- Release Notes
- Behavior Changes
- Known Issues
- System Information
- HDP Services Started Automatically on Startup
- HDP Services Not Started Automatically on Startup
- Further Reading
The VirtualBox and VMWare releases of the HDP sandbox had their OS-level packages updated, including a version update to Docker 18.03.0.
Image MD5 Checksums:
- VirtualBox – 533130d23a2f7d919559d0fbb858b09f
- VMware – 781bd8f2b55660c57e68b90c0af158d3
- Docker – 5d828411c30c9db0fd663396468e20c6
HDP Stack and Ambari: The Sandbox uses the following versions of Ambari and HDP stack. Please use the following release note links provided to view Ambari and HDP stack specific information.
- Virtual machines are now Connected Data Architecture (CDA) ready
- Kafka is on by default
- Builds now include Superset and Druid, both turned off
- Zeppelin interface on port 9995 may require clearing browser cache on some browsers
This is a list of common limitations along with their workarounds.
- RMP-3586 – Due to dependency of the underlying OS and Virtual machine application, the following may occur when suspending the virtual machine:
- Region Server service for HBase may be stopped when returning back from suspended state. It will need to be restarted.
- Ambari Metrics may be stopped when returning back from suspended state since it now uses an embedded HBase.
- Workaround: Avoid having to suspend your virtual machine.
Operating System and Java versions that the Sandbox has installed.
- OS Version (docker container)
- CentOS release 7.5.1804 (Core)
- Java Version (docker container)
- openJDK version “188.8.131.52”
- OpenJDK Runtime Environment (build 184.108.40.206-8.b10)
- OpenJDK 64-Bit Server VM (build 25.161-b14, mixed mode)
- Updated from previous version
- OS Version (Hosting Virtual Machine)
- CentOS Linux release 7.2.1511 (Core)
Image File Sizes:
- VMware – 16 GB
- VirtualBox – 16 GB
- Docker – 15 GB
These are a list of databases used within Sandbox along with the corresponding HDP components that use them.
- Ambari: Postgres
- Hive Metastore : Mysql
- Ranger: Mysql
- Oozie: derby (embedded)
HDP Supported Components Not Installed
These components are offered by the Hortonworks distribution, but not included in the Sandbox.
- Apache Accumulo
- Apache Mahout
Newly Added HDP Supported Packages
HDP Services Started Automatically on Startup
When the virtual machine is booted up, the following services are started. If not specified, assume all are java processes. The users that launch the process are the corresponding names of the component. The processes are listed with their main class.
- AmbariServer – org.apache.ambari.server.controller.AmbariServer run as root user
- Ambari Agent (non java process)
- Application – org.apache.flume.node.Application
- Portmap – org.apache.hadoop.portmap.Portmap
- NameNode – org.apache.hadoop.hdfs.server.namenode.NameNode
- DataNode – org.apache.hadoop.hdfs.server.datanode.DataNode
- Portmap – Unlike the other processes that are launched by HDFS user, these are run as root user.
- The nfs process doesn’t show up as a name for jps output
- RunJar – webhcat – org.apache.hadoop.util.RunJar Run as hcat user
- RunJar – metastore – org.apache.hadoop.util.RunJar
- RunJar – hiveserver2 – org.apache.hadoop.util.RunJar
- JobHistoryServer – org.apache.hadoop.mapreduce.v2.hs.JobHistoryServer
- mapred is the user used to launch this process
- Bootstrap – org.apache.catalina.startup.Bootstrap
- Kafka – kafka.Kafka
- UnixAuthenticationService – org.apache.ranger.authentication.UnixAuthenticationService Run as root user
- EmbededServer – org.apache.ranger.server.tomcat.EmbeddedServer
- HistoryServer – org.apache.spark.deploy.history.HistoryServer
- Livy server run as livy
- Thrift server – org.apache.spark.deploy.SparkSubmit run as hive user
- ApplicationHistoryServer – org.apache.hadoop.yarn.server.applicationhistoryservice.ApplicationHistoryServer
- ResourceManager – org.apache.hadoop.yarn.server.resourcemanager.ResourceManager
- NodeManager – org.apache.hadoop.yarn.server.nodemanager.NodeManager
- QuorumPeerMain – org.apache.zookeeper.server.quorum.QuorumPeerMain
- ZeppelinServer – org.apache.zeppelin.server.ZeppelinServer
HDP Services Not Started Automatically on Startup
Because of the limited resources available in the sandbox virtual machine environment, the following services are in maintenance mode and will not automatically start. To fully use these services, you must allocate more memory to the sandbox virtual machine. If you want these services to automatically start, turn off maintenance mode. The processes are listed with their main class.
- Ambari Infra
- Ambari Metrics
- Main – org.apache.atlas.Main
- SecondaryNameNode – org.apache.hadoop.hdfs.server.namenode.SecondaryNameNode
- Since on a single node, secondary namenode is not needed, it is not started.
- Main – org.apache.falcon.Main
- HRegionServer – org.apache.hadoop.hbase.regionserver.HRegionServer
- HMaster – org.apache.hadoop.hbase.master.HMaster
- gateway.jar – /usr/hdp/current/knox-server/bin/gateway.jar
- ldap.jar – /usr/hdp/current/knox-server/bin/ldap.jar This process is a mini ldap server
- supervisor – backtype.storm.daemon.supervisor
- nimbus – backtype.storm.daemon.nimbus
- logviewer – backtype.storm.daemon.logviewer
- core – backtype.storm.ui.core
- drpc – backtype.storm.daemon.drpc