[ 
https://issues.apache.org/jira/browse/FLINK-40083?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18119242#comment-18119242
 ] 

Martijn Visser commented on FLINK-40083:
----------------------------------------

Also backported to resolve FLINK-40802

release-1.20: 0d8fcb2ecfd528086d6665b7d9bb63ca0ff38edb and 
5998abe4a9af954703cbcfbc0ca27da40a53f402
release-2.2: 237c1f682500ffb5d7427c10c049cca957230f23 and 
1ffaa38d7655dd53c3884ca037f653a4c732cfc3
release-2.3: f68ea74de58054eee21b1471e9f4572bbc83e985 and 
1db44c21e2f8d15884c6a7f63b7089ed1e9aa775

> Launch the SeaweedFS S3 test backend via weed mini and bump the image to 4.38
> -----------------------------------------------------------------------------
>
>                 Key: FLINK-40083
>                 URL: https://issues.apache.org/jira/browse/FLINK-40083
>             Project: Flink
>          Issue Type: Improvement
>          Components: Tests
>            Reporter: Chris Lu
>            Assignee: Chris Lu
>            Priority: Major
>              Labels: pull-request-available
>             Fix For: 2.4.0
>
>
> The S3 filesystem tests use SeaweedFS as their S3 backend (FLINK-39499). Two 
> improvements:
>  # Launch the container with "weed mini", which pre-creates the bucket via 
> "-bucket" and reports readiness only after all components are up — removing 
> the client-side bucket creation, the startup-503 retry workaround, and the 
> /healthz poll from both SeaweedFsTestContainer and common_s3_seaweedfs.sh.
>  # Bump the pinned image chrislusf/seaweedfs 4.34 → 4.38. 4.38 answers S3 
> directory-path probes the way AWS does 
> (https://github.com/seaweedfs/seaweedfs/pull/10225) so Flink 
> savepoints/checkpoints restore correctly against SeaweedFS, and adds 
> multipart / truncated-PUT robustness fixes.
> PR: [https://github.com/apache/flink/pull/28560]



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to