnx)
- Avoid watch cache when using UntilWithSync (#1049,@tnozicka)
- Don't use "any" prefix for LB CQL service (#1085,@tnozicka)
- Make sure manager controller reacts to SA token changes (#1087,@Choraden)
- Fix repository paths so it can be vendored (#1098,@tnozicka)
Documentation
- Enable building documentation from v1.7 (#925,@zimnx)
- Mark 1.7 documentation as stable (#929,@zimnx)
- Bump release schedule (#930,@zimnx)
- Add link to Scylla University lesson (#949,@guy9)
- Fix "local ssd disks" link in EKS deployment docs (#970,@rzetelskik)
- Use local NVMe storage class in EKS example (#984,@fee-mendes)
- Fix Scylla Monitoring support (#1001,@rzetelskik)
- Drop unsupported releases from docs (#1008,@rzetelskik)
- Add 'Download Scylla Operator' section to 'Deploying Scylla Operator' in docs (#1014,@rzetelskik)
Flaky/Failing Test
- Add conntrack-cleaner DaemonSet to CI deployment (#973,@rzetelskik)
- Fix NodeConfig e2e flake and correct resourcequotas resource name in test machinery (#1029,@tnozicka)
- Disable shadawareness E2E giving false positives (#1048,@zimnx)
- Wait for scylla nodes to have a full quorum before testing CQL in e2e tests (#1054,@tnozicka)
- Fix e2e wait timeout for scylla quorum (#1065,@tnozicka)
- Bump gocql to scylladb/[email protected] (#1102,@zimnx)
Other
- Use shared images (#859,@tnozicka)
- Update Scylla docs theme to 1.1.0 (#942,@tnozicka)
- Bump docomentation dependencies (#945,@tnozicka)
- Pin google cloud SDK (#957,@tnozicka)
- Update Golang to 1.18 (#961,@tnozicka)
- Bump docs theme from 1.1 to 1.2 (#965,@tnozicka)
- Gather logs from previous iterations of restarted containers of non-e2e namespaced resources (#976,@rzetelskik)
- Fallback image builds to cgroupfs to workaround ubuntu systemd issues (#999,@tnozicka)
- Fix webhook service name in helm template (#1004,@petercable)
- Fix loglevel type and change image tag in example helm values for Scylla Operator Chart (#1011,@rzetelskik)
- Bump dependencies (#1022,@tnozicka)
- Update Golang to 1.19 (#1026,@tnozicka)
- Unify WaitForState implementations (#1032,@tnozicka)
- Raise loglevel for manager controller in CI (#1040,@Choraden)
- Drop migration workaround for manual ServiceAccounts (#1042,@Choraden)
- Split serial and parallel e2e suites into dedicated jobs (#1046,@tnozicka)
- Add resource version to e2e messages waiting for ScyllaCluster rollout (#1060,@tnozicka)
- Raise loglevel for manager controller in CI (#1062,@Choraden)
- Tee test logs of e2e test runner's parallel processes to files (#1064,@rzetelskik)
- Update actions in CI (#1067,@Choraden)
- Adjust rate limits and timeouts for e2e tests (#1069,@tnozicka)
- Fix running parallel processes in test runner (#1084,[@rzetelskik](https:/...