summaryzb
summaryzb
@colinmjj @jerqi PTAL
> @summaryzb For this optimization, I think there has once more `System.arraycopy` to add record, and it will impact performance a lot with previous test. For this PR, I think...
> but I prefer to avoid possible performance regression with more resources. Well, how about adding a config to make this pr as an optional strategy, while current implementation is...
no obvious benefit can be gather from this pr, close it
please assign this to me
I'll look into the GHA action, add one in this patch later
> Is it possible to reuse `ColumnarBatchSerializer` somehow ? Yes, it's almost the same. Change it to reuse `ColumnarBatchSerializer`
I'm in, Please assign this to me
@xianjingfeng @zuston Maybe the root cause is that we change the storage or storage manager in `hadoopThreadPoolExecutor` or `localFileThreadPoolExecutor` which is designed to handle specified storage type. The `storage` and...
To solve this issue and make the storage selection logic more clear, may be we could try the below way 1. Mark and get `storage` or `stageManager` only through `ShuffleDataFlushEvent`...