eolivelli opened a new pull request, #3598: URL: https://github.com/apache/celeborn/pull/3598
### What changes were proposed in this pull request? Add new configuration `celeborn.worker.reserve.slots.io.threads` (default 0) to set the number of threads to use to create the files when reserving the slots. Zero means to use a number computed from the number of available cpus. In order to have the same behavior as in 0.6.0 you can set `celeborn.worker.reserve.slots.io.threads` to 1. ### Why are the changes needed? When using S3 the createPartitionDataWriter method can be slow, because it creates files on S3. This is only IO and it can be parallelize to speed up the bootstrap of the shuffle operation. ### Does this PR resolve a correctness bug? No ### Does this PR introduce _any_ user-facing change? Yes It adds a new configuration `celeborn.worker.reserve.slots.io.threads`. The new behavior is to parallelize the creation of files. ### How was this patch tested? CI and Manual testing -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
