The parts of a Hadoop cluster are typically used for multiple ports, some for interaction between daemon, and some for RPC access and HTTP access. And with the increase of the peripheral components of Hadoop, it is completely unable to remember which port corresponds to which application, especially collect records so as to query. This includes the components we use: HDFS, YARN, HBase, Hive, ZooKeeper:
Component |
Node |
Default port |
Configuration |
Description of Use |
Hdfs |
DataNode |
50010 |
Dfs.datanode.address |
Datanode service port for data transfer |
Hdfs |
DataNode |
50075 |
Dfs.datanode.http.address |
Port for HTTP Service |
Hdfs |
DataNode |
50475 |
Dfs.datanode.https.address |
Ports for HTTPS services |
Hdfs |
DataNode |
50020 |
Dfs.datanode.ipc.address |
Ports for IPC Services |
Hdfs |
NameNode |
50070 |
Dfs.namenode.http-address |
Port for HTTP Service |
Hdfs |
NameNode |
50470 |
Dfs.namenode.https-address |
Ports for HTTPS services |
Hdfs |
NameNode |
8020 |
Fs.defaultfs |
The RPC port that receives the client connection for obtaining file system metadata information. |
Hdfs |
Journalnode |
8485 |
Dfs.journalnode.rpc-address |
RPC Service |
Hdfs |
Journalnode |
8480 |
Dfs.journalnode.http-address |
HTTP Service |
Hdfs |
Zkfc |
8019 |
Dfs.ha.zkfc.port |
ZooKeeper Failovercontroller, for nn HA |
YARN |
ResourceManager |
8032 |
Yarn.resourcemanager.address |
RM Applications Manager (ASM) port |
YARN |
ResourceManager |
8030 |
Yarn.resourcemanager.scheduler.address |
IPC Port for Scheduler components |
YARN |
ResourceManager |
8031 |
Yarn.resourcemanager.resource-tracker.address |
Ipc |
YARN |
ResourceManager |
8033 |
Yarn.resourcemanager.admin.address |
Ipc |
YARN |
ResourceManager |
8088 |
Yarn.resourcemanager.webapp.address |
HTTP Service Port |
YARN |
NodeManager |
8040 |
Yarn.nodemanager.localizer.address |
Localizer IPC |
YARN |
NodeManager |
8042 |
Yarn.nodemanager.webapp.address |
HTTP Service Port |
YARN |
NodeManager |
8041 |
Yarn.nodemanager.address |
Port of container Manager in NM |
YARN |
Jobhistory Server |
10020 |
Mapreduce.jobhistory.address |
Ipc |
YARN |
Jobhistory Server |
19888 |
Mapreduce.jobhistory.webapp.address |
HTTP Service Port |
HBase |
Master |
60000 |
Hbase.master.port |
Ipc |
HBase |
Master |
60010 |
Hbase.master.info.port |
HTTP Service Port |
HBase |
Regionserver |
60020 |
Hbase.regionserver.port |
Ipc |
HBase |
Regionserver |
60030 |
Hbase.regionserver.info.port |
HTTP Service Port |
HBase |
Hquorumpeer |
2181 |
Hbase.zookeeper.property.clientPort |
Hbase-managed ZK mode, using a standalone zookeeper cluster will not enable the port. |
HBase |
Hquorumpeer |
2888 |
Hbase.zookeeper.peerport |
Hbase-managed ZK mode, using a standalone zookeeper cluster will not enable the port. |
HBase |
Hquorumpeer |
3888 |
Hbase.zookeeper.leaderport |
Hbase-managed ZK mode, using a standalone zookeeper cluster will not enable the port. |
Hive |
Metastore |
9083 |
/etc/default/hive-metastore export port=<port> to update the default port |
|
Hive |
Hiveserver |
10000 |
/ETC/HIVE/CONF/HIVE-ENV.SH export hive_server2_thrift_port=<port> to update the default port |
|
ZooKeeper |
Server |
2181 |
/etc/zookeeper/conf/zoo.cfg in Clientport=<port> |
Ports that serve clients |
ZooKeeper |
Server |
2888 |
/etc/zookeeper/conf/zoo.cfg in server.x=[hostname]:nnnnn[:nnnnn], blue part |
The follower is used to connect to the leader and only listens on the leader on the port. |
ZooKeeper |
Server |
3888 |
/etc/zookeeper/conf/zoo.cfg in server.x=[hostname]:nnnnn[:nnnnn], blue part |
Used for the leader election. Required only if ELECTIONALG is 3 (default). |
All port protocols are based on TCP. For all Hadoop daemon that exist with the Web UI (HTTP service), there are the following url:/logs
List of log files for downloading and viewing/loglevel
Allows you to set the log4j logging level, similar to Hadoop daemonlog/stacks
stack trace for all threads, which is helpful for debug/JMX
Server-side metrics, output in JSON format. /jmx?qry=hadoop:* will return all Hadoop-related metrics.
The/jmx?get=mxbeanname::attributename query specifies the value of the specified bean property, such as/jmx?get=hadoop:service=namenode,name=namenodeinfo:: Clusterid will return to Clusterid.
The processing class for this request: Org.apache.hadoop.jmx.JMXJsonServlet and the specific daemon has a specific URL path specific corresponding information. namenode:http://:50070//dfshealth.jsp
HDFs Information page, which has links to view file system/dfsnodelist.jsp?whatnodes= (dead| LIVE)
Show dead or live status of Datanode/fsck
Running the fsck command is not recommended to use when the cluster is busy! Datanode:http://:50075//blockscannerreport
Each datanode will specify interval validation block information
A summary of common port usage configurations for BI eco-circle