nsq双机集群部署
2021年4月02日 13:27
问题
双机nsq如果部署集群,如何保证高可用性
1. 单节点nsqlookup故障?
2. 单节点nsq故障?
3. 消息丢失?
环境说明
两台机器
机器A 192.168.120.1
机器B 192.168.120.101
部署过程
部署方法
1. 机器A部署一套nsqlookup+nsqd
2. 机器B部署一套nsqlookup+nsqd
3. 生产者将消息同时写入两个nsqd
3. 消费者监听两个nsqlookup
结构图
机器A
nsqlookupd -broadcast-address 192.168.120.1 nsqd -lookupd-tcp-address=192.168.120.1:4160 -lookupd-tcp-address=192.168.120.101:4160 -broadcast-address 192.168.120.1
机器B
nsqlookupd -broadcast-address 192.168.120.101 nsqd -lookupd-tcp-address=192.168.120.1:4160 -lookupd-tcp-address=192.168.120.101:4160 -broadcast-address 192.168.120.101
pynsq测试脚本
生产者
import nsq import tornado.ioloop import time def pub_message(): writer.pub('test', str(time.strftime('%H:%M:%S')).encode("utf-8"), finish_pub) def finish_pub(conn, data): print(data) # 写入两个nsq好处: 1.防止nsqd单点故障 2.防止消息丢失 writer = nsq.Writer(['192.168.120.101:4150', '192.168.120.1:4150']) tornado.ioloop.PeriodicCallback(pub_message, 1000).start() nsq.run()
消费者
import nsq def handler(message): print(message, message.id, message.timestamp, message.attempts, message.body) return True # 防止nsqlookup故障 r = nsq.Reader(message_handler=handler, lookupd_http_addresses=['http://192.168.120.1:4161', 'http://192.168.120.101:4161'], topic='test', channel='abc', lookupd_poll_interval=15) nsq.run()
总结
只要有一个nsqlookup和一个nsqd存活,系统就不会挂