MySQL - innodb_buffer_pool_size

테이블이 엄청나게 커지면서 실험에 bottleneck이 되었는데 검색하다보니 innodb_buffer_pool_size 요놈 때문인 거 같다. index를 만들었는데도 너무 느렸는데 확인해보니 innodb_buffer_pool_size가 index데이터를 cache하는 설정이다... 서버에 메모리는 빵빵하니 크게 설정해놓으면 놓을 수록 살짝 메모리DB처럼 되간다고 보면 된다고 하네...

The size in bytes of the buffer pool, the memory area where InnoDB caches table and index data. The default value is 128MB. 

Play with sesame native store for large triples

The Sesame Native Store is reported to scale up to about 100-150 million triples (depending on hardware and data-set characteristics). And Rivuli also mentioned in his blogpost that it is becomeing slow while the data set normally over 600,0000 triples 500MB for inserting files to Native Store of sesame.

Tested for chunk commit in sesame API, see example here, that is, stop auto commit and commit in every chunk size you set.

• Test: commit every 5 triples, how many will be committed if there's exception in 14th triple(statement)?
• A: 10
• B: 13
• Answer: 13 ( It will be committed before the exception point)