連接失效問題

例子

其中，Redis常見的報(bào)錯就是：

配置項(xiàng)：timeout

報(bào)錯信息：

Error while reading line from the server

Redis可以配置如果客戶端經(jīng)過多少秒還不給Redis服務(wù)器發(fā)送數(shù)據(jù)，那么就會把連接close掉。

MySQL常見的報(bào)錯：

配置項(xiàng)：wait_timeout interactive_timeout

報(bào)錯信息：

has gone away

和Redis服務(wù)器一樣，MySQL也會定時(shí)的去清理掉沒用的連接。

如何解決

1、用的時(shí)候進(jìn)行重連 。優(yōu)點(diǎn)是簡單，缺點(diǎn)是面臨短連接的問題。

2、定時(shí)發(fā)送心跳維持連接（推薦）。

如何維持長連接

tcp協(xié)議中實(shí)現(xiàn)的tcp_keepalive

操作系統(tǒng)底層提供了一組tcp的keepalive配置：

tcp_keepalive_time (integer; default: 7200; since Linux 2.2)
The number of seconds a connection needs to be idle before TCP
begins sending out keep-alive probes. Keep-alives are sent only
when the SO_KEEPALIVE socket option is enabled. The default
value is 7200 seconds (2 hours). An idle connection is
terminated after approximately an additional 11 minutes (9
probes an interval of 75 seconds apart) when keep-alive is
enabled.

Note that underlying connection tracking mechanisms and
application timeouts may be much shorter.

tcp_keepalive_intvl (integer; default: 75; since Linux 2.4)
The number of seconds between TCP keep-alive probes.

tcp_keepalive_probes (integer; default: 9; since Linux 2.2)
The maximum number of TCP keep-alive probes to send before
giving up and killing the connection if no response is obtained
from the other end.
8

Swoole底層把這些配置開放出來了，例如：

?php

$server = new \Swoole\Server('127.0.0.1', 6666, SWOOLE_PROCESS);

$server->set([
'worker_num' => 1,
'open_tcp_keepalive' => 1,
'tcp_keepidle' => 4, // 對應(yīng)tcp_keepalive_time
'tcp_keepinterval' => 1, // 對應(yīng)tcp_keepalive_intvl
'tcp_keepcount' => 5, // 對應(yīng)tcp_keepalive_probes
]);

其中：

'open_tcp_keepalive' => 1, // 總開關(guān)，用來開啟tcp_keepalive
'tcp_keepidle' => 4, // 4s沒有數(shù)據(jù)傳輸就進(jìn)行檢測
// 檢測的策略如下：
'tcp_keepinterval' => 1, // 1s探測一次，即每隔1s給客戶端發(fā)一個包（然后客戶端可能會回一個ack的包，如果服務(wù)端收到了這個ack包，那么說明這個連接是活著的）
'tcp_keepcount' => 5, // 探測的次數(shù)，超過5次后客戶端還沒有回ack包，那么close此連接

我們來實(shí)戰(zhàn)測試體驗(yàn)一下，服務(wù)端腳本如下：

?php

$server = new \Swoole\Server('127.0.0.1', 6666, SWOOLE_PROCESS);

$server->set([
'worker_num' => 1,
'open_tcp_keepalive' => 1, // 開啟tcp_keepalive
'tcp_keepidle' => 4, // 4s沒有數(shù)據(jù)傳輸就進(jìn)行檢測
'tcp_keepinterval' => 1, // 1s探測一次
'tcp_keepcount' => 5, // 探測的次數(shù)，超過5次后還沒有回包c(diǎn)lose此連接
]);

$server->on('connect', function ($server, $fd) {
var_dump("Client: Connect $fd");
});

$server->on('receive', function ($server, $fd, $reactor_id, $data) {
var_dump($data);
});

$server->on('close', function ($server, $fd) {
var_dump("close fd $fd");
});

$server->start();

我們啟動這個服務(wù)器：

~/codeDir/phpCode/hyperf-skeleton # php server.php

然后通過tcpdump進(jìn)行抓包：

~/codeDir/phpCode/hyperf-skeleton # tcpdump -i lo port 6666
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on lo, link-type EN10MB (Ethernet), capture size 262144 bytes

我們此時(shí)正在監(jiān)聽lo上的6666端口的數(shù)據(jù)包。

然后我們用客戶端去連接它：

~/codeDir/phpCode/hyperf-skeleton # nc 127.0.0.1 6666

此時(shí)服務(wù)端會打印出消息：

~/codeDir/phpCode/hyperf-skeleton # php server.php

string(17) "Client: Connect 1"

tcpdump的輸出信息如下：

01:48:40.178439 IP localhost.33933 > localhost.6666: Flags [S], seq 43162537, win 43690, options [mss 65495,sackOK,TS val 9833698 ecr 0,nop,wscale 7], length 0

01:48:40.178484 IP localhost.6666 > localhost.33933: Flags [S.], seq 1327460565, ack 43162538, win 43690, options [mss 65495,sackOK,TS val 9833698 ecr 9833698,nop,wscale 7], length 0

01:48:40.178519 IP localhost.33933 > localhost.6666: Flags [.], ack 1, win 342, options [nop,nop,TS val 9833698 ecr 9833698], length 0

01:48:44.229926 IP localhost.6666 > localhost.33933: Flags [.], ack 1, win 342, options [nop,nop,TS val 9834104 ecr 9833698], length 0

01:48:44.229951 IP localhost.33933 > localhost.6666: Flags [.], ack 1, win 342, options [nop,nop,TS val 9834104 ecr 9833698], length 0

01:48:44.229926 IP localhost.6666 > localhost.33933: Flags [.], ack 1, win 342, options [nop,nop,TS val 9834104 ecr 9833698], length 0

01:48:44.229951 IP localhost.33933 > localhost.6666: Flags [.], ack 1, win 342, options [nop,nop,TS val 9834104 ecr 9833698], length 0

01:48:44.229926 IP localhost.6666 > localhost.33933: Flags [.], ack 1, win 342, options [nop,nop,TS val 9834104 ecr 9833698], length 0

// 省略了其他的輸出

我們會發(fā)現(xiàn)最開始的時(shí)候，會打印三次握手的包：

01:48:40.178439 IP localhost.33933 > localhost.6666: Flags [S], seq 43162537, win 43690, options [mss 65495,sackOK,TS val 9833698 ecr 0,nop,wscale 7], length 0

01:48:40.178484 IP localhost.6666 > localhost.33933: Flags [S.], seq 1327460565, ack 43162538, win 43690, options [mss 65495,sackOK,TS val 9833698 ecr 9833698,nop,wscale 7], length 0

01:48:40.178519 IP localhost.33933 > localhost.6666: Flags [.], ack 1, win 342, options [nop,nop,TS val 9833698 ecr 9833698], length 0

然后，停留了4s沒有任何包的輸出。

之后，每隔1s左右就會打印出一組：

01:52:54.359341 IP localhost.6666 > localhost.43101: Flags [.], ack 1, win 342, options [nop,nop,TS val 9859144 ecr 9858736], length 0

01:52:54.359377 IP localhost.43101 > localhost.6666: Flags [.], ack 1, win 342, options [nop,nop,TS val 9859144 ecr 9855887], length 0

其實(shí)這就是我們配置的策略：

'tcp_keepinterval' => 1, // 1s探測一次
'tcp_keepcount' => 5, // 探測的次數(shù)，超過5次后還沒有回包c(diǎn)lose此連接

因?yàn)槲覀儾僮飨到y(tǒng)底層會自動的給客戶端回ack，所以這個連接不會在5次探測后被關(guān)閉。操作系統(tǒng)底層會持續(xù)不斷的發(fā)送這樣的一組包：

01:52:54.359341 IP localhost.6666 > localhost.43101: Flags [.], ack 1, win 342, options [nop,nop,TS val 9859144 ecr 9858736], length 0

01:52:54.359377 IP localhost.43101 > localhost.6666: Flags [.], ack 1, win 342, options [nop,nop,TS val 9859144 ecr 9855887], length 0

如果我們要測試5次探測后關(guān)閉這個連接，可以禁掉6666端口的包：

~/codeDir/phpCode/hyperf-skeleton # iptables -A INPUT -p tcp --dport 6666 -j DROP

這樣會把所有從6666端口進(jìn)來的包給禁掉，自然，服務(wù)器就接收不到從客戶端那一邊發(fā)來的ack包了。

然后服務(wù)器過5秒就會打印出close（服務(wù)端主動的調(diào)用了close方法，給客戶端發(fā)送了FIN包）：

~/codeDir/phpCode/hyperf-skeleton # php server.php

string(17) "Client: Connect 1"

string(10) "close fd 1"

我們恢復(fù)一下iptables的規(guī)則：

~/codeDir/phpCode # iptables -D INPUT -p tcp -m tcp --dport 6666 -j DROP

即把我們設(shè)置的規(guī)則給刪除了。

通過tcp_keepalive的方式實(shí)現(xiàn)心跳的功能，優(yōu)點(diǎn)是簡單，不要寫代碼就可以完成這個功能，并且發(fā)送的心跳包小。缺點(diǎn)是依賴于系統(tǒng)的網(wǎng)絡(luò)環(huán)境，必須保證服務(wù)器和客戶端都實(shí)現(xiàn)了這樣的功能，需要客戶端配合發(fā)心跳包。還有一個更為嚴(yán)重的缺點(diǎn)是如果客戶端和服務(wù)器不是直連的，而是通過代理來進(jìn)行連接的，例如socks5代理，它只會轉(zhuǎn)發(fā)應(yīng)用層的包，不會轉(zhuǎn)發(fā)更為底層的tcp探測包，那這個心跳功能就失效了。

所以，Swoole就提供了其他的解決方案，一組檢測死連接的配置。

'heartbeat_check_interval' => 1, // 1s探測一次
'heartbeat_idle_time' => 5, // 5s未發(fā)送數(shù)據(jù)包就close此連接

swoole實(shí)現(xiàn)的heartbeat

我們來測試一下：

?php

$server = new \Swoole\Server('127.0.0.1', 6666, SWOOLE_PROCESS);

$server->set([
'worker_num' => 1,
'heartbeat_check_interval' => 1, // 1s探測一次
'heartbeat_idle_time' => 5, // 5s未發(fā)送數(shù)據(jù)包就close此連接
]);

$server->on('connect', function ($server, $fd) {
var_dump("Client: Connect $fd");
});

$server->on('receive', function ($server, $fd, $reactor_id, $data) {
var_dump($data);
});

$server->on('close', function ($server, $fd) {
var_dump("close fd $fd");
});

$server->start();

然后啟動服務(wù)器：

~/codeDir/phpCode/hyperf-skeleton # php server.php

然后啟動tcpdump：

~/codeDir/phpCode # tcpdump -i lo port 6666

tcpdump: verbose output suppressed, use -v or -vv for full protocol decode

listening on lo, link-type EN10MB (Ethernet), capture size 262144 bytes

然后再啟動客戶端：

~/codeDir/phpCode/hyperf-skeleton # nc 127.0.0.1 6666

此時(shí)服務(wù)器端打?。?/p>

~/codeDir/phpCode/hyperf-skeleton # php server.php

string(17) "Client: Connect 1"

然后tcpdump打?。?/p>

02:48:32.516093 IP localhost.42123 > localhost.6666: Flags [S], seq 1088388248, win 43690, options [mss 65495,sackOK,TS val 10193342 ecr 0,nop,wscale 7], length 0

02:48:32.516133 IP localhost.6666 > localhost.42123: Flags [S.], seq 80508236, ack 1088388249, win 43690, options [mss 65495,sackOK,TS val 10193342 ecr 10193342,nop,wscale 7], length 0

02:48:32.516156 IP localhost.42123 > localhost.6666: Flags [.], ack 1, win 342, options [nop,nop,TS val 10193342 ecr 10193342], length 0

這是三次握手信息。

然后過了5s后，tcpdump會打印出：

02:48:36.985027 IP localhost.6666 > localhost.42123: Flags [F.], seq 1, ack 1, win 342, options [nop,nop,TS val 10193789 ecr 10193342], length 0

02:48:36.992172 IP localhost.42123 > localhost.6666: Flags [.], ack 2, win 342, options [nop,nop,TS val 10193790 ecr 10193789], length 0

也就是服務(wù)端發(fā)送了FIN包。因?yàn)榭蛻舳藳]有發(fā)送數(shù)據(jù)，所以Swoole關(guān)閉了連接。

然后服務(wù)器端會打?。?/p>

~/codeDir/phpCode/hyperf-skeleton # php server.php

string(17) "Client: Connect 1"

string(10) "close fd 1"

所以，heartbeat和tcp keepalive還是有一定的區(qū)別的，tcp keepalive有?；钸B接的功能，但是heartbeat存粹是檢測沒有數(shù)據(jù)的連接，然后關(guān)閉它，并且只可以在服務(wù)端這邊配置，如果需要保活，也可以讓客戶端配合發(fā)送心跳。

如果我們不想讓服務(wù)端close掉連接，那么就得在應(yīng)用層里面不斷的發(fā)送數(shù)據(jù)包來進(jìn)行?；睿缥以趎c客戶端里面不斷的發(fā)送包：

~/codeDir/phpCode/hyperf-skeleton # nc 127.0.0.1 6666

ping

ping

ping

ping

ping

ping

ping

ping

ping

我發(fā)送了9個ping包給服務(wù)器，tcpdump的輸出如下：

// 省略了三次握手的包

02:57:53.697363 IP localhost.44195 > localhost.6666: Flags [P.], seq 1:6, ack 1, win 342, options [nop,nop,TS val 10249525 ecr 10249307], length 5

02:57:53.697390 IP localhost.6666 > localhost.44195: Flags [.], ack 6, win 342, options [nop,nop,TS val 10249525 ecr 10249525], length 0

02:57:55.309532 IP localhost.44195 > localhost.6666: Flags [P.], seq 6:11, ack 1, win 342, options [nop,nop,TS val 10249686 ecr 10249525], length 5

02:57:55.309576 IP localhost.6666 > localhost.44195: Flags [.], ack 11, win 342, options [nop,nop,TS val 10249686 ecr 10249686], length 0

02:57:58.395206 IP localhost.44195 > localhost.6666: Flags [P.], seq 11:16, ack 1, win 342, options [nop,nop,TS val 10249994 ecr 10249686], length 5

02:57:58.395239 IP localhost.6666 > localhost.44195: Flags [.], ack 16, win 342, options [nop,nop,TS val 10249994 ecr 10249994], length 0

02:58:01.858094 IP localhost.44195 > localhost.6666: Flags [P.], seq 16:21, ack 1, win 342, options [nop,nop,TS val 10250341 ecr 10249994], length 5

02:58:01.858126 IP localhost.6666 > localhost.44195: Flags [.], ack 21, win 342, options [nop,nop,TS val 10250341 ecr 10250341], length 0

02:58:04.132584 IP localhost.44195 > localhost.6666: Flags [P.], seq 21:26, ack 1, win 342, options [nop,nop,TS val 10250568 ecr 10250341], length 5

02:58:04.132609 IP localhost.6666 > localhost.44195: Flags [.], ack 26, win 342, options [nop,nop,TS val 10250568 ecr 10250568], length 0

02:58:05.895704 IP localhost.44195 > localhost.6666: Flags [P.], seq 26:31, ack 1, win 342, options [nop,nop,TS val 10250744 ecr 10250568], length 5

02:58:05.895728 IP localhost.6666 > localhost.44195: Flags [.], ack 31, win 342, options [nop,nop,TS val 10250744 ecr 10250744], length 0

02:58:07.150265 IP localhost.44195 > localhost.6666: Flags [P.], seq 31:36, ack 1, win 342, options [nop,nop,TS val 10250870 ecr 10250744], length 5

02:58:07.150288 IP localhost.6666 > localhost.44195: Flags [.], ack 36, win 342, options [nop,nop,TS val 10250870 ecr 10250870], length 0

02:58:08.349124 IP localhost.44195 > localhost.6666: Flags [P.], seq 36:41, ack 1, win 342, options [nop,nop,TS val 10250990 ecr 10250870], length 5

02:58:08.349156 IP localhost.6666 > localhost.44195: Flags [.], ack 41, win 342, options [nop,nop,TS val 10250990 ecr 10250990], length 0

02:58:09.906223 IP localhost.44195 > localhost.6666: Flags [P.], seq 41:46, ack 1, win 342, options [nop,nop,TS val 10251145 ecr 10250990], length 5

02:58:09.906247 IP localhost.6666 > localhost.44195: Flags [.], ack 46, win 342, options [nop,nop,TS val 10251145 ecr 10251145], length 0

有9組數(shù)據(jù)包的發(fā)送。（這里的Flags [P.]代表Push的含義）

此時(shí)服務(wù)器還沒有close掉連接，實(shí)現(xiàn)了客戶端?；钸B接的功能。然后我們停止發(fā)送ping，過了5秒后tcpdump就會輸出一組：

02:58:14.811761 IP localhost.6666 > localhost.44195: Flags [F.], seq 1, ack 46, win 342, options [nop,nop,TS val 10251636 ecr 10251145], length 0
02:58:14.816420 IP localhost.44195 > localhost.6666: Flags [.], ack 2, win 342, options [nop,nop,TS val 10251637 ecr 10251636], length 0
服務(wù)端那邊發(fā)送了FIN包，說明服務(wù)端close掉了連接。服務(wù)端的輸出如下：

~/codeDir/phpCode/hyperf-skeleton # php server.php

string(17) "Client: Connect 1"

string(5) "ping

"

string(5) "ping

"

string(5) "ping

"

string(5) "ping

"

string(5) "ping

"

string(5) "ping

"

string(5) "ping

"

string(5) "ping

"

string(5) "ping

"

string(10) "close fd 1"

然后我們在客戶端那邊ctrl + c來關(guān)閉連接：

~/codeDir/phpCode/hyperf-skeleton # nc 127.0.0.1 6666

ping

ping

ping

ping

ping

ping

ping

ping

ping

^Cpunt!

~/codeDir/phpCode/hyperf-skeleton #

此時(shí)，tcpdump的輸出如下：

03:03:02.257667 IP localhost.44195 > localhost.6666: Flags [F.], seq 46, ack 2, win 342, options [nop,nop,TS val 10280414 ecr 10251636], length 0

03:03:02.257734 IP localhost.6666 > localhost.44195: Flags [R], seq 2678621620, win 0, length 0

應(yīng)用層心跳

1、制定ping/pong協(xié)議（mysql等自帶ping協(xié)議）

2、客戶端靈活的發(fā)送ping心跳包

3、服務(wù)端OnRecive檢查可用性回復(fù)pong

例如：

$server->on('receive', function (\Swoole\Server $server, $fd, $reactor_id, $data)
{
if ($data == 'ping')
{
checkDB();
checkServiceA();
checkRedis();
$server->send('pong');
}
});

結(jié)論

1、tcp的keepalive最簡單，但是有兼容性問題，不夠靈活

2、swoole提供的keepalive最實(shí)用，但是需要客戶端配合，復(fù)雜度適中

3、應(yīng)用層的keepalive最靈活但是最麻煩

以上就是詳解PHP Swoole長連接常見問題的詳細(xì)內(nèi)容，更多關(guān)于PHP Swoole長連接常見問題的資料請關(guān)注腳本之家其它相關(guān)文章！