openstack/deb-sahara
Code Issues Proposed changes

New doc about autoconfiguration policy

Browse Source 
Change-Id: I65ea22407440be7aee6d59212231f8ac7ad9605d
Partial-bug: 1490687
This commit is contained in:
Vitaly Gridnev 2015-09-07 17:45:03 +03:00
parent f7ddd0c94a
commit 5ac5c8668d
2 changed files with 37 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
doc/source/userdoc/advanced.configuration.guide.rst
View File

@ -432,6 +432,12 @@ set to ``True`` and some extra configurations are needed:
It should be noted that in a situation when the host has no space for volume
creation, the created volume will have an ``Error`` state and can not be used.
Autoconfiguration for templates
-------------------------------
:doc:`../userdoc/configs_recommendations`
NTP service configuration
-------------------------

31
doc/source/userdoc/configs_recommendations.rst Normal file
View File

@ -0,0 +1,31 @@
Autoconfiguring templates
=========================
During the Liberty development cycle sahara implemented a tool that recommends and
applies configuration values for cluster templates and node group templates.
These recommendations are based on the number of specific instances and on
flavors of the cluster node groups. Currently the following plugins support
this feature:
* Cloudera;
* Spark;
* the Vanilla Apache Hadoop plugin.
By default this feature is enabled for all cluster templates and node group templates.
If you want to disable this feature for a particular cluster or node group template
you should set the ``use_autoconfig`` field to ``false``.
.. NOTE
Also, if you manually set configs from the list below, the recommended configs
will not be applied.
The following describes the settings for which sahara can recommend autoconfiguration:
The Cloudera, Spark and Vanilla Apache Hadoop plugin support configuring
``dfs.replication`` (``dfs_replication`` for Cloudera plugin) which is calculated as
a minimun from the amount of ``datanode`` (``HDFS_DATANODE`` for Cloudera plugin) instances in
the cluster and the default value for ``dfs.replication``.
The Vanilla Apache Hadoop plugin and Cloudera plugin support autoconfiguration
of basic YARN and MapReduce configs. These autoconfigurations are based on the following
documentation: http://docs.hortonworks.com/HDPDocuments/HDP2/HDP-2.0.9.1/bk_installing_manually_book/content/rpm-chap1-11.html