Coprocessor read should not get started before the related MPPTask get scheduled.
Nobody has claimed this yet.
- Dominant language
- C++
- Stars
- 1k
- Forks
- 423
- Avg merge
- 1d 15h
- Merged PRs (30d)
- 24
Description
Enhancement
For MPPTask with table scan, if some regions are not in current TiFlash node, we use CoprocessorReader to pull the data from other TiFlash nodes. Current implementation will start coprocessor read immediately after CoprocessorReader is constructed. After https://github.com/pingcap/tiflash/pull/4044 we introduce a scheduler for MPP queries, a MPP task should only get started after it is scheduled, since CoprocessorReader is part of mpp task, we should make sure that the coprocessor read should not get started before related MPPTask get scheduled.
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start in dbms/src/Flash/Coprocessor/DAGQueryBlockInterpreter.cpp at the CoprocessorReader reference, then read the MPP scheduler changes introduced by pull request #4044. Trace when the coprocessor read starts relative to MPPTask scheduling. Done means the read does not begin until its related MPPTask has been scheduled.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp
- Domain
- databases, distributed-systems
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 42/100