Hbase排錯：Hmaster就是無法啓動起來

hbase報錯

執行命令：

start-hbase.sh

其它節點的HRegionServer夠全部啓動， 就是 Hmaster啓動不來。
日誌查看：

Caused by: java.net.UnknownHostException
        ... 32 more
2018-01-12 14:32:38,385 INFO  [geeke3:16000.activeMasterManager] regionserver.HRegionServer: STOPPED: Unhandled exception. Starting shutdown.
2018-01-12 14:32:41,389 INFO  [master/geeke3/192.168.1.123:16000] ipc.RpcServer: Stopping server on 16000
2018-01-12 14:32:41,391 INFO  [RpcServer.listener,port=16000] ipc.RpcServer: RpcServer.listener,port=16000: stopping
2018-01-12 14:32:41,394 INFO  [RpcServer.responder] ipc.RpcServer: RpcServer.responder: stopped
2018-01-12 14:32:41,394 INFO  [RpcServer.responder] ipc.RpcServer: RpcServer.responder: stopping
2018-01-12 14:32:41,397 INFO  [master/geeke3/192.168.1.123:16000] regionserver.HRegionServer: Stopping infoServer
2018-01-12 14:32:41,408 INFO  [master/geeke3/192.168.1.123:16000] mortbay.log: Stopped SelectChannelConnector@0.0.0.0:16010
2018-01-12 14:32:41,410 INFO  [master/geeke3/192.168.1.123:16000] regionserver.HRegionServer: stopping server geeke3,16000,1515738750238
2018-01-12 14:32:41,410 INFO  [master/geeke3/192.168.1.123:16000] client.ConnectionManager$HConnectionImplementation: Closing zookeeper sessionid=0x260e8e609440005
2018-01-12 14:32:41,414 INFO  [master/geeke3/192.168.1.123:16000-EventThread] zookeeper.ClientCnxn: EventThread shut down
2018-01-12 14:32:41,415 INFO  [master/geeke3/192.168.1.123:16000] zookeeper.ZooKeeper: Session: 0x260e8e609440005 closed
2018-01-12 14:32:41,418 INFO  [master/geeke3/192.168.1.123:16000] regionserver.HRegionServer: stopping server geeke3,16000,1515738750238; all regions closed.
2018-01-12 14:32:41,419 INFO  [master/geeke3/192.168.1.123:16000] hbase.ChoreService: Chore service for: geeke3,16000,1515738750238 had [] on shutdown
2018-01-12 14:32:41,435 INFO  [master/geeke3/192.168.1.123:16000] ipc.RpcServer: Stopping server on 16000
2018-01-12 14:32:41,461 INFO  [master/geeke3/192.168.1.123:16000] zookeeper.ZooKeeper: Session: 0x260e8e609440004 closed
2018-01-12 14:32:41,461 INFO  [master/geeke3/192.168.1.123:16000] regionserver.HRegionServer: stopping server geeke3,16000,1515738750238; zookeeper connection closed.
2018-01-12 14:32:41,462 INFO  [master/geeke3/192.168.1.123:16000] regionserver.HRegionServer: master/geeke3/192.168.1.123:16000 exiting
2018-01-12 14:32:41,462 INFO  [main-EventThread] zookeeper.ClientCnxn: EventThread shut down

問題解決：

冷靜下來，認爲自己以前替換過Hadoop下jar包。
如果回滾到以前，重新來一次呢？

於是做了下面幾步：
1. 刪掉每個節點下的Hbase；
2. 重新解壓開一個新的hbase包；
3. 重新將配置文件寫一遍；
4. 分發到每個節點；
5. 再次啓動hbase。

這次報錯終於不一樣的了， 同時也沒有什麼aws3的相關問題報錯。
看看這個報錯：

// 關鍵一句放出來
hbase/tmp/local/jars, DynamicClassLoader failed to init

想象配置文件上就配置過幾個屬性， 有關“tmp”的就有一個。要不把它註釋掉吧。
於是註釋掉每個hbas配置文件中的屬性：

<property>
<name>hbase.tmp.dir</name>
<value>/opt/software/hbase/tmp</value>
</property>

再次啓動，淚奔，Hbase終於正常，執行命令也正常。