📝 6 新 Issues
✅ 5 关闭
✨ 8 新 PRs
🎉 9 合并
🔥 核心要点
- 🎉 PR #2454 新增 Lake Tiering 表级监控指标，提升系统可观测性
- 🎉 PR #2862 修复 Paimon lake 表分区过期策略从未触发的问题
- 🎉 PR #2842 优化 Iceberg 扫描，仅在存在过滤条件时启用列统计
- 🐛 PR #2860 修复 ServerConnection 重试认证任务未正确取消的问题
- ⚠️ PR #2863 回滚 Helm Chart 组件配置隔离功能，因评审意见未妥善处理
⚠️ 重点关注
- ⚠️ Issue #2856 Admin.alterClusterConfigs() 在校验失败时返回 UnknownServerException 而非 ConfigException
- ⚠️ Issue #2854 测试用例 FlussAuthorizationITCase.testDeleteProducerOffsetsAuthorization 不稳定
💬 Issue/PR 动态
Issue 讨论
Issue #2809 [KV] Support Full KV Scan
- @platinumhamburg: 指出三个问题：close_scanner请求无法发送、scanner生命周期未集成、客户端未处理扫描错误
- @polyzos: 感谢详细反馈，将逐一修复这三个问题并继续迭代
Issue #2454 [lake/tiering] Add per-table monitoring metrics for Lake Tiering
- @beryllw: 已在 Docker 环境中验证指标名称和标签值正确
Issue #2859 [lake/tiering] fallback to Flink's `io.tmp.dir` when `client.scanner.io.tmpdir` is not set
- @zuston: 请求 @luoyuxia 和 @wuchong 审阅此 PR
Issue #2865 [Helm] Restructure helm chart of Fluss
- @nhuantho: 询问这是否是改进 Fluss Helm Chart 的正确方向
Issue #2863 [Helm] Revert Chart Component Configuration Isolation (#2472)
- @swuferhong: 回滚 #2472 因为评审意见未妥善处理，如需继续请先讨论
Issue #2581 [test] Unstable test FlussAuthorizationITCase. testInitWriter
- @Prajwal-banakar: 建议将 waitUntilTableReady 超时从1分钟增加到2分钟
PR Review
PR #2711 [helm] Enable metrics reporting on helm charts
- @affo: 建议只接受端口配置，其他 Prometheus 选项通过 overrides 设置
- @morazow: 同意方案，建议在配置冲突时直接报错
PR #2454 [lake/tiering] Add per-table monitoring metrics for Lake Tiering
- @Copilot: 指出注释与断言不一致、null应改为UNKNOWN、NPE风险等多处代码问题
- @luoyuxia: 建议将指标类型改为 Counter，对 UnsupportedOperationException 使用 trace/Debug 级别
PR #2839 [metric] Support influxdb reporter
- @Copilot: 指出编译错误、NPE风险、异常处理缺失、测试方法脆弱性及模块未集成到发布包等问题
PR #2859 [lake/tiering] fallback to Flink's `io.tmp.dir` when `client.scanner.io.tmpdir` is not set
- @Copilot: 指出路径拼接使用 /fluss 绝对路径会忽略父目录，需修复为相对路径
PR #2846 [helm] Rewrite README to point to website docs
- @morazow: 建议使用渲染后的文档链接，添加本地开发指引
- @affo: README 用于开发目的，应该指向实际文件以便开发者编辑
PR #2862 [lake/paimon] Fix partition expiration never triggered in Paimon lake table managed by Tiering Service
- @Copilot: 指出注释中的配置值与实际配置格式不一致，三元表达式冗余
📝 新建 Issue/PR
Issues
- #2865 [Helm] Restructure helm chart of Fluss @nhuantho
- #2861 [lake/paimon] Partition expiration never fires in Paimon lake table when managed by Tiering Service @LiebingYu
- #2856 [server] Admin.alterClusterConfigs() returns UnknownServerException instead of ConfigException on validation failure @LiebingYu
- #2854 [test] Unstable test FlussAuthorizationITCase.testDeleteProducerOffsetsAuthorization @app/github-actions
- #2853 [Feature] Support downloading CDC log segments from remote storage during KV table recovery @swuferhong
- #2851 ListDatabaseSummary always return table count as 0 or 1 @loserwang1024
Pull Requests
- #2866 Fix Unstable test FlussAuthorizationITCase.testDeleteProducerOffsetsAuthorization @hemanthsavasere
- #2864 [flink][test] Fix OOM when startTaskManager in FlinkMetricsITCase @Prajwal-banakar
- #2860 [rpc] Fixed the issue about ServerConnection retry authentication tasks were not canceled @xuzifu666
- #2858 update docs for data formats @polyzos
- #2857 [server] Register ConfigException in Errors to fix misleading UnknownServerException from alterClusterConfigs @LiebingYu
- #2855 [kv] Support downloading CDC log segments from remote storage during KV table recovery @swuferhong
- #2852 [server] Fix get table count of database. @loserwang1024
- #2850 [docs] Update s3.region in quickstart docs of flink and lakehouse @nhuantho
✅ 关闭 Issue/PR
已关闭 Issues
- #2861 [lake/paimon] Partition expiration never fires in Paimon lake table when managed by Tiering Service @LiebingYu
- #2844 [docs] Update doc in api/java-client about complex type @nhuantho
- #2629 AbfsFileSystemBehaviorITCase fails with "Bind Address already in use" in CI @litiliu
- #2440 [lake/tiering] Add Server-Level Monitoring Metrics for Lake Tiering @beryllw
- #2191 [helm] Create region config for components in value file @nhuantho
已合并 PRs
- #2863 [Helm] Revert Chart Component Configuration Isolation (#2472) @affo
- #2862 [lake/paimon] Fix partition expiration never triggered in Paimon lake table managed by Tiering Service @LiebingYu
- #2859 [lake/tiering] fallback to Flink's `io.tmp.dir` when `client.scanner.io.tmpdir` is not set @zuston
- #2845 [docs] Update doc in api/java-client about complex type @nhuantho
- #2842 [lake/iceberg] only enable column stats when scan filter exists @zuston
- #2798 [typo] Fix typo in log loading period @zcoo
- #2630 [Fix][UT]Fix AbfsFileSystemBehaviorITCase fails with "Bind Address already in use" in CI @litiliu
- #2472 [Helm] Chart Component Configuration Isolation @hemanthsavasere
- #2454 [lake/tiering] Add per-table monitoring metrics for Lake Tiering @beryllw