[Bug]: [0105 main standalone regression]tpcc 100-1000 test oom.

[Ariznawlll]

Is there an existing issue for the same bug?

• I have checked the existing issues.

Branch Name

main

Commit ID

04a3ebb

Other Environment Information

- Hardware parameters:
- OS type:
- Others:

Actual Behavior

job url: https://github.com/matrixorigin/mo-auto-test/actions/runs/12609544410/job/35143260248

heap_profile:
LOG_7c4dccb4-4d3c-41f8-b482-5251dc7a41bf_heap_019431f2-2043-77fd-bfd0-03be3115f4df.gz
LOG_7c4dccb4-4d3c-41f8-b482-5251dc7a41bf_heap_019431f2-9574-73ff-8b03-40e71dc5889b.gz
LOG_7c4dccb4-4d3c-41f8-b482-5251dc7a41bf_heap_019431f8-b3f1-7bc9-ad78-45ba48c19568.gz

malloc_profile:
LOG_7c4dccb4-4d3c-41f8-b482-5251dc7a41bf_malloc_019431f9-297b-72af-9bc5-1ef4fcd0369c.gz
LOG_7c4dccb4-4d3c-41f8-b482-5251dc7a41bf_malloc_019431f8-b457-717d-8245-26d09d0565d2.gz
LOG_7c4dccb4-4d3c-41f8-b482-5251dc7a41bf_malloc_019431f2-2953-7ee2-ab50-7f0f09a60641.gz

Expected Behavior

No response

Steps to Reproduce

trigger workflow on 129.

Additional information

No response

[daviszhen]

[daviszhen]

[daviszhen]

[Ariznawlll]

复现了：
job：https://github.com/matrixorigin/mo-auto-test/actions/runs/12668310895/job/35303728630

metric：oom时间段2025-01-08 23:29:53-2025-01-08 23:38:05（utc+8 ）
https://shanghai.idc.matrixorigin.cn:30001/d/ce1wtrqp5urk0d/memory-metrics?orgId=1&var-interval=1m&var-datasource=prometheus-standalone&var-instance=10.222.1.129:7001&from=1736350193000&to=1736350685000

malloc profile:
LOG_7c4dccb4-4d3c-41f8-b482-5251dc7a41bf_malloc_0194468c-40d0-750c-bf6d-c6f1328eb375.gz
LOG_7c4dccb4-4d3c-41f8-b482-5251dc7a41bf_malloc_0194468c-b5c6-7612-a493-7244f0e15294.gz

heap profile:
LOG_7c4dccb4-4d3c-41f8-b482-5251dc7a41bf_heap_0194468c-379c-7854-a26b-b6336bf4158b.gz
LOG_7c4dccb4-4d3c-41f8-b482-5251dc7a41bf_heap_0194468c-accc-72d7-bcac-fdde72cee4f4.gz



[reusee]

堆外内存无异常：https://shanghai.idc.matrixorigin.cn:30001/d/ce1wtrqp5urk0d/memory-metrics?orgId=1&var-interval=1m&var-datasource=prometheus-standalone&var-instance=10.222.1.129%3A7001&from=1736350193000&to=1736350685000

go堆内存有较大上升：https://shanghai.idc.matrixorigin.cn:30001/d/be1wtrqrb9m9sa/go-runtime-metrics?orgId=1&var-interval=1m&var-datasource=prometheus-standalone&var-instance=10.222.1.129%3A7001&from=1736350193000&to=1736350685000

应该是go堆内存上升导致OOM

[reusee]

从go堆heap profile看，in-use里面，logtail占了9G左右，table scan占了3.6G左右，应该优先优化

[reusee]

对比OOM前几分钟内的两个 allocs profile，得到期间分配的信息：

LOG_7c4dccb4-4d3c-41f8-b482-5251dc7a41bf_heap_0194468a-d809-7aeb-8c55-6634c67da734.gz
LOG_7c4dccb4-4d3c-41f8-b482-5251dc7a41bf_heap_0194468c-379c-7854-a26b-b6336bf4158b.gz

总共分配了 368G：

细分：

每种对象分配的比例都不算很高，但加起来多。

[reusee]

再对比 in-use：

logtail占用的空间，上升比较多。以及观察到多了 DoMergeAndWrite 这部分。

[Ariznawlll]

commit：abe8c3298eda34c0e395684c88682699fb73aeef

job：https://github.com/matrixorigin/mo-auto-test/actions/runs/12732510081/job/35487805687

metric：https://shanghai.idc.matrixorigin.cn:30001/d/ce1wtrqp5urk0d/memory-metrics?orgId=1&var-interval=1m&var-datasource=prometheus-standalone&var-instance=10.222.1.129:7001&from=1736695802000&to=1736696042000

malloc profile：
LOG_7c4dccb4-4d3c-41f8-b482-5251dc7a41bf_malloc_01945b24-22da-7903-9dbe-99065bcf06cd.gz
LOG_7c4dccb4-4d3c-41f8-b482-5251dc7a41bf_malloc_01945b24-98bc-7d00-87e2-f1ac1986802a.gz

heap profile：
LOG_7c4dccb4-4d3c-41f8-b482-5251dc7a41bf_heap_01945b24-8e93-71cc-bab7-a9c8e9989069.gz
LOG_7c4dccb4-4d3c-41f8-b482-5251dc7a41bf_heap_01945b24-1962-72cc-8a2a-b2085d26fa56.gz

[XuPeng-SH]

fixed

[Ariznawlll]

今天测试结果正常

https://github.com/matrixorigin/mo-auto-test/actions/runs/12806494085/job/35715045035