diff options
| author | Kevin W Monroe <kevin.monroe@canonical.com> | 2016-10-05 17:43:32 +0000 |
|---|---|---|
| committer | Kevin W Monroe <kevin.monroe@canonical.com> | 2016-10-12 09:29:14 -0500 |
| commit | 312c006e227aa42e1ef3bce780a3258ec9b0299b (patch) | |
| tree | 24b8a72f08cb557fe6be207dc8b9d778c20fd0ca /bigtop-packages/src/charm | |
| parent | 89d3ac480ad38cd4006825873eb5534949721ffd (diff) | |
BIGTOP-2548: Refresh charms for Juju 2.0 and Xenial (closes #148)
Signed-off-by: Kevin W Monroe <kevin.monroe@canonical.com>
Diffstat (limited to 'bigtop-packages/src/charm')
29 files changed, 545 insertions, 319 deletions
diff --git a/bigtop-packages/src/charm/README.md b/bigtop-packages/src/charm/README.md index 1290d4b2..04213ad8 100644 --- a/bigtop-packages/src/charm/README.md +++ b/bigtop-packages/src/charm/README.md @@ -18,37 +18,41 @@ ## Overview -These are the charm layers used to build Juju charms for deploying Bigtop -components. The charms are also published to the [Juju charm store][] and -can be deployed directly from there using [bundles][], or they can be -built from these layers and deployed locally. +Juju Charms allow you to deploy, configure, and connect an Apache Bigtop cluster +on any supported cloud, which can be scaled to meet workload demands. You can +also easily connect other, non-Bigtop components from the [Juju charm store][] +that support common interfaces. -Charms allow you to deploy, configure, and connect a Apache Bigtop cluster -on any supported cloud, which can be easily scaled to meet workload demands. -You can also easily connect other, non-Bigtop components from the -[Juju charm store][] that support common interfaces. +This source tree contains the charm layers used to build charms for deploying +Bigtop components. Built charms are published to the [Juju charm store][] +and can be deployed directly from there, either individually or with +[bundles][]. They can also be built from these layers and deployed locally. +For the remainder of this guide, a working Juju installation is assumed to be +present. If Juju is not yet set up, please follow the [getting-started][] +instructions prior to deploying locally built charms and bundles. [Juju charm store]: https://jujucharms.com/ -[bundles]: https://jujucharms.com/u/bigdata-dev/hadoop-processing +[bundles]: https://jujucharms.com/hadoop-processing +[getting-started]: https://jujucharms.com/docs/stable/getting-started ## Building the Bigtop Charms -To build these charms, you will need [charm-tools][]. You should also read -over the developer [Getting Started][] page for an overview of charms and -building them. Then, in any of the charm layer directories, use `charm build`. +To build these charms, you will need [charm-tools][]. You should also read +over the developer [Getting Started][] page for an overview of developing and +building charms. Then, in any of the charm layer directories, use `charm build`. For example: export JUJU_REPOSITORY=$HOME/charms - mkdir $HOME/charms + mkdir $JUJU_REPOSITORY cd bigtop-packages/src/charms/hadoop/layer-hadoop-namenode charm build This will build the NameNode charm, pulling in the appropriate base and interface layers from [interfaces.juju.solutions][]. You can get local copies -of those layers as well using `charm pull-source`: +of those layers as well by using `charm pull-source`: export LAYER_PATH=$HOME/layers export INTERFACE_PATH=$HOME/interfaces @@ -57,19 +61,22 @@ of those layers as well using `charm pull-source`: charm pull-source layer:apache-bigtop-base charm pull-source interface:dfs -You can then deploy the locally built charms individually: +You can deploy the locally built charms individually, for example: - juju deploy local:trusty/hadoop-namenode + juju deploy $JUJU_REPOSITORY/xenial/hadoop-namenode -You can also use the local version of a bundle: +> **Note**: The above assumes Juju 2.0 or greater. If using an earlier version +of Juju, the syntax is: `juju deploy local:xenial/hadoop-namenode`. - juju deploy bigtop-deploy/juju/hadoop-processing/bundle-local.yaml +You can also deploy the local version of a bundle: -> Note: With Juju versions < 2.0, you will need to use [juju-deployer][] to -deploy the local bundle. + juju deploy ./bigtop-deploy/juju/hadoop-processing/bundle-local.yaml +> **Note**: The above assumes Juju 2.0 or greater. If using an earlier version +of Juju, use [juju-quickstart][] with the following syntax: `juju quickstart +./bigtop-deploy/juju/hadoop-processing/bundle-local.yaml`. [charm-tools]: https://jujucharms.com/docs/stable/tools-charm-tools -[Getting Started]: https://jujucharms.com/docs/devel/developer-getting-started +[Getting Started]: https://jujucharms.com/docs/stable/developer-getting-started [interfaces.juju.solutions]: http://interfaces.juju.solutions/ -[juju-deployer]: https://pypi.python.org/pypi/juju-deployer/ +[juju-quickstart]: https://launchpad.net/juju-quickstart diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/README.md b/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/README.md index bf46bf73..621a1e8d 100644 --- a/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/README.md +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/README.md @@ -14,93 +14,119 @@ See the License for the specific language governing permissions and limitations under the License. --> -## Overview +# Overview The Apache Hadoop software library is a framework that allows for the distributed processing of large data sets across clusters of computers using a simple programming model. -This charm deploys the NameNode component of the Apache Bigtop platform +This charm deploys the NameNode component of the [Apache Bigtop][] platform to provide HDFS master resources. +[Apache Bigtop]: http://bigtop.apache.org/ -## Usage -This charm is intended to be deployed via one of the -[apache bigtop bundles](https://jujucharms.com/u/bigdata-dev/#bundles). -For example: +# Deploying - juju deploy hadoop-processing +A working Juju installation is assumed to be present. If Juju is not yet set +up, please follow the [getting-started][] instructions prior to deploying this +charm. -> Note: With Juju versions < 2.0, you will need to use [juju-deployer][] to -deploy the bundle. +This charm is intended to be deployed via one of the [apache bigtop bundles][]. +For example: -This will deploy the Apache Bigtop platform with a workload node -preconfigured to work with the cluster. + juju deploy hadoop-processing -You can also manually load and run map-reduce jobs via the plugin charm -included in the bundles linked above: +> **Note**: The above assumes Juju 2.0 or greater. If using an earlier version +of Juju, use [juju-quickstart][] with the following syntax: `juju quickstart +hadoop-processing`. - juju scp my-job.jar plugin/0: - juju ssh plugin/0 - hadoop jar my-job.jar +This will deploy an Apache Bigtop cluster with this charm acting as the +NameNode. More information about this deployment can be found in the +[bundle readme](https://jujucharms.com/hadoop-processing/). +## Network-Restricted Environments +Charms can be deployed in environments with limited network access. To deploy +in this environment, configure a Juju model with appropriate proxy and/or +mirror options. See [Configuring Models][] for more information. -[juju-deployer]: https://pypi.python.org/pypi/juju-deployer/ +[getting-started]: https://jujucharms.com/docs/stable/getting-started +[apache bigtop bundles]: https://jujucharms.com/u/bigdata-charmers/#bundles +[juju-quickstart]: https://launchpad.net/juju-quickstart +[Configuring Models]: https://jujucharms.com/docs/stable/models-config -## Status and Smoke Test +# Verifying +## Status Apache Bigtop charms provide extended status reporting to indicate when they are ready: - juju status --format=tabular + juju status This is particularly useful when combined with `watch` to track the on-going progress of the deployment: - watch -n 0.5 juju status --format=tabular + watch -n 2 juju status -The message for each unit will provide information about that unit's state. -Once they all indicate that they are ready, you can perform a "smoke test" -to verify HDFS or YARN services are working as expected. Trigger the -`smoke-test` action by: +The message column will provide information about a given unit's state. +This charm is ready for use once the status message indicates that it is +ready with datanodes. - juju action do namenode/0 smoke-test - juju action do resourcemanager/0 smoke-test +## Smoke Test +This charm provides a `smoke-test` action that can be used to verify the +application is functioning as expected. Run the action as follows: -After a few seconds or so, you can check the results of the smoke test: + juju run-action namenode/0 smoke-test - juju action status +> **Note**: The above assumes Juju 2.0 or greater. If using an earlier version +of Juju, the syntax is `juju action do namenode/0 smoke-test`. -You will see `status: completed` if the smoke test was successful, or -`status: failed` if it was not. You can get more information on why it failed -via: +Watch the progress of the smoke test actions with: - juju action fetch <action-id> + watch -n 2 juju show-action-status +> **Note**: The above assumes Juju 2.0 or greater. If using an earlier version +of Juju, the syntax is `juju action status`. -## Deploying in Network-Restricted Environments +Eventually, the action should settle to `status: completed`. If it +reports `status: failed`, the application is not working as expected. Get +more information about a specific smoke test with: -Charms can be deployed in environments with limited network access. To deploy -in this environment, you will need a local mirror to serve required packages. + juju show-action-output <action-id> + +> **Note**: The above assumes Juju 2.0 or greater. If using an earlier version +of Juju, the syntax is `juju action fetch <action-id>`. + +## Utilities +This charm includes Hadoop command line and web utilities that can be used +to verify information about the cluster. + +Show the dfsadmin report on the command line with the following: + + juju run --application namenode "su hdfs -c 'hdfs dfsadmin -report'" + +To access the HDFS web console, find the `PUBLIC-ADDRESS` of the +namenode application and expose it: + juju status namenode + juju expose namenode -### Mirroring Packages +The web interface will be available at the following URL: -You can setup a local mirror for apt packages using squid-deb-proxy. -For instructions on configuring juju to use this, see the -[Juju Proxy Documentation](https://juju.ubuntu.com/docs/howto-proxies.html). + http://NAMENODE_PUBLIC_IP:50070 -## Contact Information +# Contact Information - <bigdata@lists.ubuntu.com> -## Hadoop +# Resources - [Apache Bigtop](http://bigtop.apache.org/) home page - [Apache Bigtop issue tracking](http://bigtop.apache.org/issue-tracking.html) - [Apache Bigtop mailing lists](http://bigtop.apache.org/mail-lists.html) -- [Apache Bigtop charms](https://jujucharms.com/q/apache/bigtop) +- [Juju Bigtop charms](https://jujucharms.com/q/apache/bigtop) +- [Juju mailing list](https://lists.ubuntu.com/mailman/listinfo/juju) +- [Juju community](https://jujucharms.com/community) diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/actions.yaml b/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/actions.yaml index ee93b4cb..c2d65aec 100644 --- a/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/actions.yaml +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/actions.yaml @@ -1,2 +1,2 @@ smoke-test: - description: Verify that HDFS is working by creating and removing a test directory. + description: Run a simple HDFS smoke test. diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/actions/smoke-test b/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/actions/smoke-test index 58ffce21..391b6261 100755 --- a/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/actions/smoke-test +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/actions/smoke-test @@ -22,7 +22,7 @@ from jujubigdata.utils import run_as from charms.reactive import is_state if not is_state('apache-bigtop-namenode.ready'): - hookenv.action_fail('NameNode service not yet ready') + hookenv.action_fail('Charm is not yet ready') # verify the hdfs-test directory does not already exist diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/layer.yaml b/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/layer.yaml index 332a6e3d..3fca8270 100644 --- a/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/layer.yaml +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/layer.yaml @@ -1,4 +1,4 @@ -repo: git@github.com:juju-solutions/layer-hadoop-namenode.git +repo: https://github.com/apache/bigtop/tree/master/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode includes: - 'layer:apache-bigtop-base' - 'interface:dfs' diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/metadata.yaml b/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/metadata.yaml index ab51ce43..a358a6d7 100644 --- a/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/metadata.yaml +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/metadata.yaml @@ -1,12 +1,12 @@ name: hadoop-namenode -summary: HDFS master (NameNode) for Apache Bigtop platform +summary: HDFS master (NameNode) from Apache Bigtop maintainer: Juju Big Data <bigdata@lists.ubuntu.com> description: > Hadoop is a software platform that lets one easily write and run applications that process vast amounts of data. - This charm manages the HDFS master node (NameNode). -tags: ["applications", "bigdata", "bigtop", "hadoop", "apache"] + This charm provides the HDFS master node (NameNode). +tags: [] provides: namenode: interface: dfs diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/reactive/namenode.py b/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/reactive/namenode.py index c39a6098..c8a71dae 100644 --- a/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/reactive/namenode.py +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/reactive/namenode.py @@ -15,7 +15,9 @@ # limitations under the License. from charms.reactive import is_state, remove_state, set_state, when, when_not -from charms.layer.apache_bigtop_base import Bigtop, get_layer_opts, get_fqdn +from charms.layer.apache_bigtop_base import ( + Bigtop, get_hadoop_version, get_layer_opts, get_fqdn +) from charmhelpers.core import hookenv, host from jujubigdata import utils from path import Path @@ -50,6 +52,8 @@ def send_early_install_info(remote): def install_namenode(): hookenv.status_set('maintenance', 'installing namenode') bigtop = Bigtop() + hdfs_port = get_layer_opts().port('namenode') + webhdfs_port = get_layer_opts().port('nn_webapp_http') bigtop.render_site_yaml( hosts={ 'namenode': get_fqdn(), @@ -58,6 +62,10 @@ def install_namenode(): 'namenode', 'mapred-app', ], + overrides={ + 'hadoop::common_hdfs::hadoop_namenode_port': hdfs_port, + 'hadoop::common_hdfs::hadoop_namenode_http_port': webhdfs_port, + } ) bigtop.trigger_puppet() @@ -96,6 +104,7 @@ def start_namenode(): for port in get_layer_opts().exposed_ports('namenode'): hookenv.open_port(port) set_state('apache-bigtop-namenode.started') + hookenv.application_version_set(get_hadoop_version()) hookenv.status_set('maintenance', 'namenode started') diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/tests/01-basic-deployment.py b/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/tests/01-basic-deployment.py index 15c00c9d..38aa45b9 100755 --- a/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/tests/01-basic-deployment.py +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-namenode/tests/01-basic-deployment.py @@ -28,7 +28,7 @@ class TestDeploy(unittest.TestCase): """ def test_deploy(self): - self.d = amulet.Deployment(series='trusty') + self.d = amulet.Deployment(series='xenial') self.d.add('namenode', 'hadoop-namenode') self.d.setup(timeout=900) self.d.sentry.wait(timeout=1800) diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/README.md b/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/README.md index cbea7f02..405c08ac 100644 --- a/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/README.md +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/README.md @@ -14,79 +14,109 @@ See the License for the specific language governing permissions and limitations under the License. --> -## Overview +# Overview The Apache Hadoop software library is a framework that allows for the distributed processing of large data sets across clusters of computers using a simple programming model. -This charm facilitates communication between core Apache Bigtop cluster -components and workload charms. +This charm facilitates communication between Hadoop components of an +[Apache Bigtop][] cluster and workload applications. +[Apache Bigtop]: http://bigtop.apache.org/ -## Usage -This charm is intended to be deployed via one of the -[apache bigtop bundles](https://jujucharms.com/u/bigdata-dev/#bundles). -For example: +# Deploying - juju deploy hadoop-processing +A working Juju installation is assumed to be present. If Juju is not yet set +up, please follow the [getting-started][] instructions prior to deploying this +charm. -> Note: With Juju versions < 2.0, you will need to use [juju-deployer][] to -deploy the bundle. +This charm is intended to be deployed via one of the [apache bigtop bundles][]. +For example: -This will deploy the Apache Bigtop platform with a workload node -preconfigured to work with the cluster. + juju deploy hadoop-processing -You could extend this deployment, for example, to analyze data using Apache Pig. -Simply deploy Pig and attach it to the same plugin: +> **Note**: The above assumes Juju 2.0 or greater. If using an earlier version +of Juju, use [juju-quickstart][] with the following syntax: `juju quickstart +hadoop-processing`. - juju deploy apache-pig pig - juju add-relation plugin pig +This will deploy an Apache Bigtop cluster with a client unit preconfigured to +work with the cluster. More information about this deployment can be found in the +[bundle readme](https://jujucharms.com/hadoop-processing/). +## Network-Restricted Environments +Charms can be deployed in environments with limited network access. To deploy +in this environment, configure a Juju model with appropriate proxy and/or +mirror options. See [Configuring Models][] for more information. -[juju-deployer]: https://pypi.python.org/pypi/juju-deployer/ +[getting-started]: https://jujucharms.com/docs/stable/getting-started +[apache bigtop bundles]: https://jujucharms.com/u/bigdata-charmers/#bundles +[juju-quickstart]: https://launchpad.net/juju-quickstart +[Configuring Models]: https://jujucharms.com/docs/stable/models-config -## Status and Smoke Test +# Verifying +## Status Apache Bigtop charms provide extended status reporting to indicate when they are ready: - juju status --format=tabular + juju status This is particularly useful when combined with `watch` to track the on-going progress of the deployment: - watch -n 0.5 juju status --format=tabular + watch -n 2 juju status + +The message column will provide information about a given unit's state. +This charm is ready for use once the status message indicates that it is +ready with hdfs and/or yarn. + +## Smoke Test +This charm provides a `smoke-test` action that can be used to verify the +application is functioning as expected. Run the action as follows: + + juju run-action plugin/0 smoke-test + +> **Note**: The above assumes Juju 2.0 or greater. If using an earlier version +of Juju, the syntax is `juju action do plugin/0 smoke-test`. + +Watch the progress of the smoke test actions with: + + watch -n 2 juju show-action-status + +> **Note**: The above assumes Juju 2.0 or greater. If using an earlier version +of Juju, the syntax is `juju action status`. -The message for each unit will provide information about that unit's state. -Once they all indicate that they are ready, you can perform a "smoke test" -to verify HDFS or YARN services are working as expected. Trigger the -`smoke-test` action by: +Eventually, the action should settle to `status: completed`. If it +reports `status: failed`, the application is not working as expected. Get +more information about a specific smoke test with: - juju action do namenode/0 smoke-test - juju action do resourcemanager/0 smoke-test + juju show-action-output <action-id> -After a few seconds or so, you can check the results of the smoke test: +> **Note**: The above assumes Juju 2.0 or greater. If using an earlier version +of Juju, the syntax is `juju action fetch <action-id>`. - juju action status +## Utilities +This charm includes Hadoop command line utilities that can be used +to verify information about the cluster. -You will see `status: completed` if the smoke test was successful, or -`status: failed` if it was not. You can get more information on why it failed -via: +Show the dfsadmin report on the command line with the following: - juju action fetch <action-id> + juju run --application plugin "su hdfs -c 'hdfs dfsadmin -report'" -## Contact Information +# Contact Information - <bigdata@lists.ubuntu.com> -## Resources +# Resources - [Apache Bigtop](http://bigtop.apache.org/) home page - [Apache Bigtop issue tracking](http://bigtop.apache.org/issue-tracking.html) - [Apache Bigtop mailing lists](http://bigtop.apache.org/mail-lists.html) -- [Apache Bigtop charms](https://jujucharms.com/q/apache/bigtop) +- [Juju Bigtop charms](https://jujucharms.com/q/apache/bigtop) +- [Juju mailing list](https://lists.ubuntu.com/mailman/listinfo/juju) +- [Juju community](https://jujucharms.com/community) diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/actions.yaml b/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/actions.yaml new file mode 100644 index 00000000..c2d65aec --- /dev/null +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/actions.yaml @@ -0,0 +1,2 @@ +smoke-test: + description: Run a simple HDFS smoke test. diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/actions/smoke-test b/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/actions/smoke-test new file mode 100755 index 00000000..65ba07c6 --- /dev/null +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/actions/smoke-test @@ -0,0 +1,62 @@ +#!/usr/bin/env python3 + +# Licensed to the Apache Software Foundation (ASF) under one or more +# contributor license agreements. See the NOTICE file distributed with +# this work for additional information regarding copyright ownership. +# The ASF licenses this file to You under the Apache License, Version 2.0 +# (the "License"); you may not use this file except in compliance with +# the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +import sys + +from charmhelpers.core import hookenv +from jujubigdata.utils import run_as +from charms.reactive import is_state + +if not is_state('apache-bigtop-plugin.hdfs.ready'): + hookenv.action_fail('Charm is not yet ready') + + +# verify the hdfs-test directory does not already exist +output = run_as('ubuntu', 'hdfs', 'dfs', '-ls', '/tmp', capture_output=True) +if '/tmp/hdfs-test' in output: + run_as('ubuntu', 'hdfs', 'dfs', '-rm', '-R', '/tmp/hdfs-test') + output = run_as('ubuntu', 'hdfs', 'dfs', '-ls', '/tmp', capture_output=True) + if 'hdfs-test' in output: + hookenv.action_fail('Unable to remove existing hdfs-test directory') + sys.exit() + +# create the directory +run_as('ubuntu', 'hdfs', 'dfs', '-mkdir', '-p', '/tmp/hdfs-test') +run_as('ubuntu', 'hdfs', 'dfs', '-chmod', '-R', '777', '/tmp/hdfs-test') + +# verify the newly created hdfs-test subdirectory exists +output = run_as('ubuntu', 'hdfs', 'dfs', '-ls', '/tmp', capture_output=True) +for line in output.split('\n'): + if '/tmp/hdfs-test' in line: + if 'ubuntu' not in line or 'drwxrwxrwx' not in line: + hookenv.action_fail('Permissions incorrect for hdfs-test directory') + sys.exit() + break +else: + hookenv.action_fail('Unable to create hdfs-test directory') + sys.exit() + +# remove the directory +run_as('ubuntu', 'hdfs', 'dfs', '-rm', '-R', '/tmp/hdfs-test') + +# verify the hdfs-test subdirectory has been removed +output = run_as('ubuntu', 'hdfs', 'dfs', '-ls', '/tmp', capture_output=True) +if '/tmp/hdfs-test' in output: + hookenv.action_fail('Unable to remove hdfs-test directory') + sys.exit() + +hookenv.action_set({'outcome': 'success'}) diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/layer.yaml b/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/layer.yaml index 5ddc2c9b..ceedad7f 100644 --- a/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/layer.yaml +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/layer.yaml @@ -1,8 +1,12 @@ -repo: git@github.com:juju-solutions/layer-hadoop-plugin.git -includes: ['layer:apache-bigtop-base', 'interface:hadoop-plugin', 'interface:dfs', 'interface:mapred'] +repo: https://github.com/apache/bigtop/tree/master/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin +includes: + - 'layer:apache-bigtop-base' + - 'interface:hadoop-plugin' + - 'interface:dfs' + - 'interface:mapred' options: basic: use_venv: true metadata: deletes: - - requires.java + - provides.java diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/metadata.yaml b/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/metadata.yaml index a5fd4538..4df86f1b 100644 --- a/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/metadata.yaml +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/metadata.yaml @@ -1,5 +1,5 @@ name: hadoop-plugin -summary: Simplified connection point for Apache Bigtop platform +summary: Facilitates communication with an Apache Bigtop Hadoop cluster maintainer: Juju Big Data <bigdata@lists.ubuntu.com> description: > Hadoop is a software platform that lets one easily write and @@ -8,7 +8,7 @@ description: > This charm provides a simplified connection point for client / workload services which require access to Apache Hadoop. This connection is established via the Apache Bigtop gateway. -tags: ["applications", "bigdata", "hadoop", "apache"] +tags: [] subordinate: true requires: namenode: diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/reactive/apache_bigtop_plugin.py b/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/reactive/apache_bigtop_plugin.py index e5b12758..e680002a 100644 --- a/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/reactive/apache_bigtop_plugin.py +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/reactive/apache_bigtop_plugin.py @@ -42,6 +42,7 @@ def install_hadoop_client_hdfs(principal, namenode): bigtop.render_site_yaml(hosts=hosts, roles='hadoop-client') bigtop.trigger_puppet() set_state('apache-bigtop-plugin.hdfs.installed') + hookenv.application_version_set(get_hadoop_version()) hookenv.status_set('maintenance', 'plugin (hdfs) installed') else: hookenv.status_set('waiting', 'waiting for namenode fqdn') diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/tests/01-basic-deployment.py b/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/tests/01-basic-deployment.py index 512630df..815f9fbf 100755 --- a/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/tests/01-basic-deployment.py +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-plugin/tests/01-basic-deployment.py @@ -29,7 +29,7 @@ class TestDeploy(unittest.TestCase): """ def test_deploy(self): - self.d = amulet.Deployment(series='trusty') + self.d = amulet.Deployment(series='xenial') self.d.load({ 'services': { 'client': {'charm': 'hadoop-client'}, diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/README.md b/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/README.md index 02508817..430cc974 100644 --- a/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/README.md +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/README.md @@ -14,142 +14,170 @@ See the License for the specific language governing permissions and limitations under the License. --> -## Overview +# Overview The Apache Hadoop software library is a framework that allows for the distributed processing of large data sets across clusters of computers using a simple programming model. -This charm deploys the ResourceManager component of the Apache Bigtop platform -to provide YARN master resources. +This charm deploys the ResourceManager component of the [Apache Bigtop][] +platform to provide YARN master resources. +[Apache Bigtop]: http://bigtop.apache.org/ -## Usage -This charm is intended to be deployed via one of the -[apache bigtop bundles](https://jujucharms.com/u/bigdata-dev/#bundles). -For example: +# Deploying - juju deploy hadoop-processing +A working Juju installation is assumed to be present. If Juju is not yet set +up, please follow the [getting-started][] instructions prior to deploying this +charm. -> Note: With Juju versions < 2.0, you will need to use [juju-deployer][] to -deploy the bundle. +This charm is intended to be deployed via one of the [apache bigtop bundles][]. +For example: -This will deploy the Apache Bigtop platform with a workload node -preconfigured to work with the cluster. + juju deploy hadoop-processing -You can also manually load and run map-reduce jobs via the plugin charm -included in the bundles linked above: +> **Note**: The above assumes Juju 2.0 or greater. If using an earlier version +of Juju, use [juju-quickstart][] with the following syntax: `juju quickstart +hadoop-processing`. - juju scp my-job.jar plugin/0: - juju ssh plugin/0 - hadoop jar my-job.jar +This will deploy an Apache Bigtop cluster with this charm acting as the +ResourceManager. More information about this deployment can be found in the +[bundle readme](https://jujucharms.com/hadoop-processing/). +## Network-Restricted Environments +Charms can be deployed in environments with limited network access. To deploy +in this environment, configure a Juju model with appropriate proxy and/or +mirror options. See [Configuring Models][] for more information. -[juju-deployer]: https://pypi.python.org/pypi/juju-deployer/ +[getting-started]: https://jujucharms.com/docs/stable/getting-started +[apache bigtop bundles]: https://jujucharms.com/u/bigdata-charmers/#bundles +[juju-quickstart]: https://launchpad.net/juju-quickstart +[Configuring Models]: https://jujucharms.com/docs/stable/models-config -## Status and Smoke Test +# Verifying +## Status Apache Bigtop charms provide extended status reporting to indicate when they are ready: - juju status --format=tabular + juju status This is particularly useful when combined with `watch` to track the on-going progress of the deployment: - watch -n 0.5 juju status --format=tabular + watch -n 2 juju status -The message for each unit will provide information about that unit's state. -Once they all indicate that they are ready, you can perform a "smoke test" -to verify HDFS or YARN services are working as expected. Trigger the -`smoke-test` action by: +The message column will provide information about a given unit's state. +This charm is ready for use once the status message indicates that it is +ready with nodemanagers. - juju action do namenode/0 smoke-test - juju action do resourcemanager/0 smoke-test +## Smoke Test +This charm provides a `smoke-test` action that can be used to verify the +application is functioning as expected. This action executes the 'yarn' +smoke tests provided by Apache Bigtop and may take up to +10 minutes to complete. Run the action as follows: -After a few seconds or so, you can check the results of the smoke test: + juju run-action resourcemanager/0 smoke-test - juju action status +> **Note**: The above assumes Juju 2.0 or greater. If using an earlier version +of Juju, the syntax is `juju action do resourcemanager/0 smoke-test`. -You will see `status: completed` if the smoke test was successful, or -`status: failed` if it was not. You can get more information on why it failed -via: +Watch the progress of the smoke test actions with: - juju action fetch <action-id> + watch -n 2 juju show-action-status +> **Note**: The above assumes Juju 2.0 or greater. If using an earlier version +of Juju, the syntax is `juju action status`. -## Benchmarking +Eventually, the action should settle to `status: completed`. If it +reports `status: failed`, the application is not working as expected. Get +more information about a specific smoke test with: -This charm provides several benchmarks to gauge the performance of your -environment. + juju show-action-output <action-id> -The easiest way to run the benchmarks on this service is to relate it to the -[Benchmark GUI][]. You will likely also want to relate it to the -[Benchmark Collector][] to have machine-level information collected during the -benchmark, for a more complete picture of how the machine performed. +> **Note**: The above assumes Juju 2.0 or greater. If using an earlier version +of Juju, the syntax is `juju action fetch <action-id>`. -[Benchmark GUI]: https://jujucharms.com/benchmark-gui/ -[Benchmark Collector]: https://jujucharms.com/benchmark-collector/ +## Utilities +This charm includes Hadoop command line and web utilities that can be used +to verify information about the cluster. -However, each benchmark is also an action that can be called manually: +Show the running nodes on the command line with the following: - $ juju action do resourcemanager/0 nnbench - Action queued with id: 55887b40-116c-4020-8b35-1e28a54cc622 - $ juju action fetch --wait 0 55887b40-116c-4020-8b35-1e28a54cc622 + juju run --application resourcemanager "su yarn -c 'yarn node -list'" - results: - meta: - composite: - direction: asc - units: secs - value: "128" - start: 2016-02-04T14:55:39Z - stop: 2016-02-04T14:57:47Z - results: - raw: '{"BAD_ID": "0", "FILE: Number of read operations": "0", "Reduce input groups": - "8", "Reduce input records": "95", "Map output bytes": "1823", "Map input records": - "12", "Combine input records": "0", "HDFS: Number of bytes read": "18635", "FILE: - Number of bytes written": "32999982", "HDFS: Number of write operations": "330", - "Combine output records": "0", "Total committed heap usage (bytes)": "3144749056", - "Bytes Written": "164", "WRONG_LENGTH": "0", "Failed Shuffles": "0", "FILE: - Number of bytes read": "27879457", "WRONG_MAP": "0", "Spilled Records": "190", - "Merged Map outputs": "72", "HDFS: Number of large read operations": "0", "Reduce - shuffle bytes": "2445", "FILE: Number of large read operations": "0", "Map output - materialized bytes": "2445", "IO_ERROR": "0", "CONNECTION": "0", "HDFS: Number - of read operations": "567", "Map output records": "95", "Reduce output records": - "8", "WRONG_REDUCE": "0", "HDFS: Number of bytes written": "27412", "GC time - elapsed (ms)": "603", "Input split bytes": "1610", "Shuffled Maps ": "72", "FILE: - Number of write operations": "0", "Bytes Read": "1490"}' - status: completed - timing: - completed: 2016-02-04 14:57:48 +0000 UTC - enqueued: 2016-02-04 14:55:14 +0000 UTC - started: 2016-02-04 14:55:27 +0000 UTC +To access the Resource Manager web consoles, find the `PUBLIC-ADDRESS` of the +resourcemanager application and expose it: + juju status resourcemanager + juju expose resourcemanager -## Deploying in Network-Restricted Environments +The YARN and Job History web interfaces will be available at the following URLs: -Charms can be deployed in environments with limited network access. To deploy -in this environment, you will need a local mirror to serve required packages. + http://RESOURCEMANAGER_PUBLIC_IP:8088 + http://RESOURCEMANAGER_PUBLIC_IP:19888 + + +# Benchmarking + +This charm provides several benchmarks to gauge the performance of the +cluster. Each benchmark is an action that can be run with `juju run-action`: + $ juju actions resourcemanager + ACTION DESCRIPTION + mrbench Mapreduce benchmark for small jobs + nnbench Load test the NameNode hardware and configuration + smoke-test Run an Apache Bigtop smoke test. + teragen Generate data with teragen + terasort Runs teragen to generate sample data, and then runs terasort to sort that data + testdfsio DFS IO Testing -### Mirroring Packages + $ juju run-action resourcemanager/0 nnbench + Action queued with id: 55887b40-116c-4020-8b35-1e28a54cc622 -You can setup a local mirror for apt packages using squid-deb-proxy. -For instructions on configuring juju to use this, see the -[Juju Proxy Documentation](https://juju.ubuntu.com/docs/howto-proxies.html). + $ juju show-action-output 55887b40-116c-4020-8b35-1e28a54cc622 + results: + meta: + composite: + direction: asc + units: secs + value: "128" + start: 2016-02-04T14:55:39Z + stop: 2016-02-04T14:57:47Z + results: + raw: '{"BAD_ID": "0", "FILE: Number of read operations": "0", "Reduce input groups": + "8", "Reduce input records": "95", "Map output bytes": "1823", "Map input records": + "12", "Combine input records": "0", "HDFS: Number of bytes read": "18635", "FILE: + Number of bytes written": "32999982", "HDFS: Number of write operations": "330", + "Combine output records": "0", "Total committed heap usage (bytes)": "3144749056", + "Bytes Written": "164", "WRONG_LENGTH": "0", "Failed Shuffles": "0", "FILE: + Number of bytes read": "27879457", "WRONG_MAP": "0", "Spilled Records": "190", + "Merged Map outputs": "72", "HDFS: Number of large read operations": "0", "Reduce + shuffle bytes": "2445", "FILE: Number of large read operations": "0", "Map output + materialized bytes": "2445", "IO_ERROR": "0", "CONNECTION": "0", "HDFS: Number + of read operations": "567", "Map output records": "95", "Reduce output records": + "8", "WRONG_REDUCE": "0", "HDFS: Number of bytes written": "27412", "GC time + elapsed (ms)": "603", "Input split bytes": "1610", "Shuffled Maps ": "72", "FILE: + Number of write operations": "0", "Bytes Read": "1490"}' + status: completed + timing: + completed: 2016-02-04 14:57:48 +0000 UTC + enqueued: 2016-02-04 14:55:14 +0000 UTC + started: 2016-02-04 14:55:27 +0000 UTC -## Contact Information +# Contact Information - <bigdata@lists.ubuntu.com> -## Hadoop +# Resources - [Apache Bigtop](http://bigtop.apache.org/) home page - [Apache Bigtop issue tracking](http://bigtop.apache.org/issue-tracking.html) - [Apache Bigtop mailing lists](http://bigtop.apache.org/mail-lists.html) -- [Apache Bigtop charms](https://jujucharms.com/q/apache/bigtop) +- [Juju Bigtop charms](https://jujucharms.com/q/apache/bigtop) +- [Juju mailing list](https://lists.ubuntu.com/mailman/listinfo/juju) +- [Juju community](https://jujucharms.com/community) diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/actions.yaml b/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/actions.yaml index da4fc08e..77a644bf 100644 --- a/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/actions.yaml +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/actions.yaml @@ -1,6 +1,5 @@ smoke-test: - description: > - Verify that YARN is working as expected by running a small (1MB) terasort. + description: Run an Apache Bigtop smoke test. mrbench: description: Mapreduce benchmark for small jobs params: diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/actions/smoke-test b/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/actions/smoke-test index 9ef33a9f..3280e791 100755 --- a/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/actions/smoke-test +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/actions/smoke-test @@ -1,4 +1,4 @@ -#!/bin/bash +#!/usr/bin/env python3 # Licensed to the Apache Software Foundation (ASF) under one or more # contributor license agreements. See the NOTICE file distributed with @@ -15,66 +15,34 @@ # See the License for the specific language governing permissions and # limitations under the License. -set -ex +import sys +sys.path.append('lib') -if ! charms.reactive is_state 'apache-bigtop-resourcemanager.ready'; then - action-fail 'ResourceManager not yet ready' - exit -fi +from charmhelpers.core import hookenv +from charms.layer.apache_bigtop_base import Bigtop +from charms.reactive import is_state -IN_DIR='/tmp/smoke_test_in' -OUT_DIR='/tmp/smoke_test_out' -SIZE=10000 -OPTIONS='' -MAPS=1 -REDUCES=1 -NUMTASKS=1 -COMPRESSION='LocalDefault' +def fail(msg, output=None): + if output: + hookenv.action_set({'output': output}) + hookenv.action_fail(msg) + sys.exit() -OPTIONS="${OPTIONS} -D mapreduce.job.maps=${MAPS}" -OPTIONS="${OPTIONS} -D mapreduce.job.reduces=${REDUCES}" -OPTIONS="${OPTIONS} -D mapreduce.job.jvm.numtasks=${NUMTASKS}" -if [ $COMPRESSION == 'Disable' ] ; then - OPTIONS="${OPTIONS} -D mapreduce.map.output.compress=false" -elif [ $COMPRESSION == 'LocalDefault' ] ; then - OPTIONS="${OPTIONS}" -else - OPTIONS="${OPTIONS} -D mapreduce.map.output.compress=true -D mapred.map.output.compress.codec=org.apache.hadoop.io.compress.${COMPRESSION}Codec" -fi +if not is_state('apache-bigtop-resourcemanager.ready'): + fail('Charm is not yet ready to run the Bigtop smoke test(s)') -# create dir to store results -RUN=`date +%s` -RESULT_DIR=/opt/terasort-results -RESULT_LOG=${RESULT_DIR}/${RUN}.$$.log -mkdir -p ${RESULT_DIR} -chown -R hdfs ${RESULT_DIR} +# Bigtop smoke test components +smoke_components = ['yarn'] -# clean out any previous data (must be run as the hdfs user) -su hdfs << EOF -if hadoop fs -stat ${IN_DIR} &> /dev/null; then - hadoop fs -rm -r -skipTrash ${IN_DIR} || true -fi -if hadoop fs -stat ${OUT_DIR} &> /dev/null; then - hadoop fs -rm -r -skipTrash ${OUT_DIR} || true -fi -EOF +# Env required by test components +smoke_env = { + 'HADOOP_CONF_DIR': '/etc/hadoop/conf', +} -START=`date +%s` -# NB: Escaped vars in the block below (e.g., \${HADOOP_MAPRED_HOME}) come from -# the environment while non-escaped vars (e.g., ${IN_DIR}) are parameterized -# from this outer scope -su hdfs << EOF -. /etc/default/hadoop -echo 'generating data' -hadoop jar \${HADOOP_MAPRED_HOME}/hadoop-mapreduce-examples-*.jar teragen ${SIZE} ${IN_DIR} &>/dev/null -echo 'sorting data' -hadoop jar \${HADOOP_MAPRED_HOME}/hadoop-mapreduce-examples-*.jar terasort ${OPTIONS} ${IN_DIR} ${OUT_DIR} &> ${RESULT_LOG} -EOF -STOP=`date +%s` - -if ! grep -q 'Bytes Written=1000000' ${RESULT_LOG}; then - action-fail 'smoke-test failed' - action-set log="$(cat ${RESULT_LOG})" -fi -DURATION=`expr $STOP - $START` +bigtop = Bigtop() +result = bigtop.run_smoke_tests(smoke_components, smoke_env) +if result == 'success': + hookenv.action_set({'outcome': result}) +else: + fail('{} smoke tests failed'.format(smoke_components), result) diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/layer.yaml b/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/layer.yaml index ad0b5695..c2e34205 100644 --- a/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/layer.yaml +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/layer.yaml @@ -1,4 +1,4 @@ -repo: git@github.com:juju-solutions/layer-hadoop-resourcemanager.git +repo: https://github.com/apache/bigtop/tree/master/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager includes: - 'layer:apache-bigtop-base' - 'interface:dfs' diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/metadata.yaml b/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/metadata.yaml index 82b82cd7..695d5bfa 100644 --- a/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/metadata.yaml +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/metadata.yaml @@ -1,12 +1,12 @@ name: hadoop-resourcemanager -summary: YARN master (ResourceManager) for Apache Bigtop platform +summary: YARN master (ResourceManager) from Apache Bigtop maintainer: Juju Big Data <bigdata@lists.ubuntu.com> description: > Hadoop is a software platform that lets one easily write and run applications that process vast amounts of data. - This charm manages the YARN master node (ResourceManager). -tags: ["applications", "bigdata", "bigtop", "hadoop", "apache"] + This charm provides the YARN master node (ResourceManager). +tags: [] provides: resourcemanager: interface: mapred diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/reactive/resourcemanager.py b/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/reactive/resourcemanager.py index afca26bd..3f3e9ae7 100644 --- a/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/reactive/resourcemanager.py +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/reactive/resourcemanager.py @@ -15,7 +15,9 @@ # limitations under the License. from charms.reactive import is_state, remove_state, set_state, when, when_not -from charms.layer.apache_bigtop_base import Bigtop, get_layer_opts, get_fqdn +from charms.layer.apache_bigtop_base import ( + Bigtop, get_hadoop_version, get_layer_opts, get_fqdn +) from charmhelpers.core import hookenv, host from jujubigdata import utils @@ -61,11 +63,32 @@ def install_resourcemanager(namenode): """ if namenode.namenodes(): hookenv.status_set('maintenance', 'installing resourcemanager') + # Hosts nn_host = namenode.namenodes()[0] rm_host = get_fqdn() + + # Ports + rm_ipc = get_layer_opts().port('resourcemanager') + rm_http = get_layer_opts().port('rm_webapp_http') + jh_ipc = get_layer_opts().port('jobhistory') + jh_http = get_layer_opts().port('jh_webapp_http') + bigtop = Bigtop() - hosts = {'namenode': nn_host, 'resourcemanager': rm_host} - bigtop.render_site_yaml(hosts=hosts, roles='resourcemanager') + bigtop.render_site_yaml( + hosts={ + 'namenode': nn_host, + 'resourcemanager': rm_host, + }, + roles=[ + 'resourcemanager', + ], + overrides={ + 'hadoop::common_yarn::hadoop_rm_port': rm_ipc, + 'hadoop::common_yarn::hadoop_rm_webapp_port': rm_http, + 'hadoop::common_mapred_app::mapreduce_jobhistory_port': jh_ipc, + 'hadoop::common_mapred_app::mapreduce_jobhistory_webapp_port': jh_http, + } + ) bigtop.trigger_puppet() # /etc/hosts entries from the KV are not currently used for bigtop, @@ -104,7 +127,8 @@ def start_resourcemanager(namenode): for port in get_layer_opts().exposed_ports('resourcemanager'): hookenv.open_port(port) set_state('apache-bigtop-resourcemanager.started') - hookenv.status_set('active', 'ready') + hookenv.application_version_set(get_hadoop_version()) + hookenv.status_set('maintenance', 'resourcemanager started') ############################################################################### diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/tests/01-basic-deployment.py b/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/tests/01-basic-deployment.py index 65dbbbb5..3b694548 100755 --- a/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/tests/01-basic-deployment.py +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-resourcemanager/tests/01-basic-deployment.py @@ -28,7 +28,7 @@ class TestDeploy(unittest.TestCase): """ def test_deploy(self): - self.d = amulet.Deployment(series='trusty') + self.d = amulet.Deployment(series='xenial') self.d.add('resourcemanager', 'hadoop-resourcemanager') self.d.setup(timeout=900) self.d.sentry.wait(timeout=1800) diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/README.md b/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/README.md index 25800722..4bf240dd 100644 --- a/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/README.md +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/README.md @@ -14,103 +14,116 @@ See the License for the specific language governing permissions and limitations under the License. --> -## Overview +# Overview The Apache Hadoop software library is a framework that allows for the distributed processing of large data sets across clusters of computers using a simple programming model. This charm deploys a combined slave node running the NodeManager -and DataNode components of the Apache Bigtop platform +and DataNode components of the [Apache Bigtop][] platform to provide YARN compute and HDFS storage resources. +[Apache Bigtop]: http://bigtop.apache.org/ -## Usage -This charm is intended to be deployed via one of the -[apache bigtop bundles](https://jujucharms.com/u/bigdata-dev/#bundles). -For example: +# Deploying - juju deploy hadoop-processing +A working Juju installation is assumed to be present. If Juju is not yet set +up, please follow the [getting-started][] instructions prior to deploying this +charm. -> Note: With Juju versions < 2.0, you will need to use [juju-deployer][] to -deploy the bundle. +This charm is intended to be deployed via one of the [apache bigtop bundles][]. +For example: -This will deploy the Apache Bigtop platform with a workload node -preconfigured to work with the cluster. + juju deploy hadoop-processing -You can also manually load and run map-reduce jobs via the plugin charm -included in the bundles linked above: +> **Note**: The above assumes Juju 2.0 or greater. If using an earlier version +of Juju, use [juju-quickstart][] with the following syntax: `juju quickstart +hadoop-processing`. - juju scp my-job.jar plugin/0: - juju ssh plugin/0 - hadoop jar my-job.jar +This will deploy an Apache Bigtop cluster with 3 units of this charm acting as +the combined DataNode/NodeManager application. More information about this +deployment can be found in the [bundle readme](https://jujucharms.com/hadoop-processing/). +## Network-Restricted Environments +Charms can be deployed in environments with limited network access. To deploy +in this environment, configure a Juju model with appropriate proxy and/or +mirror options. See [Configuring Models][] for more information. -[juju-deployer]: https://pypi.python.org/pypi/juju-deployer/ +[getting-started]: https://jujucharms.com/docs/stable/getting-started +[apache bigtop bundles]: https://jujucharms.com/u/bigdata-charmers/#bundles +[juju-quickstart]: https://launchpad.net/juju-quickstart +[Configuring Models]: https://jujucharms.com/docs/stable/models-config -## Status and Smoke Test +# Verifying +## Status Apache Bigtop charms provide extended status reporting to indicate when they are ready: - juju status --format=tabular + juju status This is particularly useful when combined with `watch` to track the on-going progress of the deployment: - watch -n 0.5 juju status --format=tabular + watch -n 2 juju status -The message for each unit will provide information about that unit's state. -Once they all indicate that they are ready, you can perform a "smoke test" -to verify HDFS or YARN services are working as expected. Trigger the -`smoke-test` action by: +The message column will provide information about a given unit's state. +This charm is ready for use once the status message indicates that it is +ready as a datanode/nodemanager. - juju action do namenode/0 smoke-test - juju action do resourcemanager/0 smoke-test +## Smoke Test +This charm provides a `smoke-test` action that can be used to verify the +application is functioning as expected. This action executes the 'hdfs' +and 'mapreduce' smoke tests provided by Apache Bigtop and may take up to +30 minutes to complete. Run the action as follows: -After a few seconds or so, you can check the results of the smoke test: + juju run-action slave/0 smoke-test - juju action status +> **Note**: The above assumes Juju 2.0 or greater. If using an earlier version +of Juju, the syntax is `juju action do slave/0 smoke-test`. -You will see `status: completed` if the smoke test was successful, or -`status: failed` if it was not. You can get more information on why it failed -via: +Watch the progress of the smoke test actions with: - juju action fetch <action-id> + watch -n 2 juju show-action-status +> **Note**: The above assumes Juju 2.0 or greater. If using an earlier version +of Juju, the syntax is `juju action status`. -## Scaling +Eventually, the action should settle to `status: completed`. If it +reports `status: failed`, the application is not working as expected. Get +more information about a specific smoke test with: -The slave node is the "workhorse" of the Hadoop environment. To scale your -cluster performance and storage capabilities, you can simply add more slave -units. For example, to add three more units: + juju show-action-output <action-id> - juju add-unit slave -n 3 +> **Note**: The above assumes Juju 2.0 or greater. If using an earlier version +of Juju, the syntax is `juju action fetch <action-id>`. -## Deploying in Network-Restricted Environments +# Scaling -Charms can be deployed in environments with limited network access. To deploy -in this environment, you will need a local mirror to serve required packages. +To scale the cluster compute and storage capabilities, simply add more +slave units. To add one unit: + juju add-unit slave -### Mirroring Packages +Multiple units may be added at once. For example, add four more slave units: -You can setup a local mirror for apt packages using squid-deb-proxy. -For instructions on configuring juju to use this, see the -[Juju Proxy Documentation](https://juju.ubuntu.com/docs/howto-proxies.html). + juju add-unit -n4 slave -## Contact Information +# Contact Information - <bigdata@lists.ubuntu.com> -## Hadoop +# Resources - [Apache Bigtop](http://bigtop.apache.org/) home page - [Apache Bigtop issue tracking](http://bigtop.apache.org/issue-tracking.html) - [Apache Bigtop mailing lists](http://bigtop.apache.org/mail-lists.html) -- [Apache Bigtop charms](https://jujucharms.com/q/apache/bigtop) +- [Juju Bigtop charms](https://jujucharms.com/q/apache/bigtop) +- [Juju mailing list](https://lists.ubuntu.com/mailman/listinfo/juju) +- [Juju community](https://jujucharms.com/community) diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/actions.yaml b/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/actions.yaml new file mode 100644 index 00000000..7fbb3022 --- /dev/null +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/actions.yaml @@ -0,0 +1,3 @@ +smoke-test: + description: | + Run an Apache Bigtop smoke test. Requires 3 slave units. diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/actions/smoke-test b/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/actions/smoke-test new file mode 100755 index 00000000..6dec4b5a --- /dev/null +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/actions/smoke-test @@ -0,0 +1,49 @@ +#!/usr/bin/env python3 + +# Licensed to the Apache Software Foundation (ASF) under one or more +# contributor license agreements. See the NOTICE file distributed with +# this work for additional information regarding copyright ownership. +# The ASF licenses this file to You under the Apache License, Version 2.0 +# (the "License"); you may not use this file except in compliance with +# the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +import sys +sys.path.append('lib') + +from charmhelpers.core import hookenv +from charms.layer.apache_bigtop_base import Bigtop +from charms.reactive import is_state + + +def fail(msg, output=None): + if output: + hookenv.action_set({'output': output}) + hookenv.action_fail(msg) + sys.exit() + +if not is_state('apache-bigtop-datanode.started'): + fail('Charm is not yet ready to run the Bigtop smoke test(s)') + +# Bigtop smoke test components +smoke_components = ['hdfs', 'mapreduce'] + +# Env required by test components +smoke_env = { + 'HADOOP_CONF_DIR': '/etc/hadoop/conf', + 'HADOOP_MAPRED_HOME': '/usr/lib/hadoop-mapreduce', +} + +bigtop = Bigtop() +result = bigtop.run_smoke_tests(smoke_components, smoke_env) +if result == 'success': + hookenv.action_set({'outcome': result}) +else: + fail('{} smoke tests failed'.format(smoke_components), result) diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/layer.yaml b/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/layer.yaml index 73c66e62..e10b9daa 100644 --- a/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/layer.yaml +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/layer.yaml @@ -1,2 +1,4 @@ -repo: git@github.com:juju-solutions/layer-hadoop-slave.git -includes: ['layer:hadoop-datanode', 'layer:hadoop-nodemanager'] +repo: https://github.com/apache/bigtop/tree/master/bigtop-packages/src/charm/hadoop/layer-hadoop-slave +includes: + - 'layer:hadoop-datanode' + - 'layer:hadoop-nodemanager' diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/metadata.yaml b/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/metadata.yaml index f0b6cce7..e5bbc3cf 100644 --- a/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/metadata.yaml +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/metadata.yaml @@ -1,8 +1,8 @@ name: hadoop-slave -summary: Combined slave node (DataNode + NodeManager) for Apache Bigtop. +summary: Combined slave node (DataNode + NodeManager) from Apache Bigtop. description: > Hadoop is a software platform that lets one easily write and run applications that process vast amounts of data. - This charm manages both the storage node (DataNode) for HDFS and the + This charm provides both the storage node (DataNode) for HDFS and the compute node (NodeManager) for Yarn. diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/reactive/hadoop_status.py b/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/reactive/hadoop_status.py index 1e6d38f2..8690d625 100644 --- a/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/reactive/hadoop_status.py +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/reactive/hadoop_status.py @@ -15,11 +15,10 @@ # See the License for the specific language governing permissions and # limitations under the License. -from charms.reactive import when_any, when_none, is_state +from charms.reactive import when_any, is_state from charmhelpers.core.hookenv import status_set -@when_none('namenode.spec.mismatch', 'resourcemanager.spec.mismatch') @when_any( 'bigtop.available', 'apache-bigtop-datanode.pending', diff --git a/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/tests/01-basic-deployment.py b/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/tests/01-basic-deployment.py index e4790787..5899c0fc 100755 --- a/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/tests/01-basic-deployment.py +++ b/bigtop-packages/src/charm/hadoop/layer-hadoop-slave/tests/01-basic-deployment.py @@ -28,7 +28,7 @@ class TestDeploy(unittest.TestCase): """ def test_deploy(self): - self.d = amulet.Deployment(series='trusty') + self.d = amulet.Deployment(series='xenial') self.d.add('slave', 'hadoop-slave') self.d.setup(timeout=900) self.d.sentry.wait(timeout=1800) |