Open gypleon opened 7 years ago
OMG!!! I merged these commits for several hours... but they come out again... maybe I made a ridiculous mistake that override the merge operations.... OMG.... actually, "[IO] Add Redis Connector. Fixed merge conflicts." (38ce3c2) is supposed to be the ONLY ONE left.
I am currently searching for safely merging solutions.
Hi Legend, I think the commit problem has been fixed. Please review and leave me comments. Thanks.
This PR is including other commits in master branch which are not merged into experimental branch. Could you redo the rebase?
No problem.
This branch is updated to master now. Please rebase again!
And you can run ./scripts/clang-format.py -o check
to see what does not match the coding style and run ./scripts/clang-format.py -o replace
to fix those errors automatically. Then use lint.py
to double check.
Test cases
- Preparation
- Test 1: CMake
Expectation: running well. Notes:
- Test 2: enwiki from MongoDB to Redis:
Expectation: running well. 4825150 records to be imported. around 1m30s. Notes: MUST wait for 10 minutes before Test 3, because for the first time data were loaded into Redis cluster, it takes time for master-slave synchronization.
- Test 3: enwiki from Redis to Redis:
Expectation: running well. 4825150 records to be imported. around 1m. Notes: On Husky-worker side, "waiting for keys" logs may be printed as expected, which means that some workers are waiting for keys from the next batch.
- Test 4: enwiki WordCount:
Expectation: running well. 4825150 records to be imported. around 50s. Notes: Can be performed consecutively without restart Husky-Master.