Benchmark results of LeoFS v0.14.0
Apr 18, 2013
We made benchmarks for LeoFS v0.14.0 on Apr 12th, 2013. テスト環境は以下の通りです。一般的なスペックのサーバとCentOS 6.3を使いました。Also, LeoFS’s consistency level was the same as production settings.
Test Environment
| アイテム |
値 |
|---|
| Hardware |
| CPU |
8-core |
| RAM |
16GB |
| HDD |
7200rpm (Capacity:1TB) |
| Network |
10Gbps ethernet |
| OS/Middleware |
| OS |
Linux 2.6.32-279.22.1.el6.x86_64 |
| Erlang |
Erlang R15B03-1 (erts-5.9.3.1) |
| LeoFS Cluster |
| # of Benchmarker |
1 |
| # of LeoFS-Manager |
1 |
| # of LeoFS-Gateway |
2 |
| # of LeoFS-Storage |
5 |
| Consistency Level |
| 複製の数 |
3 |
| # of successful WRITE |
2 |
| # of successful GET |
1 |
| Prerequirements |
| # of loaded objects |
100,000 |
| Stress Tool Settings (basho_bench’s configuration file) |
| R:8 W:2, “exponential_bin” |
Test Results
Test results is as follows. From each benchmark, we found the bottle-neck was disc I/O. Also, retrieving of small file size (average 128KB) has the room for an improvement. So We have been improving Leo's object-cache lib from this week.
Detail benchmark configuration for average of 128KB (file size)
| アイテム |
値 |
|---|
| Mode |
Max |
| Duration |
1,000 sec |
| # of concurrents |
64 |
| Driver |
basho_bench_driver_leofs |
| Misc |
| http_raw_request_timeout |
30,000ms (30sec) |
| value_generator_source_size |
4,194,304 bytes |
| key_generator |
partitioned_sequential_int: 100,000 |
| value_generator |
exponential_bin: min-size:16,384 bytes, avg-size:524,288 bytes |
| operations |
READ : WRITE = 80% : 20% |
|
Detail benchmark configuration for average of 1MB (file size)
| アイテム |
値 |
|---|
| Mode |
Max |
| Duration |
1,000 sec |
| # of concurrents |
64 |
| Driver |
basho_bench_driver_leofs |
| Misc |
| http_raw_request_timeout |
30,000ms (30sec) |
| value_generator_source_size |
4,194,304 bytes |
| key_generator |
partitioned_sequential_int, 100,000 |
| value_generator |
exponential_bin: min-size:16,384 bytes, avg-size:1,048,576 bytes |
| operations |
READ : WRITE = 80% : 20% |
|
索引