Hi James, That's not an unresponsive node that's killing Kafka, that's a failure to resolve the address that's killing it - my personal expectation would be that even though zookeeper-2.zookeeper.etc may be down, its name should still resolve.
Regards, Liam Clarke On Wed, Sep 26, 2018 at 10:02 PM James Yu <cyu...@gmail.com> wrote: > Hi, > > I fail to start kafka broker when the corresponding zookeeper cluster is in > working but unhealthy state. > > The zookeeper cluster is made of 3 nodes: zookeeper-0, zookeeper-1, > zookeeper-2. > > I put all 3 zookeeper nodes into kafka's server.properties, specifically > for zookeeper.connect attribute as as: > > zookeeper.connect=zookeeper-0.zookeeper.default.svc.cluster.local:2181,zookeeper-1.zookeeper.default.svc.cluster.local:2181,zookeeper-2.zookeeper.default.svc.cluster.local:2181 > > I intentionally terminate zookeeper-2 and let zookeeper cluster work with > only two nodes (zookeeper-0, zookeeper-1). Then I start start kafka and get > the following error message: > [2018-09-26 09:57:23,841] INFO Initiating client connection, > > connectString=zookeeper-0.zookeeper.default.svc.cluster.local:2181,zookeeper-1.zookeeper.default.svc.cluster.local:2181,zookeeper-2.zookeeper.default.svc.cluster.local:2181 > sessionTimeout=6000 > watcher=kafka.zookeeper.ZooKeeperClient$ZooKeeperClientWatcher$@49ec71f8 > (org.apache.zookeeper.ZooKeeper) > [2018-09-26 09:57:24,100] ERROR Fatal error during KafkaServer startup. > Prepare to shutdown (kafka.server.KafkaServer) > java.net.UnknownHostException: > zookeeper-2.zookeeper.default.svc.cluster.local: Name or service not known > at java.net.Inet6AddressImpl.lookupAllHostAddr(Native Method) > at java.net.InetAddress$2.lookupAllHostAddr(InetAddress.java:928) > at > java.net.InetAddress.getAddressesFromNameService(InetAddress.java:1323) > at java.net.InetAddress.getAllByName0(InetAddress.java:1276) > at java.net.InetAddress.getAllByName(InetAddress.java:1192) > at java.net.InetAddress.getAllByName(InetAddress.java:1126) > at > > org.apache.zookeeper.client.StaticHostProvider.<init>(StaticHostProvider.java:61) > at org.apache.zookeeper.ZooKeeper.<init>(ZooKeeper.java:445) > at org.apache.zookeeper.ZooKeeper.<init>(ZooKeeper.java:380) > at kafka.zookeeper.ZooKeeperClient.<init>(ZooKeeperClient.scala:86) > at kafka.zk.KafkaZkClient$.apply(KafkaZkClient.scala:1538) > at > > kafka.server.KafkaServer.kafka$server$KafkaServer$$createZkClient$1(KafkaServer.scala:345) > at kafka.server.KafkaServer.initZkClient(KafkaServer.scala:369) > at kafka.server.KafkaServer.startup(KafkaServer.scala:202) > at > kafka.server.KafkaServerStartable.startup(KafkaServerStartable.scala:38) > at kafka.Kafka$.main(Kafka.scala:92) > at kafka.Kafka.main(Kafka.scala) > [2018-09-26 09:57:24,104] INFO shutting down (kafka.server.KafkaServer) > [2018-09-26 09:57:24,107] WARN (kafka.utils.CoreUtils$) > java.lang.NullPointerException > at > > kafka.server.KafkaServer$$anonfun$shutdown$5.apply$mcV$sp(KafkaServer.scala:569) > at kafka.utils.CoreUtils$.swallow(CoreUtils.scala:85) > at kafka.server.KafkaServer.shutdown(KafkaServer.scala:569) > at kafka.server.KafkaServer.startup(KafkaServer.scala:326) > at > kafka.server.KafkaServerStartable.startup(KafkaServerStartable.scala:38) > at kafka.Kafka$.main(Kafka.scala:92) > at kafka.Kafka.main(Kafka.scala) > [2018-09-26 09:57:24,112] INFO shut down completed > (kafka.server.KafkaServer) > [2018-09-26 09:57:24,112] ERROR Exiting Kafka. > (kafka.server.KafkaServerStartable) > [2018-09-26 09:57:24,115] INFO shutting down (kafka.server.KafkaServer) > > I thought we can start a kafka as long as the zookeeper cluster is > working.regardless there is one zookeeper node (3 in total) fails to run? > > > This is a UTF-8 formatted mail > ----------------------------------------------- > James C.-C.Yu > +886988713275 > +8615618429976 >
