将协调员标记为组(Kafka)已死 [英] Marking the coordinator dead for group(Kafka)

查看:223
本文介绍了将协调员标记为组(Kafka)已死的处理方法,对大家解决问题具有一定的参考价值,需要的朋友们下面随着小编来一起学习吧!

问题描述

我的一项服务在几个小时前停止工作.我已经重新启动服务器,但是此使用者进程未在Kafka中注册.这是消费者启动时显示的内容:

One of my services stopped working a few hours ago. I have restarted the server but this consumer process just isn't registering with Kafka. This is what shows up when the consumer is started :

2017-04-21 10:22:54.887  INFO 18036 --- [afka-consumer-1] o.a.k.c.c.internals.ConsumerCoordinator  : Revoking previously assigned partitions [] for group mysql-conversions-group
2017-04-21 10:22:54.890  INFO 18036 --- [afka-consumer-1] o.s.k.l.KafkaMessageListenerContainer    : partitions revoked:[]
2017-04-21 10:22:54.890  INFO 18036 --- [afka-consumer-1] o.a.k.c.c.internals.AbstractCoordinator  : (Re-)joining group mysql-conversions-group
2017-04-21 10:22:54.915  INFO 18036 --- [afka-consumer-1] o.a.k.c.c.internals.AbstractCoordinator  : Marking the coordinator li819-138.members.linode.com:9092 (id: 2147483646 rack: null) dead for group mysql-conversions-group

我有4个不同的消费群体,他们的表现还不错.有什么想法吗?

I have 4 different consumer groups and they are running just fine. Any thoughts?

更新

真的很奇怪.我们必须重新启动另一个使用者,然后重新启动后,第二个工作正常的使用者现在记录了相同的错误. 基本上,我们现在有两个由于某种原因无法连接到Kafka的使用者进程.

Its odd really. We had to restart another one of our Consumer, and after restart, the other consumer that was working fine a second ago is now logging the same error. Basically, we now have 2 consumer processes that can't connect to Kafka for some reason.

推荐答案

回答我自己的问题:

最后使它起作用.这不是经纪人问题也不是消费者问题. 发生的事情是,我们在晚上8:30左右在我们的两个Kafka经纪人上启用了防火墙.已经与Kafka建立联系的消费者工作良好.但是,当这些使用者之一重新启动时,他们无法连接到协调器节点.我们在两个Kafka代理上都禁用了防火墙,并且数据立即转发给了消费者.这是一个临时性的自动提款机,但至少我们有一些东西.

Finally got it to work. This was not a broker issue or a consumer issue. What happened was, we had enabled firewall on both of our Kafka brokers around 8:30 in the night. The consumers that were already connected with Kafka were working fine. However, when one of these consumers were restarted, they were unable to connect to the coordinator node. We disabled firewall on both the Kafka brokers, and the data was immediately relayed to the consumers. It is a temporary fix ATM, but at least we have something.

这篇关于将协调员标记为组(Kafka)已死的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!

查看全文
登录 关闭
扫码关注1秒登录
发送“验证码”获取 | 15天全站免登陆