sysbench是一個(gè)非常經(jīng)典的綜合性能測(cè)試工具，它支持CPU，IO，內(nèi)存，尤其是數(shù)據(jù)庫(kù)的性能測(cè)試。那它是怎么做到通用性的呢，總結(jié)一句話(huà)是大量運(yùn)用了重載的方法。

sysbench總體架構(gòu)

sysbench是一個(gè)總體框架，它用來(lái)操作各個(gè)測(cè)性能的計(jì)算，那各個(gè)部門(mén)只需要做的一件事情是聲明需要的實(shí)現(xiàn)。只要理解了這三個(gè)struct就可以了：

/*?某個(gè)測(cè)試用例的整體結(jié)構(gòu)?*/typedef?struct?sb_test
{??const?char????????*sname;??const?char????????*lname;??/*?下面有具體說(shuō)明?*/
??sb_operations_t???ops;??sb_builtin_cmds_t?builtin_cmds;??sb_arg_t??????????*args;??sb_list_item_t????listitem;
}?sb_test_t;/*?某個(gè)測(cè)試用例的具體操作實(shí)現(xiàn)結(jié)構(gòu)?*/typedef?struct{
??sb_op_init????????????*init;????????????/*?initialization?function?*/
??sb_op_prepare?????????*prepare;?????????/*?called?after?timers?start,??but
?????????????????????????????????????????????before?thread?execution?*/
??sb_op_thread_init?????*thread_init;?????/*?thread?initialization
?????????????????????????????????????????????(called?when?each?thread?starts)?*/
??sb_op_print_mode??????*print_mode;??????/*?print?mode?function?*/
??sb_op_next_event??????*next_event;??????/*?event?generation?function?*/
??sb_op_execute_event???*execute_event;???/*?event?execution?function?*/
??sb_op_report??????????*report_intermediate;?/*?intermediate?reports?handler?*/
??sb_op_report??????????*report_cumulative;???/*?cumulative?reports?handler?*/
??sb_op_thread_run??????*thread_run;??????/*?main?thread?loop?*/
??sb_op_thread_done?????*thread_done;?????/*?thread?finalize?function?*/
??sb_op_cleanup?????????*cleanup;?????????/*?called?after?exit?from?thread,
?????????????????????????????????????????????but?before?timers?stop?*/?
??sb_op_done????????????*done;????????????/*?finalize?function?*/}?sb_operations_t;/*?某個(gè)測(cè)試用例的三階段實(shí)現(xiàn)結(jié)構(gòu)?*/typedef?struct{??sb_builtin_cmd_func_t?*help;??/*?print?help?*/
??sb_builtin_cmd_func_t?*prepare;?/*?prepare?for?the?test?*/
??sb_builtin_cmd_func_t?*run;???/*?run?the?test?*/
??sb_builtin_cmd_func_t?*cleanup;?/*?cleanup?the?test?database,?files,?etc.?*/}?sb_builtin_cmds_t;

拿最簡(jiǎn)單的CPU性能計(jì)算舉例，它需要實(shí)現(xiàn)的是：

static?sb_test_t?cpu_test?=
{
??.sname?=?"cpu",?/*case簡(jiǎn)稱(chēng)*/
??.lname?=?"CPU?performance?test",/*case全稱(chēng)*/
??.ops?=?{
????.init?=?cpu_init,?/*?初始化case?*/
????.print_mode?=?cpu_print_mode,?/*?case啟動(dòng)前，做說(shuō)明?*/
????.next_event?=?cpu_next_event,?/*?拿到下一個(gè)event的數(shù)據(jù)?*/
????.execute_event?=?cpu_execute_event,?/*?具體執(zhí)行這個(gè)event?*/
????.report_cumulative?=?cpu_report_cumulative,??/*?階段性報(bào)告輸出?*/
????.done?=?cpu_done?/*?case結(jié)束后，處理干凈?*/
??},
??.args?=?cpu_args?/*子case需要的參數(shù)說(shuō)明*/};

看到這個(gè)后，把一個(gè)case需要做的事情描述很清楚了，從需要什么參數(shù)，到初始化，逐個(gè)event執(zhí)行，函數(shù)定義很清晰。sysbench的其他case也都這樣需要一個(gè)完整的結(jié)構(gòu)說(shuō)明，如io操作，則需要多一個(gè)case的prepare和cleandown聲明。
那sysbench的完整流程是怎樣呢？黃色部分是測(cè)試用例需要實(shí)現(xiàn)的。

至此，可以清晰地看到sysbench的框架還是非常好理解。
上面struct里面有個(gè)event概念，不同的測(cè)試event的定義都不一樣：比如CPU的測(cè)試case，一個(gè)event是完成求得小于某數(shù)（默認(rèn)10000）的所有質(zhì)數(shù)。比如fileio的測(cè)試case，一次read或者一次write操作就是一個(gè)event。

sysbench的線(xiàn)程介紹

1. worker_thread具體實(shí)現(xiàn)是怎樣呢：欣賞下sysbench.c里面某子線(xiàn)程是如何執(zhí)行的，代碼非常清晰易懂：

static?int?thread_run(sb_test_t?*test,?int?thread_id){??sb_event_t????????event;??int???????????????rc?=?0;??while?(sb_more_events(thread_id)?&&?rc?==?0)
??{
????event?=?test->ops.next_event(thread_id);????if?(event.type?==?SB_REQ_TYPE_NULL)??????break;

????sb_event_start(thread_id);

????rc?=?test->ops.execute_event(&event,?thread_id);

????sb_event_stop(thread_id);
??}??return?rc;
}

1. intermediate_report線(xiàn)程：周期性輸出性能數(shù)據(jù)，參數(shù)項(xiàng)為：--report-interval=N,對(duì)CPU的測(cè)試用例舉例：sysbench cpu --report-interval=1,截取部分輸出結(jié)果如下：

Threads?started![?1s?]?thds:?1?eps:?922.10?lat?(ms,95%):?1.08[?2s?]?thds:?1?eps:?925.19?lat?(ms,95%):?1.08[?3s?]?thds:?1?eps:?926.00?lat?(ms,95%):?1.08[?4s?]?thds:?1?eps:?926.00?lat?(ms,95%):?1.08[?5s?]?thds:?1?eps:?926.00?lat?(ms,95%):?1.08[?6s?]?thds:?1?eps:?926.00?lat?(ms,95%):?1.08[?7s?]?thds:?1?eps:?925.00?lat?(ms,95%):?1.08[?8s?]?thds:?1?eps:?926.02?lat?(ms,95%):?1.08[?9s?]?thds:?1?eps:?925.99?lat?(ms,95%):?1.08[?10s?]?thds:?1?eps:?924.98?lat?(ms,95%):?1.08

每一秒輸出一個(gè)結(jié)果，eps是每一秒的event數(shù)，lat單位是毫秒，95分位延遲數(shù)據(jù)是1.08

1. checkpoints_report線(xiàn)程：如果嫌周期性輸出不夠多，那么可以在某幾個(gè)時(shí)間點(diǎn)整體輸出，參數(shù)項(xiàng)為:?--report-checkpoints=[LIST,...]
   還是對(duì)CPU測(cè)試用例舉例：sysbench cpu --report-checkpoints=3,8 run,截取部分輸出結(jié)果如下：

Threads?started!

[?3s?]?Checkpoint?report:
CPU?speed:
????events?per?second:???923.01General?statistics:
????total?time:??????????????????????????3.0001s
????total?number?of?events:??????????????2771Latency?(ms):
?????????min:??????????????????????????????????1.08
?????????avg:??????????????????????????????????1.08
?????????max:??????????????????????????????????1.22
?????????95th?percentile:??????????????????????1.08
?????????sum:???????????????????????????????3000.88Threads?fairness:
????events?(avg/stddev):???????????2773.0000/0.00
????execution?time?(avg/stddev):???3.0009/0.00[?8s?]?Checkpoint?report:
CPU?speed:
????events?per?second:???924.47General?statistics:
????total?time:??????????????????????????8.0001s
????total?number?of?events:??????????????4622Latency?(ms):
?????????min:??????????????????????????????????1.08
?????????avg:??????????????????????????????????1.08
?????????max:??????????????????????????????????1.16
?????????95th?percentile:??????????????????????1.08
?????????sum:???????????????????????????????4998.04Threads?fairness:
????events?(avg/stddev):???????????4621.0000/0.00
????execution?time?(avg/stddev):???4.9980/0.00

1. tx_rate_controll線(xiàn)程，控制每秒輸出量的一個(gè)線(xiàn)程：參數(shù)項(xiàng)為:?--rate=N,默認(rèn)是不做控制的。
   還是拿CPU測(cè)試用例舉例，控制每秒跑10個(gè)event：sysbench cpu run --rate=10,截取部分輸出結(jié)果如下：

Running?the?test?with?following?options:
Number?of?threads:?1
Target?transaction?rate:?10/sec
Initializing?random?number?generator?from?current?time


Prime?numbers?limit:?10000

Initializing?worker?threads...

Threads?started!

CPU?speed:
????events?per?second:?????8.87?#沒(méi)那么精準(zhǔn)哈

輸出速率控制在哪里呢？眼尖的人馬上可以看到是在sb_more_events函數(shù)。那sb_more_events函數(shù)主要是做什么呢：

1. 判斷是否超時(shí)，默認(rèn)是10秒

2. 判斷是否到達(dá)最大event數(shù)，如果設(shè)置了的話(huà)

3. 就是速率控制。

綜上，大概介紹了sysbench框架的總體實(shí)現(xiàn)，關(guān)于數(shù)據(jù)庫(kù)性能測(cè)試容下篇再介紹。