Skip to content

Latest commit

 

History

History
164 lines (123 loc) · 7.94 KB

release-note-3.2.md

File metadata and controls

164 lines (123 loc) · 7.94 KB

PNDA 3.2

PNDA release 3.2 contains a number of improvements and bug fixes.

Flavors

The main focus of this release is a major refactoring of how PNDA handles different profiles for different infrastructure - "flavors". Please refer to the pnda-aws-templates and pnda-heat-templates repositories for more details.

We've added a new flavor to demonstrate the new organization, named "pico". This is a very small profile suitable for evaluating and learning about PNDA on minimal infrastructure.

Example Applications

We've consolidated the example applications into one repository, example-applications.

We've also added some new examples - including a package that includes both a simple batch job and an accompanying Jupyter notebook, and a demonstration of how to launch and run the popular H2O data science platform.

Upgraded Technologies

In order to benefit from upstream bug fixes the version of Salt used has been moved to 2015.8.11.

Version Matrix

Repository Version Date
example-applications 0.1.0 21 Oct 2016
example-kafka-clients 0.2.0 07 Sep 2016
gobblin 0.1.1 13 Sep 2016
platform-console-backend 0.2.1 21 Oct 2016
platform-console-frontend 0.1.2 21 Oct 2016
platform-data-mgmnt 0.1.1 13 Sep 2016
platform-deployment-manager 0.2.0 21 Oct 2016
platform-libraries 0.1.1 09 Sep 2016
platform-package-repository 0.2.0 21 Oct 2016
platform-salt 1.0.0 21 Oct 2016
platform-salt-cloud 1.0.0 21 Oct 2016
platform-testing 0.1.1 13 Sep 2016
platform-tools 0.1.0 01 Jul 2016
pnda-aws-templates 1.0.0 21 Oct 2016
pnda-dib-elements 0.1.1 21 Oct 2016
pnda-guide 0.1.3 21 Oct 2016
pnda-heat-templates 1.0.0 21 Oct 2016
pnda-package-server-docker 0.1.1 21 Oct 2016
prod-logstash-codec-avro 0.2.0 09 Sep 2016

Change Log

Added
  • h2o-launcher application to run h2o data science platform
  • literary-word-count-app to run a classic wordcount
  • Spark streaming example app that consumes from Kafka and writes to HBase
  • Spark batch example app that consumes Gobblin produced Avro datasets from HDFS and produces Parquet for use with Impala
  • Spark streaming example app that consumes from Kafka and writes to OpenTSDB
  • Jupyter notebook example that shows some simple network data manipulation
  • Unchanged in this PNDA release
  • Unchanged in this PNDA release
Fixed
  • PNDA-2120: Rotate console backend log files to prevent disk filling up
Changes
  • PNDA-2272: Only show links to parts of the system which are present to allow modularisation of PNDA
  • Unchanged in this PNDA release
Added
  • PNDA-2233 Jupyter notebook plugin added to deployment manager
  • Unchanged in this PNDA release
Added
  • PNDA-1211: add FS repository as backend storage
Added
  • Multi-flavor mechanism, with pico flavor
  • PNDA-2320 Kafka manager port is now in pillar
  • PNDA-2272 review formulas in order to ensure no issue on deployment even if there is not all roles
  • PNDA-2233 Jupyter notebook plugin added to deployment manager
Fixed
  • Some logs were not in /var/log/pnda and so, were not shipped to the logserver
  • The 'bulk' directory in HDFS is now owned by the 'pnda' user
  • Prevent Gobblin from ingesting internal kafka __consumer_offsets topic
Changed
  • ntp:servers pillar default was removed, but can still be set
Changed
  • PNDA-2114: Updated role grains to the new set matching platform-salt
  • Install salt minions version 2015.8.11
  • Unchanged in this PNDA release
  • Unchanged in this PNDA release
Added
  • PNDA-2113: Pico flavor for small footprint clusters suitable for development or learning
  • PNDA-2266: Save provisioning logs on the cloud instances in the /home/ubuntu folder
  • PNDA-2275: '-b' branch CLI parameter to select a branch of platform-salt to use
  • PNDA-2276: Allow a local copy of platform-salt to be used instead of a git URL to clone
  • PNDA-2159: Additional .run log file saved in cli/logs
Changed
  • PNDA-2325: Use single pnda_env.yaml file instead of client_env.sh and pnda_env.sh
  • PNDA-2272: Move to Salt 2015.8.11 in order to get the fix on orchestrate #33467
Fixed
  • PNDA-2333: Allow multiple clusters to be created in parallel from the same clone of pnda-aws-templates
Added
  • Instructions to build a PNDA image from CentOS
  • TripleO and heat-templates repositories as sub-modules
  • A new environment variable to specify an alternate Ubuntu mirror
  • A new element to fix network interfaces naming
Fixed
  • sed command to update Cloud-init configuration file
Changed
  • Clarifications and improvements to many areas for 3.2 release
Changed
  • PNDA-2272: move to Salt 2015.8.11 in order to get the fix on orchestrate #33467
  • PNDA-1211: Add the ability to manage swift / s3 / volume / sshfs or local for package repository backend storage
  • PNDA-2248: Refactor heat template provisioning to enable flavors to be deployable on bare metal - firstly 'pico'
  • The optional 'NtpServers' option can be set in the pnda_env.yaml configuration file if the default servers are not reachable
Changed
  • PNDA-2326: launch the build script automatically and make the jdk available through http
  • Unchanged in this PNDA release