📝 8 新 Issues
✅ 5 关闭
✨ 9 新 PRs
🎉 11 合并
⚠️ 重点关注
- ⚠️ Issue #3206 AskAi 功能因隐私策略阻止无法加载，需回滚处理
- ⚠️ Issue #3188 CoordinatorEventManagerTest.testMetricsUpdatedImmediatelyOnStartup 测试不稳定
💬 Issue/PR 动态
Issue 讨论
Issue #3177 [server] Remove CDC log metrics from PK table write path to avoid double counting
- @binary-signal: 解释了 KvBytesIn 和 LogBytesIn 是独立指标，CDC log 确实占用实际资源，不应从监控中隐藏
Issue #3152 [fluss/ai agent] enable ai doc agent on website
- @MehulBatra: 调整了 Doris 图标可见性，后续将配合隐私团队处理域名白名单
- @wuchong: 指出 Ask AI 当前被隐私策略阻止，建议先回滚提交，待隐私团队批准后再添加
Issue #3158 [client] Fix race condition in RemoteLogDownloader causing flaky testPrefetchNum
- @ankit-khare-2015: 请求合并更改以解决问题
- @wuchong: 分析了竞态条件根因，建议将断言修改为检查 permits 在 1-2 之间
Issue #3132 [server] Optimize RemoteLogFetcher with async prefetch for recovery
- @Kaixuan-Duan: 感谢 review，正在推进测试代码和预取深度优化，参考 fluss-rust 实现
Issue #3084 Flink log table consumer threw an out-of-range offset exception
- @loserwang1024: 分析 RemoteLogDownloadFuture 只实现了 onComplete 导致异常无法传播，建议添加超时和失败处理
Issue #3179 Add documentation for FIP (Flink Improvement Proposal) contribution guide
- @luoyuxia: 指出 Confluence 已有完善的 FIP 流程文档，询问是否需要在网站复制完整内容或仅添加链接
Issue #3192 [server] local-cluster.sh starts successfully but processes fail silently when JAVA_HOME is misconfigured
- @matrixsparse: 主动请缨处理此 issue
- @Yohahaha: 感谢并确认分配给 matrixsparse
Issue #3197 [server] Validate Java version >= 11 in config.sh to prevent silent startup failures
- @matrixsparse: 提交 PR 修复 #3192，在 config.sh 中验证 Java 版本防止静默启动失败
Issue #3207 [flink] Batch query on empty datalake-enabled table should return 0 rows instead of failing with UnsupportedOperationException
- @matrixsparse: 主动请缨处理此 issue
- @binary-signal: 将 issue 分配给 matrixsparse
PR Review
PR #3116 [spark] Support filter pushdown for log tables
- @Copilot: 建议处理空 IN 谓词、迁移到 CollectionConverters，指出流式读取未实际应用过滤器的问题
- @luoyuxia: 指出非 Arrow 格式（如 compacted）可能直接抛异常的风险
- @fresh-borzoni: 已处理评审意见，lake reader 下推作为后续计划
- @YannByron: 确认 LGTM
PR #3123 [spark] Make LakeSplit extend Serializable to simplify Spark serialization
- @luoyuxia: 询问为何删除 serialVersionUID，建议也为 PaimonSplit 添加该字段
PR #3132 [server] Optimize RemoteLogFetcher with async prefetch for recovery
- @Kaixuan-Duan: 已处理 close() 调用 awaitTermination、关闭迭代器立即停止、预取匹配改用 ID 比较等评审意见
- @zuston: cc 相关人员希望提升可读性
PR #3158 [client] Fix race condition in RemoteLogDownloader causing flaky testPrefetchNum
- @Copilot: 指出 acquire() 前从队列取元素可能违反优先级顺序，建议在获取许可后再取元素
- @wuchong: 同意 Copilot 意见，指出 acquire() 在 poll() 前是保证消费顺序的关键设计
- @ankit-khare-2015: 询问是否应回滚更改并进一步调查
📝 新建 Issue/PR
Issues
- #3207 [flink] Batch query on empty datalake-enabled table should return 0 rows instead of failing with UnsupportedOperationException @binary-signal
- #3206 [docs] AskAi fails to load on Fluss website @binary-signal
- #3203 [Spark] Add set_cluster_configs and reset_cluster_configs procedures for Spark connector @XuQianJin-Stars
- #3201 [test] fix wrong error message checks in LakeEnabledTableCreateITCase @gyang94
- #3198 [docs] Titles are duplicated in the Authentication section @zhuyufeng0809
- #3192 [server] local-cluster.sh starts successfully but processes fail silently when JAVA_HOME is misconfigured @Yohahaha
- #3191 Support configurable time partition format for auto-partitioned tables @luoyuxia
- #3188 [test] Unstable test CoordinatorEventManagerTest.testMetricsUpdatedImmediatelyOnStartup @app/github-actions
Pull Requests
- #3208 [flink] Fix batch query on empty datalake-enabled table to return 0 rows instead of failing @matrixsparse
- #3205 [helm] Bump chart version to 0.9.1-incubating @polyzos
- #3204 [FLUSS][Spark] Add set_cluster_configs and reset_cluster_configs procedures @XuQianJin-Stars
- #3200 [server] Support configurable time partition format for auto-partitioned tables @wattt3
- #3199 [docs] Fix duplicate titles in the Authentication section @zhuyufeng0809
- #3197 [server] Validate Java version >= 11 in config.sh to prevent silent startup failures @matrixsparse
- #3196 [spark] Add numRowsRead custom metric and improve Spark UI operator name for FlussScan @Yohahaha
- #3194 [server] make RemoteLogManifest sorted with startOffset @zuston
- #3190 [tiering] Fix Paimon IOManager leak in MergeTreeWriter @luoyuxia
✅ 关闭 Issue/PR
已关闭 Issues
- #3122 [spark] Make LakeSplit extend Serializable to simplify Spark serialization @YannByron
- #3117 [spark] Support filter pushdown for log tables @fresh-borzoni
- #3072 Csp_check is warning @luoyuxia
- #2652 [docs] Update roadmap page to match the 2026 goals @MehulBatra
- #2599 [docs] Add "Change Data Feed" feature introduction for "Primary Key Table" page @wuchong
已合并 PRs
- #3193 [tiering] Fix Paimon IOManager leak in MergeTreeWriter causing /tmp disk exhaustion @luoyuxia
- #3189 [bp][release-0.9][helm] Add existingSecret support for SASL credentials (#3171) @fresh-borzoni
- #3184 [bp][helm] Support injecting environment variables and external secrets (backport to release-0.9) @morazow
- #3181 [flink] Push down lake filters for non-partitioned scans @luoyuxia
- #3164 [spark] Exclude slf4j from spark connector shade @Yohahaha
- #3152 [fluss/ai agent] enable ai doc agent on website @MehulBatra
- #3142 [spark][doc] Add Spark batch union read @Yohahaha
- #3123 [spark] Make LakeSplit extend Serializable to simplify Spark serialization @YannByron
- #3116 [spark] Support filter pushdown for log tables @fresh-borzoni
- #2696 [docs/website] Update website with roadmap 2026 changes @MehulBatra
- #2695 [doc] Add info for changedatafeed in Log and Pk tables to match the existing industry format @MehulBatra