🌊 Fluss 每日动态

📅 2026-06-24

🔗 apache/fluss

📝 0 新 Issues

✅ 1 关闭

✨ 1 新 PRs

🎉 3 合并

🔥 核心要点

🎉 PR #3518 Introduce Hudi compaction in tiering service - significant lakehouse integration enhancement
🎉 PR #3508 Exclude vulnerable jetty-http dependency - security fix for CVE
🎉 PR #3447 Enum conversion functionality for typed writer/reader
🐛 Issue #3446 Closed: Add enum conversion functionality for Typed Writer/Reader

⚠️ 重点关注

✅ 今日无重大问题

💬 Issue/PR 动态

Issue 讨论

Issue #3483 commitLakeTableSnapshot writes stale snapshot path after dropTable, leaving the tiering pipeline permanently stuck

@Jackeyzhe: Updated PR #3494 to scope only to case (1), rejecting commits for non-existent/queued-for-deletion tables with test verification

Issue #3415 Add per-scanner dimension for scanner metrics

@Jackeyzhe: Acknowledged and will take on this issue

Issue #3511 [server] Add negative cache for non-existent partition IDs in metadata requests

@swuferhong: Confirmed comments addressed in the PR

Issue #3440 [test] Unstable test CommitRemoteLogManifestITCase.testDeleteOutOfSyncReplicaLogAfterCommit

@github-actions[bot]: Reported test failure in CommitRemoteLogManifestITCase

Issue #3476 [lake/paimon] Support custom Paimon lake table path

@luoyuxia: Suggested scoping PR to just database mapping, removing extra changes like Hive test deps
@wzx140: Clarified table-name mapping is higher priority; willing to remove unrelated changes to keep scope manageable

Issue #3441 [test] Stabilize CommitRemoteLogManifestITCase

@zerolbsony: Asked if this problem needs to be resolved

Issue #3449 [lake] Stabilize Paimon tiering IT

@zerolbsony: Asked if this problem needs to be resolved

Issue #3420 [lake/tiering] Support reporting watermark to Paimon snapshot

@luoyuxia: Asked about motivation: whether #3314 partition readiness signal would make this watermark reporting unnecessary

Issue #3518 [lake/hudi] Introduce Hudi compaction in tiering service

@luoyuxia: Checked if XuQianJin-Stars has remaining comments
@XuQianJin-Stars: Confirmed no more comments

Issue #3515 [docs] Add Apache Polaris/Iceberg REST catalog quick-start guide

@GabrielBBaldez: Added Polaris-Realm note, vended vs static credentials section, and explicit oauth2-server-uri config
@morazow: Suggested removing Hadoop deps section since S3FileIO with vended credentials is the default
@MehulBatra: Pointed out oauth2-server-uri should be set explicitly as implicit fallback is deprecated in Iceberg 1.6.0

PR Review

PR #3494 [FLUSS-3483][coordinator] Reject lake snapshot commits for dropped tables

@Copilot: Identified thread safety issue: V2 commit path reads CoordinatorContext from ioExecutor causing data race
@luoyuxia: Suggested removing redundant test and renaming method to ensureTableNotDeleted for broader use

PR #3518 [lake/hudi] Introduce Hudi compaction in tiering service

@Copilot: Flagged snapshot ID ordering issue and nondeterministic compaction commit order from HashMap iteration
@XuQianJin-Stars: Extensive review: identified orphan inflight compactions, wrong-bucket false positives, Fluss snapshot ID issues, constructor 'this' escape, and missing rollback handling

PR #3447 [client] Enum conversion functionality for typed writer/reader

@fresh-borzoni: Suggested using `instanceof Enum` instead of `getClass().isEnum()` to handle anonymous enum subclasses
@VladBanar: Applied suggested changes to fix enum handling

PR #3459 [client][server] Support adding aggregation columns

@luoyuxia: Requested updating table options for new merge columns, fixing null aggFunction printing, and verifying merge with tests
@wzx140: Explained fields.*.agg is Flink SQL representation, not authoritative storage; schema column AggFunction is the source of truth

PR #3515 [docs] Add Apache Polaris/Iceberg REST catalog quick-start guide

@morazow: Suggested adding Polaris-Realm header note and clarifying vended credentials vs static keys paths

PR #3511 [server] Add negative cache for non-existent partition IDs in metadata requests

@swuferhong: Confirmed changed to guava cache implementation

PR #2699 [docs] Reorganize Streaming Lakehouse documentation to highlight deployment and core features

@Copilot: Noted sed replacement writes relative link without .md extension, inconsistent with other docs

📝 新建 Issue/PR

Issues

暂无

Pull Requests

#3519 [lake] Fix DV readable snapshot scanning and premature snapshot deletion @luoyuxia

✅ 关闭 Issue/PR

已关闭 Issues

#3446 [client] Add enum conversion functionality for Typed Writer/Reader @VladBanar

已合并 PRs

#3518 [lake/hudi] Introduce Hudi compaction in tiering service @fhan688
#3508 [CVE][fs] Exclude vulnerable jetty-http dependency @morazow
#3447 [client] Enum conversion functionality for typed writer/reader @VladBanar