Support in-memory segment generation and upload
Open
- Dominant language
- Java
- Stars
- 6.1k
- Forks
- 1.5k
- Avg merge
- 2d 3h
- Merged PRs (30d)
- 195
Description
Currently the SegmentWriter and SegmentUploader writes data into a TAR file and then upload to Pinot servers.
Can we optimize the write to disk by creating a in-memory TAR file format to speed up the upload?
Contributor guide
Research direction
Start by reading SegmentWriter and SegmentUploader to trace how data is written into a TAR file and uploaded to Pinot servers. Define the in-memory TAR approach and compare it with the current disk-based flow; done means segment generation and upload work without the intermediate disk write while preserving existing behavior.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- java
- Domain
- databases
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100