Hacktoberfest 2026：维护者为十月标记出来的 issue，仍然开放、适合新手。 浏览 Hacktoberfest issue

storage-cassandra: two concurrent DeleteTable calls on one table both complete

未关闭
  • 哪里不对？
#553 0 条评论 0 个 reaction 已指派 1 人 在 GitHub 查看

维护者通常 2 天内回复

@LeeroyHannigan 已经在做这个了。

开始于 2026年10月25日。

  • #598 来自 @LeeroyHannigan —— 未关闭

评估

这个 Issue 还没有评估数据。

描述

backend:cassandra

Summary

On Cassandra two concurrent DeleteTable calls on one ACTIVE table both run to completion and both answer 200. The service answers one 200 and one ResourceInUseException Attempt to change a resource which is still in use: Table is being deleted: <name> (measured 2026-10-25, five rounds of five). #550 fixes the same pair on SQLite and MongoDB; PostgreSQL already serialized it.

Cause

crates/storage-cassandra/src/delete_table.rs never writes a DELETING status: it reads the row, drops the keyspace tables and deletes the catalog rows inside the request. Nothing fences the second call. The continuous_backups and tags deletes are keyed on the table name and ARN, so a CreateTable of the same name interleaved with the second call's tail can lose the new table's PITR row and tags.

Fix shape

A lightweight-transaction claim to DELETING as the first write (UPDATE ... SET table_status = 'DELETING' IF table_status = 'ACTIVE'); the loser answers TableBeingDeleted (added in #550). Key the continuous_backups and tags deletes on the table id, or run them before the claim is released. A direct test with ten rounds of two concurrent calls asserting one Ok and one TableBeingDeleted, and that a table re-created between the calls keeps its tags and PITR row.

Found while reviewing #505.

主要语言
Rust
星标
517
派生
60
平均合并
2 天 22 小时
30 天内合并 PR
25

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。

ExtendDB/extenddb 的其他 Issue

查看 ExtendDB/extenddb 的全部 Issue

相似的 Issue

更多 Rust Issue

把新 issue 发到你的邮箱

精选适合新手参与的 GitHub issue 摘要。