Asynchronous log file module implements
Before the first time in the company to do a project, did not feel any problem; but later found the time to write the log, every time synchronization in writing, especially writing papers, this is time-consuming, so wanted to optimize it;
Optimization:
-
Asynchronous read and write
Here is the use of asynchronous queues do
Each module will log write queue, do not care about the success or failure is written; a thread is created specifically for queue read log information, write to the log file
-
Scenario Figure
[Picture dump outside the chain fails, the source station may have a security chain mechanism, it is recommended to save the pictures uploaded directly down (img-tRTy7K6R-1582939007578) (D: \ study concluded \ Network Communications \ asynchronous log file \ 1.png)]
The benefits of using queue:
- Decoupling, so that each module independent, complementary Effect
- Improve performance; each module there is no loss write files, all losses by the logging module to write files to bear;
achieve
- This implementation is that the queues used in STL queue to bottom, ensures that the global mode using a single embodiment there is only one object, to ensure use of the same queue;
- Use open source library log4cplus as reading and writing journal library
Note: Generally a different module is a thread, this time on the need to lock, otherwise there will be access to conflict issues;
Here wrote a simple example, for reference only:
#pragma once
#include <iostream>
#include <queue>
#include <windows.h>
#include <mutex>
using namespace std;
struct MSGLOG
{
//0 info;1 error;
int type;
char info[256];
};
class QueueLog
{
public:
QueueLog();
static QueueLog& instance();
~QueueLog();
bool SetLog(string logmsg,int type);
MSGLOG GetLog();
private:
MSGLOG msg;
queue<MSGLOG>que;
std::mutex _mutex;
};
#define LOG QueueLog::instance()
#include "QueueLog.h"
QueueLog::QueueLog()
{
}
QueueLog::~QueueLog()
{
}
QueueLog& QueueLog::instance()
{
static QueueLog quelog;
return quelog;
}
bool QueueLog::SetLog(string logmsg, int type)
{
std::lock_guard<mutex>lock(_mutex);
//不做判空
memset(&msg,0,sizeof(msg));
msg.type = type;
memcpy(&msg.info, logmsg.c_str(), logmsg.length());
que.push(msg);
return true;
}
MSGLOG QueueLog::GetLog()
{
std::lock_guard<mutex>lock(_mutex);
MSGLOG getMsg;
memset(&getMsg,0,sizeof(getMsg));
if (que.empty())
{
goto EXIT;
}
//while (que.empty())
//{
// Sleep(2);
//}
//Sleep(3);
getMsg = que.front();
que.pop();
EXIT:
return getMsg;
}
#pragma once
#include <iostream>
#include <string>
#include <log4cplus/logger.h>
#include <log4cplus/configurator.h>
#include <log4cplus/layout.h>
#include <log4cplus/loggingmacros.h>
#include <log4cplus/helpers/stringhelper.h>
#define MY_LOG_FILE_PATH "../log/logconfig.properites"
using namespace std;
using namespace log4cplus;
using namespace log4cplus::helpers;
class MyLogger
{
public:
static MyLogger & getInstance();
Logger logger;
private:
MyLogger();
~MyLogger();
};
#define LOG4CPLUS MyLogger::getInstance()
#include "logger.h"
MyLogger::MyLogger()
{
log4cplus::initialize();
PropertyConfigurator::doConfigure(LOG4CPLUS_TEXT(MY_LOG_FILE_PATH));
logger = Logger::getRoot();
}
MyLogger & MyLogger::getInstance()
{
static MyLogger log;
return log;
}
MyLogger::~MyLogger()
{
}
#include "QueueLog.h"
#include "test.h"
#include "logger.h"
#include <thread>
using namespace std;
void SetTest2()
{
while (1)
{
MSGLOG msg;
memset(&msg, 0, sizeof(msg));
LOG.SetLog("test", 0);
LOG.SetLog("test2", 0);
LOG.SetLog("test3", 0);
LOG.SetLog("test4", 0);
Sleep(5);
}
}
int main()
{
std::thread t1(SetTest2);
MSGLOG msg;
/*memset(&msg,0,sizeof(msg));
LOG.SetLog("test",0);*/
SetTest();
while (1)
{
memset(&msg, 0, sizeof(msg));
msg = LOG.GetLog();
switch(msg.type)
{
case 0:
LOG4CPLUS_DEBUG(LOG4CPLUS.logger, msg.info);
break;
case 1:
LOG4CPLUS_DEBUG(LOG4CPLUS.logger, msg.info);
break;
}
cout << "id:" << msg.type << endl;
cout << "msg:" << msg.info << endl;
//Sleep(1);
}
t1.join();
system("pause");
return 0;
}
to sum up:
This is a simple asynchronous logging implementation, project process is relatively simple, this can also meet the requirements; in fact, there is now a large number of message queues open source libraries, performance is relatively high, but to design their own, troubleshooting is a process of growing;
To learn more learn C ++ back-end server, please attention:
micro-channel public number:CPP backend server development
");
return 0;
}
**************
********
**总结:**
这就是一个简单的异步日志实现,项目流程比较简单,这个也可以符合要求;其实现在存在大量的消息队列的开源库,性能比较高,不过自己去设计、排错也是一个成长的过程;
************
想了解学习更多C++后台服务器方面的知识,请关注:
微信公众号:====**CPP后台服务器开发**====
**************