Github user jinxing64 commented on the issue:
https://github.com/apache/spark/pull/18482
In current change, it i fetching big chunk in memory and then writing to
disk and then release the memory. I made this change for below reasons:
1. The client shouldn't break old shuffle service, thus cannot send "stream
request" to server. We have to send `ChunkFetchRequest` and handle the
`ChunkFetchSuccess` for response.
2. It's hard to make 'ChunkFetchSuccess' to be a stream and read it to
disk. We need to implement another `TransportFrameDecoder`, which I think is
too much cost.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
