连接到Docker中运行的HBase

Question

I cannot connect to HBase running in Docker on Windows ( banno/hbase-standalone image). However, I can connect to locally installed HBase.

banno/hbase-standalone image is run using:

docker run -d -p 2181:2181 -p 60000:60000 -p 60010:60010 -p 60020:60020 -p 60030:60030 banno/hbase-standalone

I also set up the port forwarding on the boot2docker-vm (which is required when running on Windows):

I can successfully telnet to all those ports on my localhost.

Next, here is a code sample that we use in our tests:

Configuration config = HBaseConfiguration.create();
config.clear();
config.setInt("timeout", 12000);
config.set("zookeeper.znode.parent", "/hbase");
config.set("hbase.zookeeper.quorum", "127.0.0.1");
config.set("hbase.zookeeper.property.clientPort", "2181");
config.set("hbase.master", "127.0.0.1:60000");

final Configuration configuration = HBaseConfiguration.create(config);

JobDefinition.Buildable.dumpProperties(configuration, newArrayList("hbase.*"));

HBaseAdmin.checkHBaseAvailable(config);

Which causes the following exception

Exception in thread "main" org.apache.hadoop.hbase.MasterNotRunningException: com.google.protobuf.ServiceException: java.net.UnknownHostException: unknown host: a3e6c240af20
    at org.apache.hadoop.hbase.client.HConnectionManager$HConnectionImplementation$StubMaker.makeStub(HConnectionManager.java:1651)
    at org.apache.hadoop.hbase.client.HConnectionManager$HConnectionImplementation$MasterServiceStubMaker.makeStub(HConnectionManager.java:1677)
    at org.apache.hadoop.hbase.client.HConnectionManager$HConnectionImplementation.getKeepAliveMasterService(HConnectionManager.java:1885)
    at org.apache.hadoop.hbase.client.HConnectionManager$HConnectionImplementation.isMasterRunning(HConnectionManager.java:900)
    at org.apache.hadoop.hbase.client.HBaseAdmin.checkHBaseAvailable(HBaseAdmin.java:2366)
    at com.xxx.compute.hadoop.jobs.transaction.OurTest.main(OurTest.java:24)
Caused by: com.google.protobuf.ServiceException: java.net.UnknownHostException: unknown host: a3e6c240af20
    at org.apache.hadoop.hbase.ipc.RpcClient.callBlockingMethod(RpcClient.java:1674)
    at org.apache.hadoop.hbase.ipc.RpcClient$BlockingRpcChannelImplementation.callBlockingMethod(RpcClient.java:1715)
    at org.apache.hadoop.hbase.protobuf.generated.MasterProtos$MasterService$BlockingStub.isMasterRunning(MasterProtos.java:42561)
    at org.apache.hadoop.hbase.client.HConnectionManager$HConnectionImplementation$MasterServiceStubMaker.isMasterRunning(HConnectionManager.java:1688)
    at org.apache.hadoop.hbase.client.HConnectionManager$HConnectionImplementation$StubMaker.makeStubNoRetries(HConnectionManager.java:1597)
    at org.apache.hadoop.hbase.client.HConnectionManager$HConnectionImplementation$StubMaker.makeStub(HConnectionManager.java:1623)
    ... 5 more
Caused by: java.net.UnknownHostException: unknown host: a3e6c240af20
    at org.apache.hadoop.hbase.ipc.RpcClient$Connection.<init>(RpcClient.java:386)
    at org.apache.hadoop.hbase.ipc.RpcClient.createConnection(RpcClient.java:352)
    at org.apache.hadoop.hbase.ipc.RpcClient.getConnection(RpcClient.java:1526)
    at org.apache.hadoop.hbase.ipc.RpcClient.call(RpcClient.java:1438)
    at org.apache.hadoop.hbase.ipc.RpcClient.callBlockingMethod(RpcClient.java:1657)
    ... 10 more

This is explainable. We run Windows, which requires boot2docker-vm virtual machine running using NAT. The Docker container of the image is running inside the boot2docker-vm also using NAT. However, the ports are "visible" to the host machine running tests, since Docker container exports the ports, and the boot2docker-vm forwards the ports the host machine. The name a3e6c240af20 actually comes from the Docker container ID, so probably a3e6c240af20 is a hostname for the Docker container :

CONTAINER ID        IMAGE                           COMMAND                CREATED             STATUS              PORTS                                                                                                                            NAMES
a3e6c240af20        banno/hbase-standalone:latest   "/bin/sh -c '/opt/hb   24 minutes ago      Up 24 minutes       0.0.0.0:2181->2181/tcp, 0.0.0.0:60000->60000/tcp, 0.0.0.0:60010->60010/tcp, 0.0.0.0:60020->60020/tcp, 0.0.0.0:60030->60030/tcp   agitated_wozniak

I am not sure how exactly HBase communication works, but apparently it makes RPC calls to the instance. HBase Docker returns its hostname hoping that the client will call it there. But since both boot2docker-vm and Docker container running using NAT, the host machine does not see the Docker container.

I tried to add a3e6c240af20 to my hosts file:

127.0.0.1 a3e6c240af20

Then I get a different error, also during the RPC call, which actually does not help me much:

Exception in thread "main" org.apache.hadoop.hbase.MasterNotRunningException: com.google.protobuf.ServiceException: java.lang.NullPointerException
    at org.apache.hadoop.hbase.client.HConnectionManager$HConnectionImplementation$StubMaker.makeStub(HConnectionManager.java:1651)
    at org.apache.hadoop.hbase.client.HConnectionManager$HConnectionImplementation$MasterServiceStubMaker.makeStub(HConnectionManager.java:1677)
    at org.apache.hadoop.hbase.client.HConnectionManager$HConnectionImplementation.getKeepAliveMasterService(HConnectionManager.java:1885)
    at org.apache.hadoop.hbase.client.HConnectionManager$HConnectionImplementation.isMasterRunning(HConnectionManager.java:900)
    at org.apache.hadoop.hbase.client.HBaseAdmin.checkHBaseAvailable(HBaseAdmin.java:2366)
    at com.xxx.compute.hadoop.jobs.transaction.OurTest.main(OurTest.java:24)
Caused by: com.google.protobuf.ServiceException: java.lang.NullPointerException
    at org.apache.hadoop.hbase.ipc.RpcClient.callBlockingMethod(RpcClient.java:1674)
    at org.apache.hadoop.hbase.ipc.RpcClient$BlockingRpcChannelImplementation.callBlockingMethod(RpcClient.java:1715)
    at org.apache.hadoop.hbase.protobuf.generated.MasterProtos$MasterService$BlockingStub.isMasterRunning(MasterProtos.java:42561)
    at org.apache.hadoop.hbase.client.HConnectionManager$HConnectionImplementation$MasterServiceStubMaker.isMasterRunning(HConnectionManager.java:1688)
    at org.apache.hadoop.hbase.client.HConnectionManager$HConnectionImplementation$StubMaker.makeStubNoRetries(HConnectionManager.java:1597)
    at org.apache.hadoop.hbase.client.HConnectionManager$HConnectionImplementation$StubMaker.makeStub(HConnectionManager.java:1623)
    ... 5 more
Caused by: java.lang.NullPointerException
    at org.apache.hadoop.hbase.ipc.RpcClient$Connection.writeRequest(RpcClient.java:1051)
    at org.apache.hadoop.hbase.ipc.RpcClient.call(RpcClient.java:1440)
    at org.apache.hadoop.hbase.ipc.RpcClient.callBlockingMethod(RpcClient.java:1657)
    ... 10 more

Does anyone have a suggestion how this can be solved?

Answer 1

Try add [boot2docker IP] a3e6c240af20 instead of 127.0.0.1 because HBase Java client needs to reach your docker's host not exactly localhost to reach zookeeper (CMIIW). Not pretty sure if it will works but it works in my Windows.

Answer 2

I used oddpoet/hbase-cdh5 docker image to avoid this issue.

docker run -d -p 2181:2181 -p 60000:60000 -p 60010:60010 -p 60020:60020 -p 60030:60030 -h hbase oddpoet/hbase-cdh5

fig.yml

hbase:
    image: oddpoet/hbase-cdh5
    hostname: hbase
    ports:
     - "3181:2181"
     - "60000:60000"
     - "60010:60010"
     - "60020:60020"
     - "60030:60030"

my configuration file

    conf.set("hbase.zookeeper.quorum", zkPath);
    conf.set("hbase.zookeeper.property.clientPort","2181");
    conf.set("zookeeper.znode.parent", "/hbase");

    conf.set("hbase.client.retries.number", "3");  // default 35
    conf.set("hbase.rpc.timeout", "10000");  // default 60 secs
    conf.set("hbase.rpc.shortoperation.timeout", "5000"); // default 10 secs