No refreshing Iceberg credential provider on Spark 3.x builds

未关闭
#7,389 0 条评论 0 个 reaction 已指派 0 人 在 GitHub 查看

维护者通常 2 天内回复

还没有人认领这个 Issue。

评估

难度
4/5
预计耗时
3-5 天
新手友好度
35/100
Issue 类型
功能
描述清晰度
基本清楚
活跃度
活跃
技术栈
aws, scala, spark

调研方向

Start with IcebergVendedS3CredentialProvider, which exists only in the spark-4.x source set, and compare it with the Iceberg 1.8.1 AwsClientProperties refresh logic that Comet loads by reflection. Decide whether the spark-3.x set gets its own SDK v2 body or a loader-neutral way to read the credentials, and settle that with maintainers first. Done means the provider is available on Spark 3.x builds and the docs from #7363 no longer say it is missing.

由索引模型根据 Issue 内容生成。

描述

area:Iceberg area:scan enhancement spark 3.x
Is your feature request related to a problem or challenge?

#7363 adds IcebergVendedS3CredentialProvider in the spark-4.x source set only, because it links against AWS SDK v2. Iceberg 1.8.1, which the Spark 3.5 build uses, already has the same refresh logic in AwsClientProperties, and Comet loads Iceberg by reflection. On Spark 3.x, a long native Iceberg scan against a REST catalog with short-lived credentials still fails with ExpiredToken.

Describe the solution you'd like

Ship the provider in Spark 3.x builds too. Iceberg's provider is SDK v2 there as well, so the class would need its own body in the spark-3.x source set, or a loader-neutral way to read the SDK v2 credentials.

Additional context

The docs added in #7363 state that Spark 3.x builds don't include the provider.

主要语言
Scala
星标
1.3k
派生
414
平均合并
3 天 5 小时
30 天内合并 PR
210

环境准备

从这里开始

  1. 先读完整个 Issue,再读项目的贡献指南。
  2. 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
  3. Fork 仓库,在一个分支上完成修改。
  4. 提交 Pull Request,并在描述里引用这个 Issue 编号。

apache/datafusion-comet 的其他 Issue

查看 apache/datafusion-comet 的全部 Issue

相似的 Issue

更多 Scala Issue

把新 issue 发到你的邮箱

精选适合新手参与的 GitHub issue 摘要。