database is locked和SQLITE_BUSY

2021-10-29 19:27:14

这是官方网站找到的解释：http://www.sqlite.org/cvstrac/wiki?p=DatabaseIsLocked

下面我做了几个实验：

1、多个线程（pthread），使用同一个句柄（一次sqlite3_open，同一个数据库文件），在多个线程中同时使用此句柄，这些线程中有的是进行select操作的，有的是进行update操作的，使用的都是sqlite3_exec函数；结果是没有发现database is locked错误；

程序源码（mutitrhead.c ）：

#include <sqlite3.h>
#include <stdio.h>
#include <pthread.h>
#include <string.h>
#include "main.h"

#ifdef _WIN32
#include <windows.h>
#define sleep(x) Sleep((x)*1000)
#else
#define sleep(x) sleep(x)
#endif

#define CREATE_SQL              "CREATE TABLE IF NOT EXISTS \"mutiprocess\" ("                      "\"id\"  INTEGER NOT NULL PRIMARY KEY,"               "\"src\"  TEXT(1024),"                           "\"videotype\"  INTEGER NOT NULL,"                      "\"postfiledir\"  TEXT(1024)"                           ");"
#define REPLACE_SQL "REPLACE INTO mutiprocess (id,src,videotype,postfiledir) VALUES(NULL,%d,%d,%d)"
#define UPDATE_SQL "UPDATE mutiprocess SET videotype = ‘2‘ WHERE id = 1"
#define SELECT_SQL "SELECT * FROM mutiprocess"

static void* select_cf(void *param)
{
        int rc = -1;
        sqlite3 *db = (sqlite3*)param;
        char *szErrMsg;
        pthread_t tid;

        tid = pthread_self();
        while(1){
                rc = sqlite3_exec(db ,SELECT_SQL, 0, 0, &szErrMsg);
                if (rc != SQLITE_OK) {
                        printf("tid %u -- select error.[%s] -- rc=[%d]\n", tid, szErrMsg, rc);
                        if(strstr(szErrMsg, "database is locked")){
                                printf("<database is locked> when executing [%s], retry after 1 second.\n", SELECT_SQL);
                        }
                }else{
                        printf("tid %u -- successfully execute [%s].\n", tid, SELECT_SQL);
                }
                sleep(1);
        }
}

static void* update_cf(void *param)
{
        int rc = -1;
        sqlite3 *db = (sqlite3*)param;
        char *szErrMsg;
        pthread_t tid;

        tid = pthread_self();
        while(1){
                rc = sqlite3_exec(db ,UPDATE_SQL, 0, 0, &szErrMsg);
                if (rc != SQLITE_OK) {
                        printf("tid %u -- update error.[%s] -- rc=[%d]\n", tid, szErrMsg, rc);
                        if(strstr(szErrMsg, "database is locked")){
                                printf("<database is locked> when executing [%s], retry after 1 second.\n", UPDATE_SQL);
                        }
                }else{
                        printf("tid %u -- successfully execute [%s].\n", tid, UPDATE_SQL);
                }
                sleep(1);
        }
}

int main()
{
        pthread_t pid[20];
        int rc;
        sqlite3 *db = 0;
        char *szErrMsg;
        int i = 0;
        char sql[1024] = {0};

        rc = sqlite3_open("mutiprocess.db", &db);
        if (rc != SQLITE_OK) {
                printf("open sqlite3 error.\n");
        }

        rc = sqlite3_exec(db ,CREATE_SQL, 0, 0, &szErrMsg);
        if (rc != SQLITE_OK) {
                printf("create db error-[%s]\n", szErrMsg);
        }

        for(i = 0; i < 4; i++){
                sprintf(sql, REPLACE_SQL, i, i, i);
                rc = sqlite3_exec(db ,sql, 0, 0, &szErrMsg);
                if (rc != SQLITE_OK) {
                        printf("replace db error-[%s]\n", szErrMsg);
                }
        }

        for(i = 0; i < 10; i++){
                pthread_create(&(pid[i]), 0, select_cf, db);
                pthread_detach(pid[i]);
        }
        for(; i < 20; i++){
                pthread_create(&(pid[i]), 0, update_cf, db);
                pthread_detach(pid[i]);
        }

        while(1){
                sleep(-1);
        }
        return 0;
}

编译：gcc -o mutitrhead mutitrhead.c -lsqlite3 -lpthread

运行：./mutithread

注意：只运行这一个进程没有发现任何database is locked的错误提示；如果在运行一个./mutithread进程，那么两个进程都会出现database is locked错误；如果kill掉其中一个进程，那么另外一个不再出现database is locked；

注意两个进程和一个进程的区别，一个是进程数不同，一个是使用的数据库句柄连接不同；上述程序虽然在一个进程中使用了多线程，但是多个线程都是使用了同一个数据库连接（使用一个sqlite3_open返回的句柄），不好区分是什么问题（其实官方网站有一句话很重要『SQLITE_BUSY means that another database connection (probably in another process) is using the database in a way that prevents you from using it』）；这其实已经说明了产生database is locked的原因了，就是出在多个连接上，那么我们在做下面的实验；

2、同一个进程，启动多个线程，每个线程中都打开一个连接（connection，使用一个sqlite3_open返回的句柄），并且在其中做select或者update的操作；结果会出现database is locked错误；

程序源码（muticonnection.c）：

#include <sqlite3.h>
#include <stdio.h>
#include <pthread.h>
#include <string.h>
#include "main.h"

#ifdef _WIN32
#include <windows.h>
#define sleep(x) Sleep((x)*1000)
#else
#define sleep(x) sleep(x)
#endif

#define CREATE_SQL              "CREATE TABLE IF NOT EXISTS \"mutiprocess\" ("                      "\"id\"  INTEGER NOT NULL PRIMARY KEY,"               "\"src\"  TEXT(1024),"                           "\"videotype\"  INTEGER NOT NULL,"                      "\"postfiledir\"  TEXT(1024)"                           ");"
#define REPLACE_SQL "REPLACE INTO mutiprocess (id,src,videotype,postfiledir) VALUES(NULL,%d,%d,%d)"
#define UPDATE_SQL "UPDATE mutiprocess SET videotype = ‘2‘ WHERE id = 1"
#define SELECT_SQL "SELECT * FROM mutiprocess"

static void* select_cf(void *param)
{
        int rc;
        sqlite3 *db = 0;
        char *szErrMsg;
        int i = 0;
        char sql[1024] = {0};
        pthread_t tid;

        rc = sqlite3_open("mutiprocess.db", &db);
        if (rc != SQLITE_OK) {
                printf("open sqlite3 error.\n");
        }

        rc = sqlite3_exec(db ,CREATE_SQL, 0, 0, &szErrMsg);
        if (rc != SQLITE_OK) {
                printf("create db error-[%s]\n", szErrMsg);
        }

        for(i = 0; i < 4; i++){
                sprintf(sql, REPLACE_SQL, i, i, i);
                rc = sqlite3_exec(db ,sql, 0, 0, &szErrMsg);
                if (rc != SQLITE_OK) {
                        printf("replace db error-[%s]\n", szErrMsg);
                }
        }

        tid = pthread_self();
        while(1){
                rc = sqlite3_exec(db ,SELECT_SQL, 0, 0, &szErrMsg);
                if (rc != SQLITE_OK) {
                        printf("tid %u -- select error.[%s] -- rc=[%d]\n", tid, szErrMsg, rc);
                        if(strstr(szErrMsg, "database is locked")){
                                printf("<database is locked> when executing [%s], retry after 1 second.\n", SELECT_SQL);
                        }
                }else{
                        printf("tid %u -- successfully execute [%s].\n", tid, SELECT_SQL);
                }
                sleep(1);
        }
}

static void* update_cf(void *param)
{
        int rc;
        sqlite3 *db = 0;
        char *szErrMsg;
        int i = 0;
        char sql[1024] = {0};
        pthread_t tid;

        rc = sqlite3_open("mutiprocess.db", &db);
        if (rc != SQLITE_OK) {
                printf("open sqlite3 error.\n");
        }

        rc = sqlite3_exec(db ,CREATE_SQL, 0, 0, &szErrMsg);
        if (rc != SQLITE_OK) {
                printf("create db error-[%s]\n", szErrMsg);
        }

        for(i = 0; i < 4; i++){
                sprintf(sql, REPLACE_SQL, i, i, i);
                rc = sqlite3_exec(db ,sql, 0, 0, &szErrMsg);
                if (rc != SQLITE_OK) {
                        printf("replace db error-[%s]\n", szErrMsg);
                }
        }

        tid = pthread_self();
        while(1){
                rc = sqlite3_exec(db ,UPDATE_SQL, 0, 0, &szErrMsg);
                if (rc != SQLITE_OK) {
                        printf("tid %u -- update error.[%s] -- rc=[%d]\n", tid, szErrMsg, rc);
                        if(strstr(szErrMsg, "database is locked")){
                                printf("<database is locked> when executing [%s], retry after 1 second.\n", UPDATE_SQL);
                        }
                }else{
                        printf("tid %u -- successfully execute [%s].\n", tid, UPDATE_SQL);
                }
                sleep(1);
        }
}

int main()
{
        pthread_t pid[20];
        int i = 0;

        for(i = 0; i < 10; i++){
                pthread_create(&(pid[i]), 0, select_cf, 0);
                pthread_detach(pid[i]);
        }
        for(; i < 20; i++){
                pthread_create(&(pid[i]), 0, update_cf, 0);
                pthread_detach(pid[i]);
        }

        while(1){
                sleep(-1);
        }
        return 0;
}

编译：gcc -o muticonnection muticonnection.c -lsqlite3 -lpthread

运行：./muticonnection

总结：看来出现这个错误是因为多数是因为使用了多个连接(多个sqlite3_open返回的句柄)导致的，这里我的说法是“多数”，因为还有别的原因，请见官方网站的解释：http://www.sqlite.org/cvstrac/wiki?p=DatabaseIsLocked

还有一点要注意：

database is locked是SQLITE_BUSY的错误，不要和SQLITE_LOCKED混淆；

SQLITE_LOCKED（6） means the source of contention is internal and comes from the same database connection that received the SQLITE_LOCKED error.

SQLITE_BUSY（5） means that another database connection (probably in another process) is using the database in a way that prevents you from using it.

附：SQLITE_LOCKED(6)错误见：http://sqlite.1065341.n5.nabble.com/Begin-immediate-transaction-gt-SQLITE-BUSY-database-is-locked-td64878.html

database is locked和SQLITE_BUSY

码农公寓

相关文章