前面七講我們基本上介紹完了 Zookeeper C API 的所有內容,本文將結合一個小例子講講如何在你的實際項目中使用 Zookeeper 服務。
設想如下場景:
假設程序 A 需要 7* 24 小時在線對外提供服務,但是 A 程序在生產環境下總是不穩定,時常崩潰,不過幸運的是解決方案很簡單,在 A 程序崩潰以后只需要重啟它就可以了。當然如此簡單的問題你可以提出多種解決方案,比方說自己實現一個服務程序,每隔一定時間去輪詢 A 的狀態,如果發現 A 崩潰了,立即重啟它,並向管理人員報告問題。不過我們並不打算這么做,畢竟本文主題是講 Zookeeper C API 的應用,所以我們采用 Zookeeper 服務來解決該問題。
若采用 Zookeeper 服務可以按照如下方案解決問題,程序 A 在啟動時創建一個臨時(ZOO_EPHEMERAL) znode 節點 /A,然后按照正常流程對外提供服務。另外監控程序對 /A 節點設置監視,當 /A 節點消失(說明 A 程序已經崩潰)時,重啟 A 程序。假設 A 的名稱是 QueryServer,即對外提供查詢服務的程序,具體提供什么查詢服務由應用自身決定,我們這里只是簡單地模擬一下。QueryServer 在啟動時創建一個 /QueryServer 的臨時節點(ZOO_EPHEMERAL),然后,程序 QueryServerd 監控 /QueryServer 節點,當 /QueryServer 節點消失(說明 A 程序已經崩潰)時,重啟 QueryServer 程序。
下面是 QueryServer 的實現代碼:
/* * ============================================================================= * * Filename: QueryServer.c * * Description: QueryServer * * Created: 02/15/2013 08:48:49 PM * * Author: Fu Haiping (forhappy), haipingf@gmail.com * Company: ICT ( Institute Of Computing Technology, CAS ) * * ============================================================================= */ #include <stdbool.h> #include <stdio.h> #include <stdlib.h> #include <string.h> #include <zookeeper/zookeeper.h> #include <zookeeper/zookeeper_log.h> void QueryServer_watcher_g(zhandle_t* zh, int type, int state, const char* path, void* watcherCtx) { if (type == ZOO_SESSION_EVENT) { if (state == ZOO_CONNECTED_STATE) { printf("[[[QueryServer]]] Connected to zookeeper service successfully!\n"); } else if (state == ZOO_EXPIRED_SESSION_STATE) { printf("Zookeeper session expired!\n"); } } } void QueryServer_string_completion(int rc, const char *name, const void *data) { fprintf(stderr, "[%s]: rc = %d\n", (char*)(data==0?"null":data), rc); if (!rc) { fprintf(stderr, "\tname = %s\n", name); } } void QueryServer_accept_query() { printf("QueryServer is running...\n"); } int main(int argc, const char *argv[]) { const char* host = "127.0.0.1:2181,127.0.0.1:2182," "127.0.0.1:2183,127.0.0.1:2184,127.0.0.1:2185"; int timeout = 30000; zoo_set_debug_level(ZOO_LOG_LEVEL_WARN); zhandle_t* zkhandle = zookeeper_init(host, QueryServer_watcher_g, timeout, 0, "hello zookeeper.", 0); if (zkhandle == NULL) { fprintf(stderr, "Error when connecting to zookeeper servers...\n"); exit(EXIT_FAILURE); } // struct ACL ALL_ACL[] = {{ZOO_PERM_ALL, ZOO_ANYONE_ID_UNSAFE}}; // struct ACL_vector ALL_PERMS = {1, ALL_ACL}; int ret = zoo_acreate(zkhandle, "/QueryServer", "alive", 5, &ZOO_OPEN_ACL_UNSAFE, ZOO_EPHEMERAL, QueryServer_string_completion, "zoo_acreate"); if (ret) { fprintf(stderr, "Error %d for %s\n", ret, "acreate"); exit(EXIT_FAILURE); } do { // 模擬 QueryServer 對外提供服務. // 為了簡單起見, 我們在此調用一個簡單的函數來模擬 QueryServer. // 然后休眠 5 秒,程序主動退出(即假設此時已經崩潰). QueryServer_accept_query(); sleep(5); } while(false); zookeeper_close(zkhandle); }
Makefile如下:
all:QueryServer QueryServer:QueryServer.o gcc -L/usr/local/lib/ -lzookeeper_mt -o $@ $^ QueryServer.o:QueryServer.c gcc -DTHREADED -I/usr/local/include/zookeeper -o $@ -c $^ .PHONY:clean clean: rm QueryServer.o QueryServer
QueryServerd 代碼如下:
/* * ============================================================================= * * Filename: QueryServerd.c * * Description: QueryServer daemon using zookeeper. * * Created: 02/15/2013 08:48:49 PM * * Author: Fu Haiping (forhappy), haipingf@gmail.com * Company: ICT ( Institute Of Computing Technology, CAS ) * * ============================================================================= */ #include <unistd.h> #include <stdio.h> #include <stdlib.h> #include <string.h> #include <sys/types.h> #include <sys/wait.h> #include <zookeeper/zookeeper.h> #include <zookeeper/zookeeper_log.h> void QueryServerd_watcher_global(zhandle_t * zh, int type, int state, const char *path, void *watcherCtx); static void QueryServerd_dump_stat(const struct Stat *stat); void QueryServerd_stat_completion(int rc, const struct Stat *stat, const void *data); void QueryServerd_watcher_awexists(zhandle_t *zh, int type, int state, const char *path, void *watcherCtx); static void QueryServerd_awexists(zhandle_t *zh); void QueryServerd_watcher_global(zhandle_t * zh, int type, int state, const char *path, void *watcherCtx) { if (type == ZOO_SESSION_EVENT) { if (state == ZOO_CONNECTED_STATE) { printf("Connected to zookeeper service successfully!\n"); } else if (state == ZOO_EXPIRED_SESSION_STATE) { printf("Zookeeper session expired!\n"); } } } static void QueryServerd_dump_stat(const struct Stat *stat) { char tctimes[40]; char tmtimes[40]; time_t tctime; time_t tmtime; if (!stat) { fprintf(stderr, "null\n"); return; } tctime = stat->ctime / 1000; tmtime = stat->mtime / 1000; ctime_r(&tmtime, tmtimes); ctime_r(&tctime, tctimes); fprintf(stderr, "\tctime = %s\tczxid=%llx\n" "\tmtime=%s\tmzxid=%llx\n" "\tversion=%x\taversion=%x\n" "\tephemeralOwner = %llx\n", tctimes, stat->czxid, tmtimes, stat->mzxid, (unsigned int) stat->version, (unsigned int) stat->aversion, stat->ephemeralOwner); } void QueryServerd_stat_completion(int rc, const struct Stat *stat, const void *data) { // fprintf(stderr, "%s: rc = %d Stat:\n", (char *) data, rc); // QueryServerd_dump_stat(stat); } void QueryServerd_watcher_awexists(zhandle_t *zh, int type, int state, const char *path, void *watcherCtx) { if (state == ZOO_CONNECTED_STATE) { if (type == ZOO_DELETED_EVENT) { printf("QueryServer gone away, restart now...\n"); // re-exists and set watch on /QueryServer again. QueryServerd_awexists(zh); pid_t pid = fork(); if (pid < 0) { fprintf(stderr, "Error when doing fork.\n"); exit(EXIT_FAILURE); } if (pid == 0) { /* child process */ // 重啟 QueryServer 服務. execl("/tmp/QueryServer/QueryServer", "QueryServer", NULL); exit(EXIT_SUCCESS); } sleep(1); /* sleep 1 second for purpose. */ } else if (type == ZOO_CREATED_EVENT) { printf("QueryServer started...\n"); } } // re-exists and set watch on /QueryServer again. QueryServerd_awexists(zh); } static void QueryServerd_awexists(zhandle_t *zh) { int ret = zoo_awexists(zh, "/QueryServer", QueryServerd_watcher_awexists, "QueryServerd_awexists.", QueryServerd_stat_completion, "zoo_awexists"); if (ret) { fprintf(stderr, "Error %d for %s\n", ret, "aexists"); exit(EXIT_FAILURE); } } int main(int argc, const char *argv[]) { const char *host = "127.0.0.1:2181,127.0.0.1:2182," "127.0.0.1:2183,127.0.0.1:2184,127.0.0.1:2185"; int timeout = 30000; zoo_set_debug_level(ZOO_LOG_LEVEL_WARN); zhandle_t *zkhandle = zookeeper_init(host, QueryServerd_watcher_global, timeout, 0, "QueryServerd", 0); if (zkhandle == NULL) { fprintf(stderr, "Error when connecting to zookeeper servers...\n"); exit(EXIT_FAILURE); } QueryServerd_awexists(zkhandle); // Wait for asynchronous zookeeper call done. getchar(); zookeeper_close(zkhandle); return 0; }
Makefile 如下:
all:QueryServerd QueryServerd:QueryServerd.o gcc -L/usr/local/lib/ -lzookeeper_mt -o $@ $^ QueryServerd.o:QueryServerd.c gcc -g -DTHREADED -I/usr/local/include/zookeeper -o $@ -c $^ .PHONY:clean clean: rm QueryServerd.o QueryServerd
首先執行 QueryServerd,
forhappy@haiping-ict:/tmp/QueryServerd$ ./QueryServerd
Connected to zookeeper service successfully!
然后執行 QueryServer,
forhappy@haiping-ict:/tmp/QueryServer$ ./QueryServer QueryServer is running... [[[QueryServer]]] Connected to zookeeper service successfully! [zoo_acreate]: rc = 0 name = /QueryServer
可見 Queryerver 創建了 /QueryServer 節點,5 秒后 QueryServer 模擬程序崩潰而退出,那么此時在 QueryServerd 端輸出如下:
Connected to zookeeper service successfully! QueryServer started... # QueryServerd 感知到 QueryServer 已正常啟動. QueryServer gone away, restart now... # 5 秒鍾后,QueryServer 崩潰,QueryServerd 准備重啟 QueryServer. QueryServer is running... # QueryServer 正在運行,以下 3 行是 QueryServer 輸出結果。 [[[QueryServer]]] Connected to zookeeper service successfully! [zoo_acreate]: rc = 0 name = /QueryServer QueryServer started... # QueryServerd 感知到 QueryServer 已正常啟動. QueryServer gone away, restart now...# 又過了 5 秒鍾后,QueryServer 崩潰,QueryServerd 准備重啟 QueryServer. QueryServer is running... # QueryServer 再次運行,以下 3 行是 QueryServer 輸出結果。 [[[QueryServer]]] Connected to zookeeper service successfully! [zoo_acreate]: rc = 0 name = /QueryServer QueryServer started... # QueryServerd 再次感知到 QueryServer 已正常啟動,如此反復. QueryServer gone away, restart now... QueryServer is running... [[[QueryServer]]] Connected to zookeeper service successfully! [zoo_acreate]: rc = 0 name = /QueryServer QueryServer started...
即 QueryServer 每 5 秒鍾崩潰一次,然后又被 QueryServerd 重啟,模擬了上面的應用場景。
好了 Zookeeper C API 的應用小示例講完了,可能應用場景選取的不好,不過大致可以一些說明問題吧,如果你想看 Zookeeper 更貼近現實的應用場景,可以參考淘寶的一篇文章《ZooKeeper典型應用場景一覽》和 IBM developerWorks 的一篇博文《分布式服務框架 Zookeeper -- 管理分布式環境中的數據》。