This is the home of the Myriad space.

Apache Myriad enables the co-existence of Apache Hadoop and Apache Mesos on the physical infrastructure. By running Hadoop YARN as a Mesos framework, YARN applications and Mesos frameworks can run side-by-side, dynamically sharing cluster resources. 

With Apache Myriad, you can:

  • Run your operational applications (including those running in Docker) side-by-side with your analytic applications.
  • Achieve Hadoop multi-tenancy by provisioning logical Hadoop clusters for each user or group.

Key features include:

  • YARN running as a Mesos Framework, with resource manager and node managers running inside Mesos containers.
  • YARN clusters running on Mesos that can allocate resources in one of the following ways:
    • Static - Administrators can use an API or a GUI to add or remove node managers or auxiliary services like the JobHistoryServer.
    • Fine-grained - Administrators can provision thin node managers that are dynamically resized based on application demand.
  • High Availability (HA) and graceful restart of YARN daemons.
  • Ability to launch multiple YARN clusters on the same set of nodes.
  • Support for YARN FairScheduler and all functionality such as hierarchical queues with weights. 
  • Ability to deploy YARN Resource Manager using Marathon. This feature leverages Marathon's dynamic scheduling, process supervision, and integration with service discovery (Mesos-DNS). 
  • Ability to run MapReduce v2 and associated libraries such as Hive, Pig, and Mahout.

Topics

 

https://wiki.apache.org/incubator/MyriadProposal

Incubator Status: https://incubator.apache.org/projects/myriad.html

Source: https://github.com/apache/incubator-myriad

Issue tracking: https://issues.apache.org/jira/browse/MYRIAD

Patch Submission: https://github.com/apache/incubator-myriad/pulls

Mailing list: http://mail-archives.apache.org/mod_mbox/incubator-myriad-dev/

  • No labels