[Bug] version 3.0.3 has setting for muzzy_decay_ms and dirty_decay_ms as 15000 (vs 5000)

[patrickdung2022]

Search before asking

• I had searched in the issues and found no similar issues.

Version

3.0.3

What's Wrong?

In the admin manual:
https://doris.apache.org/docs/3.0/admin-manual/trouble-shooting/memory-management/memory-analysis/jemalloc-memory-analysis

It is mentioned that
Reduce dirty_decay_ms of JEMALLOC_CONF in be.conf to 2000 ms or less. The default dirty_decay_ms in be.conf is 5000 ms. Jemalloc will release dirty pages according to a smooth gradient curve within the time specified by dirty_decay_ms. For reference, Jemalloc opt.dirty_decay_ms. When the BE process has insufficient available memory and triggers Minor GC or Full GC, it will actively release all dirty pages according to a certain strategy.

Before Doris 2.1, the default value of dirty_decay_ms in JEMALLOC_CONF in be.conf is 15000, which will cause the Jemalloc Cache to be too large in some scenarios. After Doris 2.1, the default value is 5000.

When compare below config (v2.1.7 and v3.0.3):
https://github.com/apache/doris/blob/2.1.7-rc03/conf/be.conf
https://github.com/apache/doris/blob/3.0.3-rc04/conf/be.conf

It is found that v2.1.7 has updated JEMALLOC_CONF to muzzy_decay_ms:5000,dirty_decay_ms:5000
But the same setting in v3.0.3 is still muzzy_decay_ms:15000,dirty_decay_ms:15000

Does Doris version 3.0.x actually needs to update these settings to 5000?

What You Expected?

Setting same as manual

How to Reproduce?

No response

Anything Else?

No response

Are you willing to submit PR?

• Yes I am willing to submit a PR!

Code of Conduct

• I agree to follow this project's Code of Conduct