Releases: kopeio/etcd-manager
3.0.20210228
Latest patch release in the 3.0 series, supporting the etcd 3 series and migration from the etcd 2 series.
Contains an important bugfix: etcd-manager pods before version 3.0.20200428 will stop working after 1 year, when some certificates expire. Users of prior versions of etcd-manager must upgrade to this version to avoid this.
Please see the release notes for the full list of changes.
Important changes
-
Improved diagnostics and support for metrics, particularly for OpenStack
-
Add etcd-manager metrics, add openstack API metrics @zetaab #396
-
Add log levels to prevent too verbose logging @ottosulin #394
Please see the release notes for the full changes.
3.0.20210122
Latest patch release in the 3.0 series, supporting the etcd 3 series and migration from the etcd 2 series.
Contains an important bugfix: etcd-manager pods before version 3.0.20200428 will stop working after 1 year, when some certificates expire. Users of prior versions of etcd-manager must upgrade to this version to avoid this.
Please see the release notes for the full list of changes.
Important changes
- Fixes potential bug where clusters occasionally got "stuck" during an upgrade, requiring manual intervention (typically terminating the affected process/machine)
- Fixes some issues causing slower cluster updates
Changes since 3.0.20201202
- Release notes for 3.0.20201209 @justinsb #384
- Handling hanging list command from etcd client during reconciliation loop @mmerrill3 #388
- Update dependencies to match kOps 1.19 @hakman #385
- Include IPs in peer certificates @justinsb #389
- Use etcd v3.4.13 in tests @hakman #390
- Eager-broadcast leadership on every leader iteration @justinsb #391
- Update Go to 1.15.7 @hakman #359
Please see the release notes for the full changes.
3.0.20201209
Latest patch release in the 3.0 series, supporting the etcd 3 series and migration from the etcd 2 series.
Contains an important bugfix: etcd-manager pods before version 3.0.20200428 will stop working after 1 year, when some certificates expire. Users of prior versions of etcd-manager must upgrade to this version to avoid this.
Please see the release notes for the full list of changes.
Important changes
- Initial (experimental) Azure support
- Fix images on arm64, which were mistakenly built with amd64 bases
- Avoid certificate expiration by eagerly renewing certificates #309, #313. This fixes problems where an etcd-manager process will stop working after 1 year, when the certificate expires. Users of prior versions of etcd-manager must upgrade to avoid this.
Changes since 3.0.20201202
- Update Kops dependency for Azure Blob Storage support @rifelpet #372
- Exclude gazelle from tools/deb-tools @justinsb #373
- Regenerate bazel in tools/deb-tools @justinsb #374
- Release notes for 3.0.20201202 @justinsb #375
- Remove travis CI @justinsb #377
- Fix vendor generation for tools/deb-tools subproject @justinsb #376
- Add script to verify image hashes @justinsb #380
- Fix some incorrect base image hashes for arm64 @justinsb #379
- Support Azure @kenji-cloudnatix #378
- Add more descriptions to wait loops @justinsb #383
- Rename fields in the azure client struct @kenji-cloudnatix #382
- Fix small typo in code comment @justinsb #381
Please see the release notes for the full changes.
3.0.20201202
Latest patch release in the 3.0 series, supporting the etcd 3 series and migration from the etcd 2 series.
Contains an important bugfix: etcd-manager pods before version 3.0.20200428 will stop working after 1 year, when some certificates expire. Users of prior versions of etcd-manager must upgrade to this version to avoid this.
Please see the release notes for the full list of changes.
Important changes
- Workaround bug where we are unable to restore backups when the ETCD_LISTEN_METRICS_URL environment variable is set
- Avoid certificate expiration by eagerly renewing certificates #309, #313. This fixes problems where an etcd-manager process will stop working after 1 year, when the certificate expires. Users of prior versions of etcd-manager must upgrade to avoid this.
Changes since 3.0.20201117
- Release notes for 3.0.20201117 @justinsb #364
- Fix gofmt @justinsb #365
- Add gofmt check to github actions @justinsb #366
- Add boilerplate to tools/deb-tools/main.go @justinsb #367
- Do not set ListenMetricsURLS @granular-ryanbonham #369
- Fix bazel formatting @justinsb #370
Please see the release notes for the full changes.
3.0.20201117
Latest patch release in the 3.0 series, supporting the etcd 3 series and migration from the etcd 2 series.
Contains an important bugfix: etcd-manager pods before version 3.0.20200428 will stop working after 1 year, when some certificates expire. Users of prior versions of etcd-manager must upgrade to this version to avoid this.
Please see the release notes for the full list of changes.
Important changes
- Avoid certificate expiration by eagerly renewing certificates #309, #313. This fixes problems where an etcd-manager process will stop working after 1 year, when the certificate expires. Users of prior versions of etcd-manager must upgrade to avoid this.
Changes since 3.0.20200531
- Release notes for 3.0.20200531 @justinsb #330
- Adds support for using OS application credentials @olemarkus #336
- Fixes usage of OpenStack Swift reauthentication @kciredor #338
- Move from debian-hyperkube-base to debian-base @rdrgmnzs #340
- Add license headers to each file @justinsb #341
- Fix some typos picked up by verify-spelling @justinsb #342
- Fix some problems with trailing spaces @justinsb #343
- Add support for etcd 3.4.13 @hakman #344
- Switch to gcr.io/cloud-marketplace-containers/google/debian10 - Fix for #340 option 1 @rdrgmnzs #345
- Support for ARM64 @hakman,@justinsb #334
- BUG: OpenStack ignore AvailabilityZone in discovery @zetaab #351
- Added full cinder ID to candidateDeviceNodes @goberle #358
- feat(etcd-manager-ctl): use backupname to delete backup instead of timestamp @sudoforge #352
- Update kops to pick up AllowAuth Openstack @justinsb #347
- Build base image by raw expansion of deb packages @justinsb #355
- Switch the cloudbuild docker image, locking to 2.2.0 @justinsb #363
- Fix build on case-insensitive file systems (MacOS) @justinsb #361
- Set AltNames on server certificates @justinsb #362
- govet: Fix a log message @justinsb #360
Please see the release notes for the full changes.
3.0.20200531
Latest patch release in the 3.0 series, supporting the etcd 3 series and migration from the etcd 2 series.
Contains an important bugfix: etcd-manager pods before version 3.0.20200428 will stop working after 1 year, when some certificates expire. Users of prior versions of etcd-manager must upgrade to this version to avoid this.
Please see the release notes for the full list of changes.
Important changes
- Avoid certificate expiration by eagerly renewing certificates #309, #313. This fixes problems where an etcd-manager process will stop working after 1 year, when the certificate expires. Users of prior versions of etcd-manager must upgrade to avoid this.
Highlights
- Removes old backups even when using an S3 bucket with versioning enabled. Previously although the backups were removed on-schedule, the version retention meant that all backups were retained. The configured policy is now honored.
Changes since 3.0.20200527
- Simplify downloading backups @johngmyers #327
- Remove all versions for backups when using S3 @hakman #328
- Update vendored kops version to 1.17.0 @justinsb #329
Please see the release notes for the full changes.
3.0.20200527
Latest patch release in the 3.0 series, supporting the etcd 3 series and migration from the etcd 2 series.
Contains an important bugfix: etcd-manager pods before version 3.0.20200428 will stop working after 1 year, when some certificates expire. Users of prior versions of etcd-manager must upgrade to this version to avoid this.
Please see the release notes for the full list of changes.
Important changes
- Avoid certificate expiration by eagerly renewing certificates #309, #313. This fixes problems where an etcd-manager process will stop working after 1 year, when the certificate expires. Users of prior versions of etcd-manager must upgrade to avoid this.
Highlights
- Configurable backup retention period
- Faster startup by avoiding mountpoint collisions
- More efficient backups by avoiding a copy
Changes since 3.0.20200429
- Use env vars to customize backup retention @hakman #322
- Use next attachment point when device already in use @hakman #323
- Simplify uploading backups @johngmyers #324
Please see the release notes for the full changes.
3.0.20200429
Latest patch release in the 3.0 series, supporting the etcd 3 series and migration from the etcd 2 series.
Contains an important bugfix: etcd-manager pods before version 3.0.20200428 will stop working after 1 year, when some certificates expires. Users of prior versions of etcd-manager must upgrade to this version to avoid this.
Please see the release notes for the full list of changes.
Important changes
- Avoid certificate expiration by eagerly renewing certificates #309, #313. This fixes problems where an etcd-manager process will stop working after 1 year, when the certificate expires. Users of prior versions of etcd-manager must upgrade to avoid this.
Highlights
- Support for new af-south-1 and eu-south-1 AWS regions
Changes since 3.0.20200428
Please see the release notes for the full changes.
3.0.20200428
Latest patch release in the 3.0 series, supporting the etcd 3 series and migration from the etcd 2 series.
Contains an important bugfix: Prior etcd-manager pods will stop working after 1 year, when some certificates expires. Users of prior versions of etcd-manager must upgrade to this version to avoid this.
Please see the release notes for the full list of changes.
Important changes
- Avoid certificate expiration by eagerly renewing certificates #309, #313. This fixes problems where an etcd-manager process will stop working after 1 year, when the certificate expires. Users of prior versions of etcd-manager must upgrade to avoid this.
Highlights
Changes since 3.0.20200307
- Release notes for 3.0.20200307 @justinsb #303
- Add support for etcd 3.3.17 @rifelpet #304
- Adding client usage extension for server cert (#305) @mmerrill3 #306
- Add a check to renew certificates on startup if they expire in 60 days or less @coderanger #309
- Try github actions @justinsb #310
- Upgrade bazel to 2.2.0 @justinsb #311
- Update to go 1.13.10 @justinsb #314
- Bazel: update dependency @justinsb #316
- e2e tests should wait for cluster readiness @justinsb #318
- Remove old bazel versions from travis @justinsb #317
- Always renew certificates @justinsb #313
Please see the release notes for the full changes.