AVStrem rtmp推流中断

[qian_e_]

你好，我们测试过程中发现，avstream模块rtmp推流模块有流中断的显现，排查了系统日志发现Cluster的日志中显示：

740687:2018-01-11 12:25:44.309  - DEBUG: ClusterManager - workerJoin, purpose: session worker: session.64437be4-0b1d-7085-25bc-5b5548921bc2 info: { ip: '172.17.44.221',

740692:2018-01-11 12:25:44.310  - DEBUG: Scheduler - Add worker: session.64437be4-0b1d-7085-25bc-5b5548921bc2 info: { ip: '172.17.44.221',

740697:2018-01-11 12:25:44.310  - WARN: Scheduler - Double adding worker: session.64437be4-0b1d-7085-25bc-5b5548921bc2

740698:2018-01-11 12:25:44.310  - DEBUG: ClusterManager - workerJoin, purpose: session worker: session.64437be4-0b1d-7085-25bc-5b5548921bc2 info: { ip: '172.17.44.221',

740703:2018-01-11 12:25:44.310  - DEBUG: Scheduler - Add worker: session.64437be4-0b1d-7085-25bc-5b5548921bc2 info: { ip: '172.17.44.221',

740708:2018-01-11 12:25:44.310  - WARN: Scheduler - Double adding worker: session.64437be4-0b1d-7085-25bc-5b5548921bc2

740709:2018-01-11 12:25:44.311  - DEBUG: ClusterManager - workerJoin, purpose: session worker: session.64437be4-0b1d-7085-25bc-5b5548921bc2 info: { ip: '172.17.44.221',

740714:2018-01-11 12:25:44.311  - DEBUG: Scheduler - Add worker: session.64437be4-0b1d-7085-25bc-5b5548921bc2 info: { ip: '172.17.44.221',

740719:2018-01-11 12:25:44.311  - WARN: Scheduler - Double adding worker: session.64437be4-0b1d-7085-25bc-5b5548921bc2

740720:2018-01-11 12:25:44.311  - DEBUG: ClusterManager - workerJoin, purpose: session worker: session.64437be4-0b1d-7085-25bc-5b5548921bc2 info: { ip: '172.17.44.221',

740725:2018-01-11 12:25:44.311  - DEBUG: Scheduler - Add worker: session.64437be4-0b1d-7085-25bc-5b5548921bc2 info: { ip: '172.17.44.221',

740730:2018-01-11 12:25:44.311  - WARN: Scheduler - Double adding worker: session.64437be4-0b1d-7085-25bc-5b5548921bc2

740753:2018-01-11 12:25:44.337  - DEBUG: ClusterManager - workerJoin, purpose: session worker: session.64437be4-0b1d-7085-25bc-5b5548921bc2 info: { ip: '172.17.44.221',

740758:2018-01-11 12:25:44.337  - DEBUG: Scheduler - Add worker: session.64437be4-0b1d-7085-25bc-5b5548921bc2 info: { ip: '172.17.44.221',

740763:2018-01-11 12:25:44.337  - WARN: Scheduler - Double adding worker: session.64437be4-0b1d-7085-25bc-5b5548921bc2

740764:2018-01-11 12:25:44.338  - DEBUG: ClusterManager - workerJoin, purpose: session worker: session.64437be4-0b1d-7085-25bc-5b5548921bc2 info: { ip: '172.17.44.221',

740769:2018-01-11 12:25:44.338  - DEBUG: Scheduler - Add worker: session.64437be4-0b1d-7085-25bc-5b5548921bc2 info: { ip: '172.17.44.221',

740774:2018-01-11 12:25:44.338  - WARN: Scheduler - Double adding worker: session.64437be4-0b1d-7085-25bc-5b5548921bc2

显示有很多的double worker，系统清理worker的时候，会把正在使用的worker清理掉，就会引起avstream推流终止，不知道该如何解决这种问题呢？

[qian_e_]

session.64437be4-0b1d-7085-25bc-5b5548921bc2.4101.log:

2018-01-11 11:05:13.235  - INFO: ErizoJS - Connecting to rabbitMQ server...
2018-01-11 11:05:13.389  - INFO: AmqpClient - Connecting to rabbitMQ server OK, hostPort: { host: '172.19.113.88', port: 5672 }
2018-01-11 11:05:13.854  - INFO: ErizoJS - session.64437be4-0b1d-7085-25bc-5b5548921bc2.4101 as rpc server ready
2018-01-11 11:05:13.992  - INFO: ErizoJS - session.64437be4-0b1d-7085-25bc-5b5548921bc2.4101 as monitor ready
2018-01-11 12:25:44.082  - WARN: AmqpClient - Late rpc reply: { data: 'ok', corrID: 639, type: 'callback' }
2018-01-11 12:25:44.086  - WARN: AmqpClient - Late rpc reply: { data: 'ok', corrID: 666, type: 'callback' }
2018-01-11 12:25:44.087  - WARN: AmqpClient - Late rpc reply: { data: 'ok', corrID: 675, type: 'callback' }
2018-01-11 12:25:44.354  - WARN: ErizoJS - Exiting on SIGTERM

avstream.9d2db18e-bd3e-4ff9-632b-ff69c843d9ae.4510.log:

2018-01-11 11:50:04.352  - INFO: ErizoJS - Connecting to rabbitMQ server...
2018-01-11 11:50:04.374  - INFO: AmqpClient - Connecting to rabbitMQ server OK, hostPort: { host: '172.19.113.88', port: 5672 }
2018-01-11 11:50:04.407  - INFO: ErizoJS - avstream.9d2db18e-bd3e-4ff9-632b-ff69c843d9ae.4510 as rpc server ready
2018-01-11 11:50:04.414  - INFO: ErizoJS - avstream.9d2db18e-bd3e-4ff9-632b-ff69c843d9ae.4510 as monitor ready
2018-01-11 12:17:46,459  - INFO: woogeen.RtspOut - Initial audio options channels 2, sample rate: 48000
2018-01-11 12:17:47,216  - INFO: woogeen.RtspOut - Initial video options: 1280x720
2018-01-11 12:25:44.400  - ERROR: Connections - Fault detected on controller (type: node id: session.64437be4-0b1d-7085-25bc-5b5548921bc2.4101 ) of connection: 398035680909833800 and remove it
2018-01-11 12:25:44.401  - ERROR: Connections - Fault detected on controller (type: node id: session.64437be4-0b1d-7085-25bc-5b5548921bc2.4101 ) of connection: 543392283920735200 and remove it
2018-01-11 12:25:44,850  - INFO: woogeen.RtspOut - Closing rtmp://cc1.cssltest.net/origin/C0968BBA2745F5089C33DC5901307461
2018-01-11 12:25:44,850  - WARN: woogeen.RtspOut - No input frames available
2018-01-11 12:25:44.851  - WARN: ErizoJS - Exiting on SIGTERM

avstream.9d2db18e-bd3e-4ff9-632b-ff69c843d9ae.4509.log:

2018-01-11 11:10:31.269  - INFO: ErizoJS - Connecting to rabbitMQ server...
2018-01-11 11:10:31.290  - INFO: AmqpClient - Connecting to rabbitMQ server OK, hostPort: { host: '172.19.113.88', port: 5672 }
2018-01-11 11:10:31.320  - INFO: ErizoJS - avstream.9d2db18e-bd3e-4ff9-632b-ff69c843d9ae.4509 as rpc server ready
2018-01-11 11:10:31.326  - INFO: ErizoJS - avstream.9d2db18e-bd3e-4ff9-632b-ff69c843d9ae.4509 as monitor ready
2018-01-11 12:17:36,252  - INFO: woogeen.RtspOut - Initial audio options channels 2, sample rate: 48000
2018-01-11 12:17:37,006  - INFO: woogeen.RtspOut - Initial video options: 1280x720
2018-01-11 12:17:41,424  - INFO: woogeen.RtspOut - Closing rtmp://cc1.cssltest.net/origin/C0968BBA2745F5089C33DC5901307461
2018-01-11 12:17:41,424  - WARN: woogeen.RtspOut - No input frames available
2018-01-11 12:17:41.425  - WARN: ErizoJS - Exiting on SIGTERM

[qian_e_]

不知道为什么会有两个avstream日志推同一份流这个流实际开始时间是 12：17 ，正常应该是13：13结束，但是12：25就终止了，从日志分析说是session冲突，不清楚为什么这个会冲突。

[Qiujiao_W_Intel]

你使用的MCU是什么版本，部署环境是怎样的，多台机器分布式部署的吗？

[qian_e_]

Qiujiao W. (Intel) wrote:

你使用的MCU是什么版本，部署环境是怎样的，多台机器分布式部署的吗？

我们MCU是3.4的版本，目前是分布式部署，服务器中间使用专线连接的，延时在30ms内