matrixone icon indicating copy to clipboard operation
matrixone copied to clipboard

[Bug]: some `select` against `statement_info` failed during bvt test

Open sukki37 opened this issue 2 years ago • 46 comments

Is there an existing issue for the same bug?

  • [X] I have checked the existing issues.

Environment

- Version or commit-id (e.g. v0.1.0 or 8b23a93):
- Hardware parameters:
- OS type:
- Others:

Actual Behavior

https://github.com/matrixorigin/matrixone/actions/runs/6541067487/job/17762029555

image

Expected Behavior

No response

Steps to Reproduce

No response

Additional information

No response

sukki37 avatar Oct 17 '23 14:10 sukki37

repro: https://github.com/matrixorigin/matrixone/actions/runs/6548014027/job/17781709874

sukki37 avatar Oct 17 '23 14:10 sukki37

今天看看

gavinyue avatar Oct 20 '23 16:10 gavinyue

试了一下这个, 就是又个select statement 没有被记录。 我需要开了race 再测试一下。

gavinyue avatar Oct 21 '23 07:10 gavinyue

https://github.com/matrixorigin/matrixone/issues/12042

都是写入偶尔的问题。 看看使用load之后, 是不是还存在

gavinyue avatar Oct 25 '23 19:10 gavinyue

Pending

gavinyue avatar Oct 31 '23 18:10 gavinyue

Pending

gavinyue avatar Nov 06 '23 15:11 gavinyue

Pending

gavinyue avatar Nov 11 '23 18:11 gavinyue

Pending

gavinyue avatar Nov 16 '23 15:11 gavinyue

Pending

gavinyue avatar Nov 21 '23 15:11 gavinyue

Pending

gavinyue avatar Nov 25 '23 18:11 gavinyue

When running MO for BVT testing with race enabled, errors related to statement_info queries occur frequently.

https://github.com/matrixorigin/matrixone/actions/runs/7036372795/job/19149429051 https://github.com/matrixorigin/matrixone/actions/runs/7043577076/job/19170458791 https://github.com/matrixorigin/matrixone/actions/runs/7040613001/job/19162186013

sukki37 avatar Nov 30 '23 09:11 sukki37

Pending

gavinyue avatar Dec 05 '23 14:12 gavinyue

这个如果需要ctl cmd来flush的话, 需要很大的工作。

1.1 恐怕完成不了

gavinyue avatar Dec 08 '23 04:12 gavinyue

主要原因是 stmt是分散在各个cn上的。

而且就算有了这个命令, 只是缩短了决定要写的时间, 但是真正写的时间还是不可控。

gavinyue avatar Dec 08 '23 04:12 gavinyue

Pending

gavinyue avatar Dec 13 '23 14:12 gavinyue

Pending

gavinyue avatar Dec 20 '23 12:12 gavinyue

暂时没来及的解决

gavinyue avatar Dec 25 '23 14:12 gavinyue

Pending

gavinyue avatar Dec 29 '23 14:12 gavinyue

repro: https://github.com/matrixorigin/matrixone/actions/runs/7346060546/job/20000210682?pr=13857 https://github.com/matrixorigin/matrixone/actions/runs/7396619125/job/20122259309 https://github.com/matrixorigin/matrixone/actions/runs/7397232166/job/20123977845 https://github.com/matrixorigin/matrixone/actions/runs/7396619125/job/20122259309

sukki37 avatar Jan 03 '24 14:01 sukki37

正在看

gavinyue avatar Jan 04 '24 03:01 gavinyue

在不聚合一个statement info pr 里 解决中

gavinyue avatar Jan 04 '24 03:01 gavinyue

repro: https://github.com/matrixorigin/matrixone/actions/runs/7404727910/job/20146654826?pr=13923 https://github.com/matrixorigin/matrixone/actions/runs/7404421431/job/20145908049?pr=13963

xzxiong avatar Jan 04 '24 03:01 xzxiong

下载了error report

gavinyue avatar Jan 05 '24 00:01 gavinyue

repro: https://github.com/matrixorigin/matrixone/actions/runs/7408832551/job/20157854807?pr=14003

sukki37 avatar Jan 05 '24 01:01 sukki37

问题已经找到, 和聚合没有关系,主要是在写入statementinfo的时候, 会出现失败,然后数据去了csv, 然后因为测试的case是select 然后对比结果, 所以它就会和预期结果不一样。

gavinyue avatar Jan 05 '24 03:01 gavinyue

https://github.com/matrixorigin/matrixone/pull/13876

对老失败的地方做了一些修改。

整体来说select from statementinfo 是个比较容易被影响的测试。

gavinyue avatar Jan 05 '24 21:01 gavinyue

https://github.com/matrixorigin/matrixone/actions/runs/7435891547/attempts/1?pr=13876

基本上修复, 正在多跑几遍验证

gavinyue avatar Jan 07 '24 04:01 gavinyue

跑了五六遍, 都没什么问题, 准备merge了

gavinyue avatar Jan 07 '24 14:01 gavinyue

https://github.com/matrixorigin/matrixone/actions/runs/7435891547/attempts/1?pr=13876

等待approval 然后 merge

gavinyue avatar Jan 09 '24 17:01 gavinyue

正在等merge

gavinyue avatar Jan 11 '24 13:01 gavinyue