getAllGatewayAddressesFromRegister fail. tcp://0.0.0.0:1236 return false

binzhizhu

在workman中跑常规的ws服务，register、gateway、business偶发会出现上面的报错:Exception: getAllGatewayAddressesFromRegister fail. tcp://0.0.0.0:1236

之前没有遇到过，看报错的描述是因为连不上注册中心，导致拿不到gateway的通讯地址？请问下这是什么原因导致的？是偶发，不是致命的错误。

review代码是通过注册中心获取gateway通讯地址出错

/**
 * 获取通过注册中心获取所有 gateway 通讯地址
 *
 * @return array
 * @throws Exception
 */
protected static function getAllGatewayAddressesFromRegister()
{
    static $addresses_cache, $last_update;
    $time_now = time();
    $expiration_time = 1;
    $register_addresses = (array)static::$registerAddress;
    if(empty($addresses_cache) || $time_now - $last_update > $expiration_time) {
        foreach ($register_addresses as $register_address) {
            set_error_handler(function(){});
            $client = stream_socket_client('tcp://' . $register_address, $errno, $errmsg, static::$connectTimeout);
            restore_error_handler();
            if ($client) {
                break;
            }
        }
        if (!$client) {
            throw new Exception('Can not connect to tcp://' . $register_address . ' ' . $errmsg);
        }

        fwrite($client, '{"event":"worker_connect","secret_key":"' . static::$secretKey . '"}' . "\n");
        stream_set_timeout($client, 5);
        $ret = fgets($client, 655350);
        if (!$ret || !$data = json_decode(trim($ret), true)) {
            throw new Exception('getAllGatewayAddressesFromRegister fail. tcp://' .
                $register_address . ' return ' . var_export($ret, true));
        }
        $last_update = $time_now;
        $addresses_cache = $data['addresses'];
    }
    if (!$addresses_cache) {
        throw new Exception('Gateway::getAllGatewayAddressesFromRegister() with registerAddress:' .
            json_encode(static::$registerAddress) . '  return ' . var_export($addresses_cache, true));
    }
    return $addresses_cache;
}
6886 5 0
5个回答
walkor

发下你的调用代码

  • binzhizhu 2021-01-14

    就是常规的sendToGroup、Gateway::joinGroup、bindUid这种用法，没有特别的，然后日志会偶发出现上述报错：

    2021-01-14 17:17:46 客户端:7f0000010b5800000006登录
    2021-01-14 17:17:46 用户ID:[1]已绑定客户端ID:[7f0000010b5800000006]
    2021-01-14 17:17:46 客户端[7f0000010b5800000006]发送了消息{"jsonrpc":"2.0","method":"ws:subscribe","id":35204388.75511863,"params":{"topic":"im_live_room_30593_admin","uid":1}}，消息uuid是7f0000010b58000000065186636166760911202
    2021-01-14 17:17:46 客户端:7f0000010b5800000006订阅频道
    2021-01-14 17:17:46 客户端:[7f0000010b5800000006]订阅了频道[im_live_room_30593_admin]
    Exception 'Exception' with message 'getAllGatewayAddressesFromRegister fail. tcp://0.0.0.0:1236 return false'

    in /home/wwwroot/live-ecs.demo1.umworks.com/vendor/workerman/gatewayclient/Gateway.php:1307

    Stack trace:

    0 /home/wwwroot/live-ecs.demo1.umworks.com/vendor/workerman/gatewayclient/Gateway.php(726): GatewayClient\Gateway::getAllGatewayAddressesFromRegister()

    1 /home/wwwroot/live-ecs.demo1.umworks.com/vendor/workerman/gatewayclient/Gateway.php(701): GatewayClient\Gateway::getAllGatewayAddress()

    2 /home/wwwroot/live-ecs.demo1.umworks.com/vendor/workerman/gatewayclient/Gateway.php(498): GatewayClient\Gateway::getBufferFromAllGateway(Array)

    3 /home/wwwroot/live-ecs.demo1.umworks.com/modules/websocket/commands/RegisterWorker.php(77): GatewayClient\Gateway::getAllGroupIdList()

    4 /home/wwwroot/live-ecs.demo1.umworks.com/vendor/workerman/workerman/Events/Select.php(232): pengpai\modules\websocket\commands\RegisterWorker->pengpai\modules\websocket\commands{closure}()

    5 /home/wwwroot/live-ecs.demo1.umworks.com/vendor/workerman/workerman/Events/Select.php(280): Workerman\Events\Select->tick()

    6 /home/wwwroot/live-ecs.demo1.umworks.com/vendor/workerman/workerman/Worker.php(2408): Workerman\Events\Select->loop()

    7 /home/wwwroot/live-ecs.demo1.umworks.com/vendor/workerman/gateway-worker/src/Register.php(86): Workerman\Worker->run()

    8 /home/wwwroot/live-ecs.demo1.umworks.com/vendor/workerman/workerman/Worker.php(1541): GatewayWorker\Register->run()

    9 /home/wwwroot/live-ecs.demo1.umworks.com/vendor/workerman/workerman/Worker.php(1371): Workerman\Worker::forkOneWorkerForLinux(Object(GatewayWorker\Register))

    10 /home/wwwroot/live-ecs.demo1.umworks.com/vendor/workerman/workerman/Worker.php(1345): Workerman\Worker::forkWorkersForLinux()

    11 /home/wwwroot/live-ecs.demo1.umworks.com/vendor/workerman/workerman/Worker.php(1675): Workerman\Worker::forkWorkers()

    12 /home/wwwroot/live-ecs.demo1.umworks.com/vendor/workerman/workerman/Worker.php(1624): Workerman\Worker::monitorWorkersForLinux()

    13 /home/wwwroot/live-ecs.demo1.umworks.com/vendor/workerman/workerman/Worker.php(548): Workerman\Worker::monitorWorkers()

    14 /home/wwwroot/live-ecs.demo1.umworks.com/app/commands/DaemonController.php(134): Workerman\Worker::runAll()

    15 [internal function]: pengpai\commands\DaemonController->actionService('ws', 'start')

    16 /home/wwwroot/live-ecs.demo1.umworks.com/libs/yii2/base/InlineAction.php(57): call_user_func_array(Array, Array)

    17 /home/wwwroot/live-ecs.demo1.umworks.com/libs/yii2/base/Controller.php(180): yii\base\InlineAction->runWithParams(Array)

    18 /home/wwwroot/live-ecs.demo1.umworks.com/libs/yii2/console/Controller.php(179): yii\base\Controller->runAction('service', Array)

    19 /home/wwwroot/live-ecs.demo1.umworks.com/libs/yii2/base/Module.php(528): yii\console\Controller->runAction('service', Array)

    20 /home/wwwroot/live-ecs.demo1.umworks.com/libs/yii2/console/Application.php(180): yii\base\Module->runAction('daemon/service', Array)

    21 /home/wwwroot/live-ecs.demo1.umworks.com/libs/yii2/console/Application.php(147): yii\console\Application->runAction('daemon/service', Array)

    22 /home/wwwroot/live-ecs.demo1.umworks.com/libs/yii2/base/Application.php(387): yii\console\Application->handleRequest(Object(pengpai\components\console\Request))

    23 /home/wwwroot/live-ecs.demo1.umworks.com/app/components/console/Application.php(24): yii\base\Application->run()

    24 /home/wwwroot/live-ecs.demo1.umworks.com/yii(9): pengpai\components\console\Application->run()

    25 {main}

    Worker[30228] process terminated
    worker[ws-register:30228] exit with status 256
  • binzhizhu 2021-01-14

    你可以看看上面日志，下一次又是正常,看status也是正常的

    Workerman version:3.5.31 PHP version:7.2.24
    start time:2021-01-14 17:19:47 run 0 days 0 hours
    load average: 0.37, 1, 1 event-loop:\Workerman\Events\Select
    3 workers 33 processes
    worker_name exit_status exit_count
    ws-register 0 0
    ws-gateway 0 0
    ws-business 0 0
    ----------------------------------------------PROCESS STATUS---------------------------------------------------
    pid memory listening worker_name connections send_fail timers total_request qps status
    858 24M text://0.0.0.0:1236 ws-register 32 0 1 34 0 [idle]
    859 24M websocket://0.0.0.0:8484 ws-gateway 17 0 1 18 0 [idle]
    860 24M websocket://0.0.0.0:8484 ws-gateway 17 0 1 18 0 [idle]
    863 24M websocket://0.0.0.0:8484 ws-gateway 17 0 1 18 0 [idle]
    864 24M websocket://0.0.0.0:8484 ws-gateway 17 0 1 18 0 [idle]
    865 24M websocket://0.0.0.0:8484 ws-gateway 18 0 1 18 0 [idle]
    866 24M websocket://0.0.0.0:8484 ws-gateway 17 0 1 18 0 [idle]
    867 24M websocket://0.0.0.0:8484 ws-gateway 19 0 1 18 0 [idle]
    868 24M websocket://0.0.0.0:8484 ws-gateway 18 0 1 28 0 [idle]
    869 24M websocket://0.0.0.0:8484 ws-gateway 17 0 1 18 0 [idle]
    870 24M websocket://0.0.0.0:8484 ws-gateway 18 0 1 31 0 [idle]
    871 24M websocket://0.0.0.0:8484 ws-gateway 18 0 1 18 0 [idle]
    873 24M websocket://0.0.0.0:8484 ws-gateway 17 0 1 18 0 [idle]
    874 24M websocket://0.0.0.0:8484 ws-gateway 18 0 1 18 0 [idle]
    875 24M websocket://0.0.0.0:8484 ws-gateway 17 0 1 18 0 [idle]
    876 24M websocket://0.0.0.0:8484 ws-gateway 18 0 1 18 0 [idle]
    877 24M websocket://0.0.0.0:8484 ws-gateway 19 0 1 18 0 [idle]
    879 28M none ws-business 17 0 0 9 0 [idle]
    880 24M none ws-business 17 0 0 4 0 [idle]
    881 24M none ws-business 17 0 0 5 0 [idle]
    883 24M none ws-business 17 0 0 6 0 [idle]
    884 24M none ws-business 17 0 0 5 0 [idle]
    885 24M none ws-business 17 0 0 3 0 [idle]
    886 24M none ws-business 17 0 0 4 0 [idle]
    887 24M none ws-business 17 0 0 3 0 [idle]
    888 24M none ws-business 17 0 0 5 0 [idle]
    889 24M none ws-business 17 0 0 5 0 [idle]
    890 24M none ws-business 17 0 0 5 0 [idle]
    891 24M none ws-business 17 0 0 5 0 [idle]
    892 24M none ws-business 17 0 0 5 0 [idle]
    893 26M none ws-business 17 0 0 9 0 [idle]
    895 24M none ws-business 17 0 0 5 0 [idle]
    896 24M none ws-business 17 0 0 3 0 [idle]
    ----------------------------------------------PROCESS STATUS---------------------------------------------------
    Summary 798M - - 586 0 17 426 0 [Summary]
  • walkor 2021-01-15

    gatewayClient么？发下调用代码
  • binzhizhu 2021-01-15

    @1: 

            try {
            $sendData = [
                'id' => StringHelper::uuid(),
                'result' => $statsInfo,
                'jsonrpc' => '2.0',
                'method' => 'apiGetLiveRoomStatsInfo'
            ];

    // $clientIdList = Gateway::getClientIdListByGroup($channel);
    // $clientIdList = array_unique($clientIdList);
    // echo "当前分组{$channel}客户端ID列表" . Json::encode($clientIdList) . "\n";
    // if (!$clientIdList) {
    // continue;
    // }
    // foreach ($clientIdList as $client_id) {
    // Gateway::sendToClient($client_id, Json::encode($sendData));
    // }
    echo "发送数据:" . Json::encode($sendData) . "\n";
    Gateway::sendToGroup($channel, Json::encode($sendData));
    } catch (Exception $e) {
    $this->handleError("广播直播间[{$roomId}]统计数据发生错误", ['e' => $e]);
    continue;
    }

    GatewayClient\Gateway 的简单调用 sendToGroup ，不单是这处调用有问题，偶发的
  • binzhizhu 2021-01-15

    @1: 可以帮忙看看什么问题吗 调用上没什么特别的吧 自己也没做一些中转处理之类的
  • walkor 2021-01-15

    设置 Gateway::$registerAddress = xxx; 的代码在哪？贴下代码。
  • binzhizhu 2021-01-15

    @1: 我们在bootstrap的时候 定义了注册中心地址以及端口

    \GatewayClient\Gateway::$registerAddress = $yii2->params['websocketRegisterIP'] . ':' . $yii2->params['websocketRegisterPort'];
  • binzhizhu 2021-01-15

    @1:同时 在gatewayworker 以及businessworker 启动前也初始化了一次 跟上面一致

    websocketRegisterIP 就是0.0.0.0 websocketRegisterPort 是1236
  • walkor 2021-01-15

    手册写了，\GatewayClient\Gateway::$registerAddress 不能用 0.0.0.0，要用ip。如果gatewayClient和GatewayWorker在一台服务器，就用127.0.0.1:xxx。
  • binzhizhu 2021-01-15

    @1: 首先是单机，刚刚把registerAddress改成了127.0.0.1

    Workerman[ws] start in DEBUG mode
    ------------------------------------------- WORKERMAN --------------------------------------------
    Workerman version:3.5.31 PHP version:7.2.24
    -------------------------------------------- WORKERS ---------------------------------------------
    proto user worker listen processes status
    tcp root ws-register text://127.0.0.1:1236 1 [OK]
    tcp root ws-gateway websocket://0.0.0.0:8484 16 [OK]
    tcp root ws-business none

    gateway暴露0.0.0.0监听所有网卡，这样还是会出现同样的报错

    getAllGatewayAddressesFromRegister fail. tcp://127.0.0.1:1236 return false
chapin

你好，请问这个问题有解决嘛，我也遇到了相同的问题

  • binzhizhu 2021-01-14

    东哥 你好 暂时没有解决，等作者看看是什么原因吧，我定位不到什么原因，注册中心、gateway都是正常的
binzhizhu

没人回复我吗 处理这个问题

  • 暂无评论
walkor

使用gatewayClient时，\GatewayClient\Gateway::$registerAddress 不能用 0.0.0.0。
截图
手册地址：http://doc2.workerman.net/push-in-other-project.html

科普下基础，0.0.0.0 只能在监听的时候用(作为服务端)，其它时候(作为客户端)都得用实际ip。

  • binzhizhu 2021-01-15

    谢谢作者大大~~~

  • binzhizhu 2021-01-15

    Workerman[ws] start in DEBUG mode
    ------------------------------------------- WORKERMAN --------------------------------------------
    Workerman version:3.5.31 PHP version:7.2.24
    -------------------------------------------- WORKERS ---------------------------------------------
    proto user worker listen processes status
    tcp root ws-register text://127.0.0.1:1236 1 [OK]
    tcp root ws-gateway websocket://0.0.0.0:8484 16 [OK]
    tcp root ws-business none

    gateway暴露0.0.0.0监听所有网卡，这样还是会出现同样的报错

    getAllGatewayAddressesFromRegister fail. tcp://127.0.0.1:1236 return false

    还是报错
  • walkor 2021-01-15

    gatewayClient 和 GatewayWorker 是同一台服务器么？
  • binzhizhu 2021-01-15

    @1:是的 同一台 单机部署的，我们正式环境才会用分布式
  • binzhizhu 2021-01-15

    @1: 单机部署，registerAddress注册中心的地址按照手册修正为127.0.0.1 然后gateway 0.0.0.0监听 现在是这样也会出现那样的报错 tpc:127.0.0.1:1236 return false
  • walkor 2021-01-15

    运行 ps auxf，看下是不是运行了多个gatewayWorker。
  • binzhizhu 2021-01-15

    @1:

    [root@vm]# ps aux | grep Worker
    root 4405 0.0 0.4 308408 33932 ? S 13:01 0:00 WorkerMan: worker process ws-business none
    root 4847 0.0 0.4 308408 33936 ? S 14:04 0:00 WorkerMan: worker process ws-business none
    root 5008 0.0 0.4 308408 33936 ? S 14:25 0:00 WorkerMan: worker process ws-business none
    root 9327 0.0 0.4 308408 33928 ? S 13:25 0:00 WorkerMan: worker process ws-business none
    root 9758 0.0 0.3 306360 31712 ? S 14:28 0:00 WorkerMan: worker process ws-business none
    root 15324 0.0 0.4 308408 33920 ? S 13:08 0:00 WorkerMan: worker process ws-business none
    root 15769 0.0 0.4 308408 33936 ? S 14:11 0:00 WorkerMan: worker process ws-business none
    root 15924 0.0 0.3 306360 31716 ? S 14:32 0:00 WorkerMan: worker process ws-business none
    root 16788 0.0 0.3 304312 29880 ? S 12:28 0:00 WorkerMan: master process start_file=/home/wwwroot/live-ecs.demo1.umworks.com/yii
    root 16792 0.0 0.3 304312 30580 ? S 12:28 0:01 WorkerMan: worker process ws-gateway websocket://0.0.0.0:8484
    root 16793 0.0 0.3 304312 30580 ? S 12:28 0:01 WorkerMan: worker process ws-gateway websocket://0.0.0.0:8484
    root 16794 0.0 0.3 304312 30580 ? S 12:28 0:01 WorkerMan: worker process ws-gateway websocket://0.0.0.0:8484
    root 16795 0.0 0.3 304312 30580 ? S 12:28 0:01 WorkerMan: worker process ws-gateway websocket://0.0.0.0:8484
    root 16796 0.0 0.3 304312 30580 ? S 12:28 0:01 WorkerMan: worker process ws-gateway websocket://0.0.0.0:8484
    root 16797 0.0 0.3 304312 30580 ? S 12:28 0:01 WorkerMan: worker process ws-gateway websocket://0.0.0.0:8484
    root 16798 0.0 0.3 304312 30580 ? S 12:28 0:01 WorkerMan: worker process ws-gateway websocket://0.0.0.0:8484
    root 16799 0.0 0.3 304312 30580 ? S 12:28 0:01 WorkerMan: worker process ws-gateway websocket://0.0.0.0:8484
    root 16800 0.0 0.3 304312 30580 ? S 12:28 0:01 WorkerMan: worker process ws-gateway websocket://0.0.0.0:8484
    root 16802 0.0 0.3 304312 30580 ? S 12:28 0:01 WorkerMan: worker process ws-gateway websocket://0.0.0.0:8484
    root 16803 0.0 0.3 304312 30580 ? S 12:28 0:01 WorkerMan: worker process ws-gateway websocket://0.0.0.0:8484
    root 16804 0.0 0.3 304312 30580 ? S 12:28 0:01 WorkerMan: worker process ws-gateway websocket://0.0.0.0:8484
    root 16805 0.0 0.3 304312 30580 ? S 12:28 0:01 WorkerMan: worker process ws-gateway websocket://0.0.0.0:8484
    root 16806 0.0 0.3 304312 30580 ? S 12:28 0:01 WorkerMan: worker process ws-gateway websocket://0.0.0.0:8484
    root 16807 0.0 0.3 304312 30580 ? S 12:28 0:01 WorkerMan: worker process ws-gateway websocket://0.0.0.0:8484
    root 16808 0.0 0.3 304312 30580 ? S 12:28 0:01 WorkerMan: worker process ws-gateway websocket://0.0.0.0:8484
    root 20590 0.0 0.3 306360 31716 ? S 14:35 0:00 WorkerMan: worker process ws-business none
    root 26318 0.0 0.4 308408 33940 ? S 13:36 0:00 WorkerMan: worker process ws-business none
    root 26448 0.0 0.4 308408 33920 ? S 13:57 0:00 WorkerMan: worker process ws-business none
    root 26592 0.0 0.4 308408 33924 ? S 14:18 0:00 WorkerMan: worker process ws-business none
    root 26744 0.0 0.3 306360 31716 ? S 14:39 0:00 WorkerMan: worker process ws-business none
    root 28287 0.0 0.3 304312 30168 ? S 14:40 0:00 WorkerMan: worker process ws-register text://127.0.0.1:1236
    root 28289 0.0 0.0 112720 972 pts/0 S+ 14:40 0:00 grep --color=auto Worker
    root 30925 0.0 0.4 308408 33936 ? S 13:39 0:00 WorkerMan: worker process ws-business none
    root 31063 0.0 0.4 308408 33920 ? S 14:00 0:00 WorkerMan: worker process ws-business none
    root 31207 0.0 0.4 308408 33940 ? S 14:21 0:00 WorkerMan: worker process ws-business none
    root 31583 0.0 0.3 304312 29668 ? S Jan14 0:00 WorkerMan: master process start_file=/home/wwwroot/live-ecs.demo1.umworks.com/yii
    root 31584 0.1 0.5 318780 44736 ? S Jan14 1:41 WorkerMan: worker process live-room-stats-info none

    没有
  • binzhizhu 2021-01-15

    @1: 只有一个项目用到了workman，没有占用的一些问题
  • binzhizhu 2021-01-15

    @1: 我们刚刚更新了正式环境 分布式部署 也存在这样的问题 这个问题就是这两周才有的 之前没有出现过的 我们的机器都是克隆的 配置都没改过
  • binzhizhu 2021-01-15

    @1: 分部署部署，一台机作为注册中心，跑了register、gateway、business，其他几台机器单独跑gateway、business，连的是注册中心的IP，lanIp是本机内网IP
  • walkor 2021-01-15

    看到有2个 start_file=/home/wwwroot/live-ecs.demo1.umworks.com/yii 主进程，应该是开了2个gatewayWorker或者workerman程序。运行 php start.php stop，然后运行 ps auxf，kill掉重复启动的gatewayWorker进程。
binzhizhu

已解决 是我们在注册启动后onWorkerStart 加了一个定时器导致的 具体原因不知 就会产生上诉问题

  • walkor 2021-01-15

    可能是你定时器里的业务逻辑影响了register服务的正常执行，比如死循环或者长时间阻塞等
  • binzhizhu 2021-01-15

    @1: 不是主进程问题，我全部kill掉，重启还是一样，我们在onWorkerStart的时候加了一个定时器 

    Timer::add(30, function () {
            $logFile = Yii::getAlias('@runtime/ws-counter/' . date('Y-m-d') . '.log');
            @mkdir(dirname($logFile), 0777, true);

            $time = date('Y-m-d H:i:s', time());

            $groupIds = Gateway::getAllGroupIdList();

            $output = new ConsoleOutput;
            $table = new Table($output);

            $table->setHeaders(['时间', '分组', '客户端数', '用户数']);

            foreach ($groupIds as $groupId) {
                $cCount = Gateway::getClientIdCountByGroup($groupId);
                $uCount = Gateway::getUidCountByGroup($groupId);
                if ($cCount === 0 && $uCount === 0) {

                    Gateway::ungroup($groupId);
                    continue;
                }

                $table->addRow([
                    $time,
                    $groupId,
                    $cCount,
                    $uCount,
                ]);
            }

            $table->addRow([
                $time,
                '总计',
                Gateway::getAllClientCount(),
                Gateway::getAllUidCount(),
            ]);

            $table->render();

        }, [], true);

    是想定时统计每个分组具体的client数 没有死循环的代码 很奇怪 把这逻辑还原就正常
  • walkor 2021-01-15

    你的报错应该就是这个定时器报的，register自己调用自己了，不要在register上运行业务，不支持。
  • binzhizhu 2021-01-15

    @1: 谢谢作者大大~~ 我们到时单独起个worker做这样的统计

相似问题
getAllGatewayAddressesFromRegister fail. tcp://127.0.0.1:2346 return false
getAllGatewayAddressesFromRegister fail. tcp://0.0.0.0:8282 return 'HTTP/1.1 200 Websocket '
Exception' with message 'getAllGatewayAddressesFromRegister fail. tcp://127.0.0.1:1238 return false
getAllGatewayAddressesFromRegister fail. tcp://127.0.0.1:54990 return '你好呀 7f0000010b5700000003'
getAllGatewayAddressesFromRegister fail. tcp://127.0.0.1:5408 return 'HTTP/1.1 400 Bad Request
Gateway::getAllGatewayAddressesFromRegister() with registerAddress return array
年代过于久远，无法发表回答
月贡献榜
1
walkor
🚀
2
rbb
3
six
4
nitron
5
Tinywan
♨️ 请叫我 webman 推广大使
赞助商
+成为赞助商

midjouney中文版 AI画图

一句话生成一张漂亮的图片

app,小程序,H5聊天源码

语音|视频|文字聊天

客服系统源码

小程序客服|h5客服|网页客服
🔝