【问题解决】Kafka报错 Bootstrap broker x.x.x.x:9092 (id: -1 rack: null) disconnected

问题复现

近日针对某一客户需求开发了一个需要使用Kafka的功能，功能是什么暂且不论，在本地虚机的Kafka连接一切正常遂放到测试服务器上验证功能，以下是监听topic成功和警告报错：

2023-05-09 10:22:23 [localhost-startStop-1] INFO  org.apache.kafka.clients.consumer.ConsumerConfig - ConsumerConfig values: allow.auto.create.topics = trueauto.commit.interval.ms = 5000auto.offset.reset = earliestbootstrap.servers = [10.39.48.113:9092]check.crcs = trueclient.dns.lookup = use_all_dns_ipsclient.id = consumer-enn-jiuqi-1client.rack = connections.max.idle.ms = 540000default.api.timeout.ms = 60000enable.auto.commit = falseexclude.internal.topics = truefetch.max.bytes = 52428800fetch.max.wait.ms = 500fetch.min.bytes = 1group.id = enn-jiuqigroup.instance.id = nullheartbeat.interval.ms = 3000interceptor.classes = []internal.leave.group.on.close = trueinternal.throw.on.fetch.stable.offset.unsupported = falseisolation.level = read_uncommittedkey.deserializer = class org.apache.kafka.common.serialization.StringDeserializermax.partition.fetch.bytes = 1048576max.poll.interval.ms = 300000max.poll.records = 500metadata.max.age.ms = 300000metric.reporters = []metrics.num.samples = 2metrics.recording.level = INFOmetrics.sample.window.ms = 30000partition.assignment.strategy = [class org.apache.kafka.clients.consumer.RangeAssignor, class org.apache.kafka.clients.consumer.CooperativeStickyAssignor]receive.buffer.bytes = 65536reconnect.backoff.max.ms = 1000reconnect.backoff.ms = 50request.timeout.ms = 30000retry.backoff.ms = 100sasl.client.callback.handler.class = nullsasl.jaas.config = nullsasl.kerberos.kinit.cmd = /usr/bin/kinitsasl.kerberos.min.time.before.relogin = 60000sasl.kerberos.service.name = nullsasl.kerberos.ticket.renew.jitter = 0.05sasl.kerberos.ticket.renew.window.factor = 0.8sasl.login.callback.handler.class = nullsasl.login.class = nullsasl.login.refresh.buffer.seconds = 300sasl.login.refresh.min.period.seconds = 60sasl.login.refresh.window.factor = 0.8sasl.login.refresh.window.jitter = 0.05sasl.mechanism = GSSAPIsecurity.protocol = PLAINTEXTsecurity.providers = nullsend.buffer.bytes = 131072session.timeout.ms = 45000socket.connection.setup.timeout.max.ms = 30000socket.connection.setup.timeout.ms = 10000ssl.cipher.suites = nullssl.enabled.protocols = [TLSv1.2]ssl.endpoint.identification.algorithm = httpsssl.engine.factory.class = nullssl.key.password = nullssl.keymanager.algorithm = SunX509ssl.keystore.certificate.chain = nullssl.keystore.key = nullssl.keystore.location = nullssl.keystore.password = nullssl.keystore.type = JKSssl.protocol = TLSv1.2ssl.provider = nullssl.secure.random.implementation = nullssl.trustmanager.algorithm = PKIXssl.truststore.certificates = nullssl.truststore.location = nullssl.truststore.password = nullssl.truststore.type = JKSvalue.deserializer = class org.apache.kafka.common.serialization.StringDeserializer2023-05-09 10:22:23 [localhost-startStop-1] INFO  org.apache.kafka.common.utils.AppInfoParser - Kafka version: 3.0.1
2023-05-09 10:22:23 [localhost-startStop-1] INFO  org.apache.kafka.common.utils.AppInfoParser - Kafka commitId: 8e30984f43e64d8b
2023-05-09 10:22:23 [localhost-startStop-1] INFO  org.apache.kafka.common.utils.AppInfoParser - Kafka startTimeMs: 1683598943212
2023-05-09 10:22:23 [localhost-startStop-1] INFO  org.apache.kafka.clients.consumer.KafkaConsumer - [Consumer clientId=consumer-test-1, groupId=test-group] Subscribed to topic(s): sync_user2023-05-09 10:23:50 [org.springframework.kafka.KafkaListenerEndpointContainer#0-0-C-1] WARN  org.apache.kafka.clients.NetworkClient - [Consumer clientId=consumer-test-1, groupId=test-group] Bootstrap broker 10.39.48.113:9092 (id: -1 rack: null) disconnected

可以注意到日志最后有1条警告，其实我只放了一条，这个信息在日志里多滴很。

分析问题

客户对接方发送了一条消息问我们消费到没，我一查日志，满屏是WARN提示 Bootstrap broker 10.39.48.113:9092 (id: -1 rack: null) disconnected，真是小刀拉屁股——开了眼。

打开Google Bard机器人问问这是啥原因导致的：

可能是网络原因导致的，查了下其他博客说也有可能是开启了某种认证机制导致的。

通过ping发现能ping通broker，但是通过telnet却无法连接了，也就是说：这台测试服务器与kafka broker间的网络被限制为可以ping但不能访问broker的端口号！

解决办法

查看了下测试服务器的iptables链没发现问题，自己通过tcpdump抓包用wireshark分析发现只有SYN包没响应，找到客户网络工程师定位到是公司的网络策略限制了。至此问题解决，希望能给读者一个思路。