Lily Hbase Indexers 无故退出答案

【问题标题】：Lily Hbase Indexers exit for no reasonLily Hbase Indexers 无故退出
【发布时间】：2014-09-10 06:19:48
【问题描述】：

我正在运行 Cloudera/Solr 集群，并尝试使用 hbase-solr (Lily) 索引器将 Hbase 的 NRT 索引到 Solr。批处理模式索引工作正常。

但是，在我开始以恒定流的方式加载数据后，Lily 索引器开始一个接一个地死机。他们不会打印出跳出我的特定错误消息，但都以相同的方式结束：

2014-09-10 16:04:56,770 INFO org.apache.hadoop.hbase.zookeeper.RecoverableZooKeeper: Process identifier=ip-172-31-1-204.ap-southeast-2.compute.internal,44013,1410329096767 connecting to ZooKeeper ensemble=ip-172-31-1-205.ap-southeast-2.compute.internal:2181,ip-172-31-1-206.ap-southeast-2.compute.internal:2181,ip-172-31-1-204.ap-southeast-2.compute.internal:2181
2014-09-10 16:04:56,771 INFO org.apache.zookeeper.ClientCnxn: Opening socket connection to server ip-172-31-1-206.ap-southeast-2.compute.internal/172.31.1.206:2181. Will not attempt to authenticate using SASL (unknown error)
2014-09-10 16:04:56,772 INFO org.apache.hadoop.ipc.RpcServer: RpcServer.listener,port=44013: starting
2014-09-10 16:04:56,771 INFO org.apache.hadoop.ipc.RpcServer: RpcServer.responder: starting
2014-09-10 16:04:56,773 INFO org.apache.zookeeper.ClientCnxn: Socket connection established to ip-172-31-1-206.ap-southeast-2.compute.internal/172.31.1.206:2181, initiating session
2014-09-10 16:04:56,775 INFO com.ngdata.hbaseindexer.supervisor.IndexerSupervisor: Started indexer for indexFeature
2014-09-10 16:04:56,776 INFO org.apache.zookeeper.ClientCnxn: Session establishment complete on server ip-172-31-1-206.ap-southeast-2.compute.internal/172.31.1.206:2181, sessionid = 0x1485c7ff13602fd, negotiated timeout = 60000
2014-09-10 16:04:56,813 INFO org.kitesdk.morphline.api.MorphlineContext: Importing commands
2014-09-10 16:04:57,287 INFO org.kitesdk.morphline.api.MorphlineContext: Done importing commands
2014-09-10 16:04:57,289 INFO org.apache.solr.client.solrj.impl.HttpClientUtil: Creating new http client, config:
2014-09-10 16:04:57,297 INFO org.apache.hadoop.ipc.RpcServer: regionserver/ip-172-31-1-204.ap-southeast-2.compute.internal/172.31.1.204:0: started 10 reader(s).
2014-09-10 16:04:57,299 INFO org.apache.zookeeper.ZooKeeper: Initiating client connection, connectString=ip-172-31-1-205.ap-southeast-2.compute.internal:2181,ip-172-31-1-206.ap-southeast-2.compute.internal:2181,ip-172-31-1-204.ap-southeast-2.compute.internal:2181 sessionTimeout=60000 watcher=ip-172-31-1-204.ap-southeast-2.compute.internal,44713,1410329097297, quorum=ip-172-31-1-205.ap-southeast-2.compute.internal:2181,ip-172-31-1-206.ap-southeast-2.compute.internal:2181,ip-172-31-1-204.ap-southeast-2.compute.internal:2181, baseZNode=/hbase
2014-09-10 16:04:57,301 INFO org.apache.hadoop.hbase.zookeeper.RecoverableZooKeeper: Process identifier=ip-172-31-1-204.ap-southeast-2.compute.internal,44713,1410329097297 connecting to ZooKeeper ensemble=ip-172-31-1-205.ap-southeast-2.compute.internal:2181,ip-172-31-1-206.ap-southeast-2.compute.internal:2181,ip-172-31-1-204.ap-southeast-2.compute.internal:2181
2014-09-10 16:04:57,302 INFO org.apache.zookeeper.ClientCnxn: Opening socket connection to server ip-172-31-1-204.ap-southeast-2.compute.internal/172.31.1.204:2181. Will not attempt to authenticate using SASL (unknown error)
2014-09-10 16:04:57,303 INFO org.apache.hadoop.ipc.RpcServer: RpcServer.responder: starting
2014-09-10 16:04:57,303 INFO org.apache.zookeeper.ClientCnxn: Socket connection established to ip-172-31-1-204.ap-southeast-2.compute.internal/172.31.1.204:2181, initiating session
2014-09-10 16:04:57,304 INFO org.apache.hadoop.ipc.RpcServer: RpcServer.listener,port=44713: starting
2014-09-10 16:04:57,306 INFO com.ngdata.hbaseindexer.supervisor.IndexerSupervisor: Started indexer for indexSeenBlock
2014-09-10 16:04:57,307 INFO org.apache.zookeeper.ClientCnxn: Session establishment complete on server ip-172-31-1-204.ap-southeast-2.compute.internal/172.31.1.204:2181, sessionid = 0x3485c7fee8f0374, negotiated timeout = 60000
2014-09-10 16:04:57,349 INFO org.mortbay.log: Logging to org.slf4j.impl.Log4jLoggerAdapter(org.mortbay.log) via org.mortbay.log.Slf4jLog
2014-09-10 16:04:57,536 INFO org.mortbay.log: jetty-6.1.26.cloudera.2
2014-09-10 16:04:58,663 INFO org.mortbay.log: Started SelectChannelConnector@0.0.0.0:11060
2014-09-10 16:05:01,591 INFO org.kitesdk.morphline.api.MorphlineContext: Importing commands
2014-09-10 16:05:01,597 INFO org.kitesdk.morphline.api.MorphlineContext: Importing commands
2014-09-10 16:05:01,641 INFO org.kitesdk.morphline.api.MorphlineContext: Importing commands
2014-09-10 16:05:01,650 INFO org.kitesdk.morphline.api.MorphlineContext: Importing commands
2014-09-10 16:05:01,688 INFO org.kitesdk.morphline.api.MorphlineContext: Importing commands
2014-09-10 16:05:01,726 INFO org.kitesdk.morphline.api.MorphlineContext: Importing commands
2014-09-10 16:05:01,732 INFO org.kitesdk.morphline.api.MorphlineContext: Importing commands
2014-09-10 16:05:01,740 INFO org.kitesdk.morphline.api.MorphlineContext: Importing commands
2014-09-10 16:05:01,752 INFO org.kitesdk.morphline.api.MorphlineContext: Importing commands

除了说进程已经退出之外，Cloudera 管理器也没有提供任何有用的信息。 Solr 索引中的某些记录确实得到了更新，这表明索引器至少在一段时间内正常工作。

我在 RHEL6.5 和 JDK7 上运行最新的 CDH 5.1。

【问题讨论】：

嗨，你能解决这个问题吗？解决方法是什么？就我而言，自从我关闭了 WAL 以来，我的 cmets 下面就对我有所帮助。当我打开然后它开始工作。想知道其他情况是什么。请回复
@RamPrasadG 我建议不要使用 Hbase，尤其是 Lily。如果您要将数据放入 Solr，请尝试使用直接 Solr JSON API。如果您想要比 HBase 更好的数据库，请尝试 Kudu（测试版）。

标签： hadoop solr cloudera cloudera-cdh

【解决方案1】：

如果您使用 WAL，即在您的 hbase 客户端程序（如 mapreduce）中写入头日志选项为 FALSE（出于性能原因）。它不会写入 HLOG。

NRT 适用于 WAL/HLOG。

批处理索引工作正常，因为它在 WAL 上不起作用

【讨论】：