開源C擴展輕量級PHP數據庫ORM框架ycdatabase : 構建穩定的PHP數據庫連接池

XFLY 發布于2019-06-25 19:09 / 1281人閱讀

摘要：安全能通過參數綁定的方式解決注入的問題。強大便捷的函數，支持所有數據庫操作。中文文檔男性女性身高可選可選

ycdatabase

github: https://github.com/caohao-php...

Catalogue

Instruction

Requirement

Create test table

Compire ycdatabase in linux

Start ycdatabase

Init ycdb connection

Native SQL query

Error Info

Where statement

Select statement

Insert statement

Replace statement

Update statement

Delete statement

Whole Example

Database Transaction

Data Caching

PHP Database Connection Pool

Redis Connection Pool

Instruction

1、Fast : ycdb is an mysql database ORM written in c, built in php extension, as we known, database ORM is a very time-consuming operation, especially for interpretive languages such as PHP, and for a project, the proportion of ORM is very high,so here I will implement the MySQL ORM operation in C language, and use the performance of C language to improve the performance of ORM.

2、Safe : ycdb can solve SQL injection through parameter binding.

3、Powerful : concise and powerful usage , support any operation in database.

4、Easy : Extremely easy to learn and use, friendly construction.

5、Data-cache : ycdb supports data caching. You can use redis as a medium to cache database data, but remember that when the update, insert, and delete operations involve caching data, you need to delete your cache to ensure data consistency.

6、Connection-pool : ycdb uses a special way to establish a stable connection pool with MySQL. performance can be increased by at least 30%, According to PHP"s operating mechanism, long connections can only reside on top of the worker process after establishment, that is, how many work processes are there. How many long connections, for example, we have 10 PHP servers, each launching 1000 PHP-FPM worker processes, they connect to the same MySQL instance, then there will be a maximum of 10,000 long connections on this MySQL instance, the number is completely Out of control! And PHP"s connection pool heartbeat mechanism is not perfect

1、快速 - ycdb是一個為PHP擴展寫的純C語言寫的mysql數據庫ORM擴展，眾所周知，數據庫ORM是一個非常耗時的操作，尤其對于解釋性語言如PHP，而且對于一個項目來說，ORM大多數情況能占到項目很大的一個比例，所以這里我將MySQL的ORM操作用C語言實現，利用C語言的性能，提升ORM的性能。

2、安全 - ycdb能通過參數綁定的方式解決SQL注入的問題。

3、強大 - 便捷的函數，支持所有數據庫操作。

4、簡單 - 使用和學習非常簡單，界面友好。

5、數據緩存 - ycdb支持數據緩存，你可以采用redis作為介質來緩存數據庫的數據，但是記得在update、insert、delete 操作涉及到與緩存數據相關的數據修改時，需要按key刪除您的緩存，以保證數據一致性。

6、連接池 - ycdb通過一種特殊的方式來建立一個穩定的與MySQL之間的連接池，性能至少能提升30%，按照 PHP 的運行機制，長連接在建立之后只能寄居在工作進程之上，也就是說有多少個工作進程，就有多少個長連接，打個比方，我們有 10 臺 PHP 服務器，每臺啟動 1000 個 PHP-FPM 工作進程，它們連接同一個 MySQL 實例，那么此 MySQL 實例上最多將存在 10000 個長連接，數量完全失控了！而且PHP的連接池心跳機制不完善。

中文文檔(Chinese Document)： https://blog.csdn.net/caohao0... Requirement

PHP 7.0 +

need support PDO for mysql

Create test table

CREATE TABLE `user_info_test` (
  `uid` int(11) NOT NULL COMMENT "userid" AUTO_INCREMENT,
  `username` varchar(64) NOT NULL COMMENT "username",
  `sexuality` varchar(8) DEFAULT "male" COMMENT "sexuality：male - 男性  female - 女性",
  `age` int(11) DEFAULT 0 COMMENT "age",
  `height` double(11,2) DEFAULT 0 COMMENT "height of a person, 身高",
  `bool_flag` int(11) DEFAULT 1 COMMENT "flag",
  `remark` varchar(11) DEFAULT NULL,
  PRIMARY KEY (`uid`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_bin COMMENT="userinfo";

Compire ycdatabase in linux

////  path to is your PHP install dir ////
$cd ~/ycdatabase/ycdatabase_extension
$/path/to/phpize
$chmod +x ./configure
$./configure --with-php-config=/path/to/php-config
$make
$make install

Start ycdatabase

new ycdb()

$db_conf = array("host" => "127.0.0.1", 
                 "username" => "root", 
                 "password" => "test123123", 
                 "dbname" => "userinfo", 
                 "port" => "3306", 
                 "option" => array(
                        PDO::ATTR_CASE => PDO::CASE_NATURAL,
                        PDO::ATTR_TIMEOUT => 2));

$ycdb = new ycdb($db_conf);

we can start by creating a ycdatabase object (ycdb) from the obove code, db_conf include host,username,password,dbname,port and option, option is a pdo attribution, you can get the detail from http://php.net/manual/en/pdo.... For example, PDO::ATTR_TIMEOUT in the above code is specifies the timeout duration in seconds, and PDO::ATTR_CASE is forcing column names to a specific case.

Init ycdb connection

we need to init pdo connection before we use ycdatabase.

try{
    $ycdb->initialize();
} catch (PDOException $e) {
    echo "find PDOException when initialize
";
    var_dump($e);
    exit;
}

Native SQL query

We can directly execute the sql statement through the exec() function,the return value is the number of rows affected by the execution, or return insert_id if it is insert statement, when the table has not AUTO_INCREMENT field, the insert_id should be zero, and execute select statement through the query() function, If $ret = -1 indicates that the sql execution error occurs, we can pass $ycdb->errorCode(), $ycdb- >errorInfo() returns the error code and error description respectively.

insert data

$insert_id = $ycdb->exec("insert into user_info_test(username, sexuality, age, height) 
                    values("smallhow", "male", 29, 180)");
if($insert_id == -1) {
    $code = $ycdb->errorCode();
    $info = $ycdb->errorInfo();
    echo "code:" . $code . "
";
    echo "info:" . $info[2] . "
";
} else {
    echo $insert_id;
}

update data

if we execute the following update statement, $ret returns 3 if the current data is the above image.

$ret = $ycdb->exec("update user_info_test set remark="test" where height>=180");
echo $ret; //ret is 3

select data

$ret = $ycdb->query("select * from user_info_test where bool_flag=1");
echo json_encode($ret);

 $ret = $ycdb->query("select username from user_info_test where bool_flag=1");
echo json_encode($ret);

Error Info

Error codes and error messages can be obtained through the errorCode and errorInfo function

$code = $ycdb->errorCode();
$info = $ycdb->errorInfo();

Where statement

Basic usage

$ycdb->select("user_info_test", "*", ["sexuality" => "male"]);
// WHERE sexuality = "male"

$ycdb->select("user_info_test", "*", ["age" => 29]);  // WHERE age = 29

$ycdb->select("user_info_test", "*", ["age[>]" => 29]); // WHERE age > 29

$ycdb->select("user_info_test", "*", ["age[>=]" => 29]); // WHERE age >= 29

$ycdb->select("user_info_test", "*", ["age[!]" => 29]); // WHERE age != 29

$ycdb->select("user_info_test", "*", ["age[<>]" => [28, 29]]); // WHERE age  BETWEEN 28 AND 29

$ycdb->select("user_info_test", "*", ["age[><]" => [28, 29]]); // WHERE age NOT BETWEEN 28 AND 29

$ycdb->select("user_info_test", "*", ["username" => ["Tom", "Red", "carlo"]]); // WHERE username in ("Tom", "Red", "carlo")

//Multiple conditional query
$data = $ycdb->select("user_info_test", "*", [
    "uid[!]" => 10,
    "username[!]" => "James",
    "height[!]" => [165, 168, 172],
    "bool_flag" => true,
    "remark[!]" => null
]);
// WHERE uid != 10 AND username != "James" AND height NOT IN ( 165, 168, 172) AND bool_flag = 1 AND remark IS NOT NULL

Conditional Query

You can use "AND" or "OR" to make up very complex SQL statements.

$data = $ycdb->select("user_info_test", "*", [
  "OR" => [
    "uid[>]" => 3,
    "age[<>]" => [28, 29],
    "sexuality" => "female"
  ]
]);
// WHERE uid > 3 OR age BETWEEN 29 AND 29 OR sexuality = "female"

$data = $ycdb->select("user_info_test", "*", [
  "AND" => [
    "OR" => [
      "age" => 29,
      "sexuality" => "female"
    ],
    "height" => 177
  ]
]);
// WHERE (age = 29 OR sexuality="female") AND height = 177

//Attention： Because ycdb uses array arguments, the first OR is overwritten, the following usage is wrong, 
$data = $ycdb->select("user_info_test", "*", [
  "AND" => [
    "OR" => [
      "age" => 29,
      "sexuality" => "female"
    ],
    "OR" => [
      "uid[!]" => 3,
      "height[>=]" => 170
    ],
  ]
]);
// [X] SELECT * FROM user_info_test WHERE (uid != 3 OR height >= 170)

//We can use # and comments to distinguish between two diffrents OR
$data = $ycdb->select("user_info_test", "*", [
  "AND" => [
    "OR #1" => [
      "age" => 29,
      "sexuality" => "female"
    ],
    "OR #2" => [
      "uid[!]" => 3,
      "height[>=]" => 170
    ],
  ]
]);
// [√] SELECT * FROM user_info_test WHERE (age = 29 OR sexuality = "female") AND (uid != 3 OR height >= 170)

Fuzzy Matching Like

LIKE USAGE [~].

$data = $ycdb->select("user_info_test", "*", [ "username[~]" => "%ide%" ]);
// WHERE username LIKE "%ide%"

$data = $ycdb->select("user_info_test", "*", ["username[~]" => ["%ide%", "Jam%", "%ace"]]);
// WHERE username LIKE "%ide%" OR username LIKE "Jam%" OR username LIKE "%ace"

$data = $ycdb->select("user_info_test", "*", [ "username[!~]" => "%ide%" ]);
// WHERE username NOT LIKE "%ide%"

Use of wildcards

$ycdb->select("user_info_test", "*", [ "username[~]" => "Londo_" ]); // London, Londox, Londos...

$ycdb->select("user_info_test", "id", [ "username[~]" => "[BCR]at" ]); // Bat, Cat, Rat

$ycdb->select("user_info_test", "id", [    "username[~]" => "[!BCR]at" ]); // Eat, Fat, Hat...

ORDER BY And LIMIT

$data = $ycdb->select("user_info_test", "*", [
  "sexuality" => "male",
  "ORDER" => [
    "age",
    "height" => "DESC",
    "uid" => "ASC"
  ],
  "LIMIT" => 100,  //Get the first 100 of rows (overwritten by next LIMIT)
  "LIMIT" => [20, 100]  //Started from the top 20 rows, and get the next 100
]);
//SELECT * FROM `user_info_test` WHERE `sexuality` = "male" ORDER BY `age`, `height` DESC, `uid` ASC LIMIT 100 OFFSET 20

GROUP And HAVING

$ycdb->select("user_info_test", "sexuality,age,height", [
  "GROUP" => "sexuality",
 
  // GROUP by array of values
  "GROUP" => [
    "sexuality",
    "age",
    "height"
  ],
 
  // Must have to use it with GROUP together
  "HAVING" => [
    "age[>]" => 30
  ]
]);
//SELECT uid FROM `user_info_test` GROUP BY sexuality,age,height HAVING `age` > 30

Select statement

usage

select($table, $columns, $where)

table [string]

table name

columns [string/array]

Columns to be queried.

where (optional) [array]

The conditions of the query.

select($table, $join, $columns, $where)

table [string]

table name

join [array]

Multi-table query, can be ignored if not used.

columns [string/array]

Columns to be queried.

where (optional) [array]

The conditions of the query.

return: [array]

Fail if -1 is returned, otherwise result array is returned

example

You can use * to match all fields, but if you specify columns you can improve performance.

$datas = $ycdb->select("user_info_test", [
  "uid",
  "username"
], [
  "age[>]" => 31
]);

// $datas = array(
//  [0] => array(
//      "uid" => 6,
//      "username" => "Aiden"
//  ),
//  [1] => array(
//      "uid" => 11,
//      "username" => "smallhow"
//  )
// )

// Select all columns
$datas = $ycdb->select("user_info_test", "*");

// Select a column
$datas = $ycdb->select("user_info_test", "username");
 
// $datas = array(
//  [0] => "lucky",
//  [1] => "Tom",
//  [2] => "Red"
// )

Table join

Multi-table query SQL is more complicated, and it can be easily solved with ycdb.

// [>] == RIGH JOIN
// [<] == LEFT JOIN
// [<>] == FULL JOIN
// [><] == INNER JOIN

$ycdb->select("user_info_test",
[ // Table Join Info
  "[>]account" => ["uid" => "userid"], // RIGHT JOIN `account` ON `user_info_test`.`uid`= `account`.`userid`
 
  // This is a shortcut to declare the relativity if the row name are the same in both table.
  "[>]album" => "uid", //RIGHT JOIN `album` USING (`uid`) 
  
  // Like above, there are two row or more are the same in both table.
  "[<]detail" => ["uid", "age"], // LEFT JOIN `detail` USING (`uid`,`age`)
 
  // You have to assign the table with alias.
  "[<]address(addr_alias)" => ["uid" => "userid"], //LEFT JOIN `address` AS `addr_alias` ON `user_info_test`.`uid`=`addr_alias`.`userid`
 
  // You can refer the previous joined table by adding the table name before the column.
  "[<>]album" => ["account.userid" => "userid"], //FULL JOIN `album` ON  `account`.`userid` = `album`.`userid`
 
  // Multiple condition
  "[><]account" => [
    "uid" => "userid",
    "album.userid" => "userid"
  ]
], [ // columns
  "user_info_test.uid",
  "user_info_test.age",
  "addr_alias.country",
  "addr_alias.city"
], [ // where condition
  "user_info_test.uid[>]" => 3,
  "ORDER" => ["user_info_test.uid" => "DESC"],
  "LIMIT" => 50
]);


// SELECT 
//   user_info_test.uid,
//   user_info_test.age,
//   addr_alias.country,
//   addr_alias.city 
// FROM `user_info_test` 
// RIGHT JOIN `account` ON `user_info_test`.`uid`= `account`.`userid`  
// RIGHT JOIN `album` USING (`uid`) 
// LEFT JOIN `detail` USING (`uid`,`age`) 
// LEFT JOIN `address` AS `addr_alias` ON `user_info_test`.`uid`=`addr_alias`.`userid` 
// FULL JOIN `album` ON  `account`.`userid` = `album`.`userid` 
// INNER JOIN `account` ON `user_info_test`.`uid`= `account`.`userid` 
//   AND `album`.`userid` = `account`.`userid`  
// WHERE `user_info_test`.`uid` > 3 
// ORDER BY  `user_info_test`.`uid` DESC 
// LIMIT 50

alias

You can use aliases to prevent field conflicts

$data = $ycdb->select("user_info_test(uinfo)", [
  "[<]account(A)" => "userid",
], [
  "uinfo.uid(uid)",
  "A.userid"
]);

// SELECT uinfo.uid AS `uid`, A.userid 
// FROM `user_info_test` AS `uinfo` 
// LEFT JOIN `account` AS `A` USING (`userid`)

Insert statement

insert($table, $data, $cache_info)

table [string]

table name

data [array]

insert data

cache_info (optional) [array]

cache info

return [int]

Fail if -1 is returned, otherwise insert_id is returned, if the table has no AUTO_INCREMENT field, the insert_id is zero

$data = array("username" => "smallhow","sexuality" => "male","age" => 35, "height" => "168");
$insert_id = $ycdb->insert("user_info_test", $data);
if($insert_id == -1) {
  $code = $ycdb->errorCode();
  $info = $ycdb->errorInfo();
  echo "code:" . $code . "
";
  echo "info:" . $info[2] . "
";
} else {
  echo $insert_id;
}

Replace statement

replace($table, $data, $cache_info)

table [string]

table name

data [array]

replace data

cache_info (optional) [array]

cache info

return [int]

Fail if -1 is returned, otherwise insert_id is returned

$data = array("username" => "smallhow","sexuality" => "male","age" => 35, "height" => "168");
$insert_id = $ycdb->replace("user_info_test", $data);
if($insert_id == -1) {
  $code = $ycdb->errorCode();
  $info = $ycdb->errorInfo();
  echo "code:" . $code . "
";
  echo "info:" . $info[2] . "
";
} else {
  echo $insert_id;
}

Update statement

update($table, $data, $where)

table [string]

table name

data [array]

update data

where (optional) [array]

where condition [可選]

return [int]

Fail if -1 is returned, otherwise the number of update records is returned

$data = array("height" => 182,"age" => 33);
$where = array("username" => "smallhow");
$ret = $ycdb->update("user_info_test", $data, $where);

Delete statement

delete($table, $where)

table [string]

table name

where (optional) [array]

where condition [可選]

return [int]

Fail if -1 is returned, otherwise the number of delete records is returned

$where = array("username" => "smallhow");
$ret = $ycdb->delete("user_info_test", $where);

Whole Example

$table = "table_a(a)";

$join = [
  "[>]AAAA(a1)" => "id",
  "[<]BBBB" => ["E1", "E2", "E3"],
  "[>]CCCC(c1)" => [ "GG" => "HH", "II.KK" => "LL"]
];

$columns = ["name(a)", "avatar(b)", "age"];

$where =  [
  "user.email[!]" => ["foo@bar.com", "cat@dog.com", "admin@ycdb.in"],
  "user.uid[<]" => 11111,
  "uid[>=]" => 222,
  "uid[!]" => null,
  "count[!]" => [36, 57, 89],
  "id[!]" => true,
  "int_num[!]" => 3,
  "double_num[!]" => 3.76,
  "AA[~]" => "%saa%",
  "BB[!~]" => "%sbb",
  "CC[~]" => ["11%", "22_", "33%"],
  "DD[!~]" => ["%44%", "55%", "66%"],
  "EE[~]" => ["AND" => ["%E11", "E22"]],
  "FF[~]" => ["OR" => ["%F33", "F44"]],
  "GG[!~]" => ["AND" => ["%G55", "G66"]],
  "HH[!~]" => ["OR" => ["H77", "H88"]],
  "II[<>]" => ["1", "12"],
  "LL[><]" => ["1", "12"],
    "AND #1" => [
        "OR #1" => [
          "user_name" => null,
          "email" => "foo@bar.com",
        ],
        "OR #2" => [
          "user_name" => "bar",
          "email" => "bar@foo.com"
        ]
    ],
    "OR" => [
        "user_name[!]" => "foo",
        "promoted[!]" => true
    ],
    "GROUP" => "userid",
    "GROUP" => ["type", "age", "gender"],
    "HAVING" => [
        "uid.num[>]" => 111,
        "type[>]" => "smart",
        "id[!]" => false,
        "god3[!]" => 9.86,
        "uid[!]" => null,
        "AA[~]" => "SSA%",
        "CC[~]" => ["11%", "22%", "%33"],
    ],
    "ORDER" => [
        "user.score",
        "user.uid" => "ASC",
        "time" => "DESC",
    ],
    "LIMIT" => 33,
];

$ycdb->select($table, $join, $columns, $where);

Database transaction

$ycdb->begin();

$ret1 = $ycdb->exec("insert into user_info_test(username, sexuality, age, height) values("smallhow", "male", 29, 180)");
$ret2 = $ycdb->exec("insert into user_info_test(username, sexuality, age, height) values("jesson", "female", 28, 175)");

if($ret1 == -1 || $ret2 == -1 ) {
  $ycdb->rollback();
} else {
  $ycdb->commit()
}

Data Caching

We can use redis, or any other cache system that supports set/get/del/expire function as the medium to store the data returned by the database. If you do not specify the expiration time, the default storage expiration time is 5 minutes. if The cache is specified. When we call data update function such as update/delete/insert, we should pass in the same cache key so that ycdb can clear the cache to ensure data consistency.

//we want cache data by redis
$redis = new Redis();
$redis->connect("/home/redis/pid/redis.sock");

$option = array("host" => "127.0.0.1", 
                "username" => "test", 
                "password" => "test", 
                "dbname" => "test", 
                "port" => "3306",
                "cache" => $redis,  //cache instance
                "option" => array(
                    PDO::ATTR_CASE => PDO::CASE_NATURAL,
                    PDO::ATTR_TIMEOUT => 2));


$ycdb = new ycdb($option);

try{
  $ycdb->initialize();
} catch (PDOException $e) {
  echo "find PDOException when initialize
";
  exit;
}

// I want to keep the 29-year-old user data queried by the database in the cache, and keep it for 10 minutes.
$age = 29;
$cache_key = "pre_cache_key_" . $age;

$data = $ycdb->select("user_info_test", "*", [
    "age" => $age,
    "CACHE" => ["key" => $cache_key, "expire" => 600]   //cache key an expire time (seconds)
]);

echo $redis->get($cache_key) . "
";

// If I update these 29-year-old user data, or even add a new 29-year-old user information, 
// it"s best to enter the cache key to clean up the cache to keep the data consistent.
$ycdb->update("user_info_test", ["remark" => "29-year-old"], [
  "age" => $age,
  "CACHE" => ["key" => $cache_key]  //cache key
]);

echo $redis->get($cache_key) . "
";

//If you are going to delete the relevant data, it is best to also clean up the cache by cache_key.
$ycdb->delete("user_info_test", [
  "age" => $age,
  "CACHE" => ["key" => $cache_key]  //cache key
]);

echo $redis->get($cache_key) . "
";

//Clean up the cache by cache_key when the data you insert is related to the cached data.
$insert_data = array();
$insert_data["username"] = "test";
$insert_data["sexuality"] = "male";
$insert_data["age"] = 29;
$insert_data["height"] = 176;

$insert_id = $ycdb->insert("user_info_test", $insert_data, ["key" => $cache_key]);

echo $redis->get($cache_key) . "
";

PHP Database Connection Pool

Short connection performance is generally not available. CPU resources are consumed by the system. Once the network is jittered, there will be a large number of TIME_WAIT generated. The service has to be restarted periodically or the machine is restarted periodically. The server is unstable, QPS is high and low, and the connection is stable and efficient. The pool can effectively solve the above problems, it is the basis of high concurrency. ycdb uses a special way to establish a stable connection pool with MySQL. performance can be increased by at least 30%, According to PHP"s operating mechanism, long connections can only reside on top of the worker process after establishment, that is, how many work processes are there. How many long connections, for example, we have 10 PHP servers, each launching 1000 PHP-FPM worker processes, they connect to the same MySQL instance, then there will be a maximum of 10,000 long connections on this MySQL instance, the number is completely Out of control! And PHP"s connection pool heartbeat mechanism is not perfect

How ?

Let"s focus on Nginx, its stream module implements load balancing of TCP/UDP services, and with the stream-lua module, we can implement programmable stream services, that is, custom TCP/N with Nginx. UDP service! Of course, you can write TCP/UDP services from scratch, but standing on Nginx"s shoulder is a more time-saving and labor-saving choice. We can choose the OpenResty library to complete the MySQL connection pool function. OpenResty is a very powerful and well-functioning Nginx Lua framework. It encapsulates Socket, MySQL, Redis, Memcache, etc. But what is the relationship between Nginx and PHP connection pool? And listen to me slowly: Usually most PHP is used with Nginx, and PHP and Nginx are mostly on the same server. With this objective condition, we can use Nginx to implement a connection pool, connect to services such as MySQL on Nginx, and then connect to Nginx through a local Unix Domain Socket, thus avoiding all kinds of short links. Disadvantages, but also enjoy the benefits of the connection pool.

OpenResty Install

OpenResty Document: https://moonbingbing.gitbooks...
OpenResty Official Website : http://www.openresty.org/

CentOS 6.8 Install :

###### Install the necessary libraries ######
$yum install readline-devel pcre-devel openssl-devel perl

###### Install OpenResty ######
$cd ~/ycdatabase/openresty
$tar -xzvf openresty-1.13.6.1.tar.gz
$cd openresty-1.13.6.1
$./configure --prefix=/usr/local/openresty.1.13 --with-luajit --without-http_redis2_module --with-http_iconv_module
$gmake 
$gmake install

###### open mysql pool ######
$cp -rf ~/ycdatabase/openresty/openresty-pool ~/
$mkdir ~/openresty-pool/logs
$/usr/local/openresty.1.13/nginx/sbin/nginx -p ~/openresty-pool

MySQL Database Connection Pool Config

~/openresty-pool/conf/nginx.conf :

worker_processes  1;        #nginx worker process num

error_log logs/error.log;   #nginx error log path

events {
    worker_connections 1024;
}

stream {
  lua_code_cache on;

  lua_check_client_abort on;
  
  server {
    listen unix:/tmp/mysql_pool.sock;
        
    content_by_lua_block {
      local mysql_pool = require "mysql_pool"
            
      local config = {host = "127.0.0.1", 
                      user = "root", 
                      password = "test", 
                      database = "collect", 
                      timeout = 2000, 
                      max_idle_timeout = 10000, 
                      pool_size = 200}
                           
      pool = mysql_pool:new(config)
            
      pool:run()
    }
  }
}

If you have more than a MySQL Server, you can start another server and add a new listener to unix domain socket.

PHP Code

Except the option is array("unix_socket" => "/tmp/mysql_pool.sock") , Php mysql connection pool usage is exactly the same as before,But, MySQL does not support transactions in unix domain socket mode.

$option = array("unix_socket" => "/tmp/mysql_pool.sock");
$ycdb = new ycdb($option);
$ret = $ycdb->select("user_info_test", "*", ["sexuality" => "male"]);

if($ret == -1) {
  $code = $ycdb->errorCode();
  $info = $ycdb->errorInfo();
  echo "code:" . $code . "
";
  echo "info:" . $info[2] . "
";
} else {
  print_r($ret);
}

Redis Connection Pool

Similarly, Redis can solve the connection pool problem in the same way.

Redis Connection Pool Config

~/openresty-pool/conf/nginx.conf

worker_processes  1;        #nginx worker process num
 
error_log logs/error.log;   #error log path
 
events {
    worker_connections 1024;
}
 
stream {
  lua_code_cache on;
 
  lua_check_client_abort on;
 
  server {
    listen unix:/tmp/redis_pool.sock;
 
    content_by_lua_block {
      local redis_pool = require "redis_pool"
            
      pool = redis_pool:new({ip = "127.0.0.1", port = 6379, auth = "password"})
 
      pool:run()
    }
  }
    
  server {
 
    listen unix:/tmp/mysql_pool.sock;
        
    content_by_lua_block {
      local mysql_pool = require "mysql_pool"
            
      local config = {host = "127.0.0.1", 
                      user = "root", 
                      password = "test", 
                      database = "collect", 
                      timeout = 2000, 
                      max_idle_timeout = 10000, 
                      pool_size = 200}
                           
      pool = mysql_pool:new(config)
            
      pool:run()
    }
  }
}

PHP Code

$redis = new Redis();
$redis->pconnect("/tmp/redis_pool.sock");
var_dump($redis->hSet("foo1", "vvvvv42", 2));
var_dump($redis->hSet("foo1", "vvvv", 33));
var_dump($redis->expire("foo1", 111));
var_dump($redis->hGetAll("foo1"));

GPU云服務器云服務器 php數據庫連接池 php數據源連接池 php的重量級框架開源ORM框架

文章版權歸作者所有，未經允許請勿轉載,若此文章存在違規行為，您可以聯系管理員刪除。

轉載請注明本文地址：http://specialneedsforspecialkids.com/yun/17952.html

開源C擴展輕量級PHP數據庫ORM框架ycdatabase : 構建穩定的PHP數據庫連接池

摘要：安全能通過參數綁定的方式解決注入的問題。強大便捷的函數，支持所有數據庫操作。中文文檔男性女性身高可選可選 ycdatabase github: https://github.com/caohao-php... Catalogue Instruction Requirement Create test table Compire ycdatabase in linux Start yc...

Sleepy 2019-07-01 12:33 評論0 收藏0
輕量級高性能PHP框架ycroute

摘要：數據交互層可選如果你習慣了層結構，你可以加載層，作為與數據庫交互的層，而層作為業務層。 YCRoute github: https://github.com/caohao-php... 目錄框架介紹運行環境代碼結構路由配置過濾驗簽控制層加載器模型層數據交互dao層(可選) Redis緩存操作數據庫操作配置加載公共類加載公共函數日志模塊視圖層 RPC 介紹...

dailybird 2019-07-01 12:34 評論0 收藏0
IMI 基于 Swoole 開發的協程 PHP 開發框架常駐內存、協程異步非阻塞

摘要：介紹是基于開發的協程開發框架，擁有常駐內存協程異步非阻塞等優點。宇潤我在年開發并發布了第一個框架，一直維護使用至今，非常穩定，并且有文檔。于是我走上了開發的不歸路 showImg(https://segmentfault.com/img/bVbcxQH?w=340&h=160); 介紹 IMI 是基于 Swoole 開發的協程 PHP 開發框架，擁有常駐內存、協程異步非阻塞IO等優點。...

airborne007 2019-07-01 10:10 評論0 收藏0
php 學習指南及技術干貨

摘要：安全生成安全的隨機數，加密數據，掃描漏洞的庫一個兼容標準的過濾器一個生成隨機數和字符串的庫使用生成隨機數的庫一個安全庫一個純安全通信庫一個簡單的鍵值加密存儲庫一個結構化的安全層一個試驗的面向對象的包裝庫一個掃描文件安全的庫 Security 安全生成安全的隨機數，加密數據，掃描漏洞的庫 HTML Purifier-一個兼容標準的HTML過濾器 RandomLib-一個生成隨機數和字...

lifefriend_007 2019-06-27 14:12 評論0 收藏0