Skip to content

Latest commit

 

History

History
179 lines (100 loc) · 3.71 KB

ReleaseNotes.md

File metadata and controls

179 lines (100 loc) · 3.71 KB
Raw

0.1.24

  • Guard hdfs-format on previous.checkpoint/VERSION This is in addition to the current current/VERSION check.

0.1.23

  • Add hadoop-configure phase to write hadoop config The phase writes all hadoop configuration files.

0.1.22

  • Add daemon specific service phases Adds phases to allow control of individual hadoop services.

    {namenode|jobtracker|tasktracker|datanode}-{start|stop|restart}

0.1.21

  • Fix guard for hdfs format

0.1.20

  • Add service phases for hadoop services Adds :start, :stop, and :restart phases for all hadoop daemons.

0.1.19

  • Enable running hadoop jars as background tasks

0.1.18

  • Fix quoting in hadoop-exec-script

0.1.17

  • Fix hadoop-exec-script when no emv passed

0.1.16

  • Refactor expressions to simplify

  • Allow passing of environment vars to hadoop-exec To allow invocation specific environment variables, allow passing of a map of values from Named items to values. This changes the signature of hadoop-script-exec.

  • Allow string joins on non string property values When passing numeric values for env vars, etc, make sure these handled correctly by turning them into strings.

0.1.15

  • Open tasktracker web port between tasktrackers Seems to be required for reduce to work.

0.1.14

  • Allow task trackers to talk to each other This seems to be necessary to allow the reducer to run.

0.1.13

  • Create more of the required hadoop directories Ensures the hadoop directories that are configured actually exist.

0.1.12

  • Make jar file upload have a stable name This ensures that when a jar is run many times, it is not uploaded each time if it has not changed.

  • Allow datanodes to access each other

0.1.11

  • Allow jobtracker to access tasktracker

0.1.10

  • Fix net-rules for datanodes Opens up more of the required ports.

0.1.9

  • Add net rules to control intra-node ports

0.1.8

  • Make setup-etc-hosts more robust

0.1.7

  • Update pallet and locos versions

  • Write config files with :overwrite-changes true This is a workaround for the tarball install overwriting local configuration files.

  • Simplify hadoop-jar arguments Remove the special cased :main, :input and :output fields.

  • Fix apache hadoop download url

  • Update to latest dependency versions

  • Fix #1 -- Create .bash_profile in ~hadoop.

  • Fix #1: .bash_profile being created in the wrong path

0.1.6

  • fix regression where deploying on vmfest would not create /etc/hosts

  • Update .gitignore to remove generated files.

  • Filter out terminated nodes, it breaks things in AWS.

  • Fix checks on whether to write and use /etc/hosts

  • Prevent installs fail when no md5 is available.

0.1.5

  • Setup repo for releases with lein-pallet-release.

  • Remove release code in prep for lein-pallet-release.

  • Update to pallet 0.8.0-RC.9, vmfest 0.4.0-alpha.1 and locos 0.1.1.

  • Release as EPL (was All Rights Reserved).

  • Update to latest locos snapshot

0.1.4

  • Update for pallet-0.8.0-beta.8

  • Make use of /etc/hosts provider specific

0.1.3

  • Modify hadoop-config-file to not log arguments

0.1.2

  • Reduce potentially sensitive logs to trace

0.1.1

  • Allow for running tests without a compute service

  • Update to pallet 0.8.0-alpha.7

  • Default s3 and s3n credentials Use the compute-service credentials for s3 credentials, if the compute service is ec2 based.

  • Add missing use of debugf

  • Reduce log level on unclassified properties

  • Filter metrics properties correctly

  • Use ips for namenode and jobtracker urls

  • Update cloudera version

  • Move url implementation for :cloudera

  • Fix handling of args in hadoop-jar

  • Add hadoop-role-ports A multi-method to report the required open ports for each role.

0.1.0

Initial release.