- Compatible with Chef 12
- Support disk fix for MapR cluster
- Support Hive 1.0.0 in Bigtop 1.0.0
Enhancement:
- Use postgresql 9.4 for Hive
- Support for Hadoop 2.x in MapR 4.0
New features:
- Add roles and cookbooks for Mesos and Kubernetes
New features:
- Add support for configuration of topology.data
- Add support for configuring Namenode address for HBase Only Cluster
- Add support for configuring Jobtracker or ResourceManager address for Tasktracker/NodeManager Only Cluster
Enhancement:
- Set Namenode listening port to default port 8020
- Each HBase cluster has different hbase.rootdir
- Add retry for installing rpms which have big file size
- Other bug fix
New features:
- Add support for deploying Hortonworks and Apache Bigtop distro via yum
- Add support for deploying Intel 3.x and other Bigtop based Hadop 2.x distro
- Add Chef Error Reporting Handler to save chef-client exception to Chef Nodes
Enhancement:
- Make some check on the role order due to role depencendies
- Other bug fix
New features:
- Upgrade to Chef Client 11
- Enable SSL certificate validation between Chef Server and Chef nodes
- Add support for Intel Hadoop
- Add support for Fedora and Oracle Linux
- Add Support for Multi Network on Hadoop Nodes
Enhancement:
- Use Chef 11 Partial Search API to reduce memory consumption and response time when doing cluster service discovery
- Other bug fix
Enhancement:
- Use non root account to start hadoop service
- Format all attached data disks in parallel instead of one by one to speed up bootstrap
- Reduce unneccessary Chef Search API calls when registering a service to support large scale cluster provision (200+ nodes)
- Other bug fix
New Features:
- Add support for deploying CDH4 Hadoop/YARN/HBase cluster
- Add support for deploying Pivotal HD Hadoop/YARN/HBase cluster
- Add support for deploying MapR HBase cluster
- Add support for deploying on CentOS 6 in addition to CentOS 5
Serveral bug fixes.
New Features:
- Add cookbooks and roles for deploying CDH4(MR1) cluster w/ or w/o HDFS HA and HDFS Federation
- Add cookbooks and roles for deploying MapR cluster.
- Some enhancement and bug fix.
New Features:
- Add Zookeeper and HBase cookbooks/roles for deploying a Zookeeper or HBase cluster
- Enhance service register framework to ensure service is registered after the service daemon is started and ready to handle requests. This can eliminate intermittent synchronization issue between dependent service daemons
- Add support for tuning fair-scheduler.xml, capacity-scheduler.xml and mapred-queue-acls.xml
- Add ulimit settings in hadoop-daemon.sh
- Auto detect JAVA_HOME from /etc/profile instead of hardcode.
- Enable configuration for Hadoop Topology Rack Awareness and Hadoop Virtualization Extensions
Bug Fix:
- When bootstrapping, either service start or service restart should be triggerred, but not both, because restart after start is unnecessary.
- For hadoop daemons: set HADOOP_ROOT_LOGGER to hadoop.root.logger value in log4j.properties.
- Fix issue: hadoop services restart returns zero instead of non-zero when restart failed.
New Features:
- Support for User Specified Hadoop Configuration
- Generate user specified yum repo file in /etc/yum.repo.d/
- Enable WebHDFS REST API by default
- Add cookbook/role for postgresql_server and hive_server
Bug Fix:
- Fix issue: hadoop services restart returns zero instead of non-zero when restart failed
- Make log4j.properties work for hadoop 0.20, 0.23 and 1.x
- Restart hadoop services only when related configuration files are changed
- Initial import for Serengeti 0.5.0
- Contains cookbooks and roles for hadoop 0.20.x, hive, pig, etc.