From: <tho...@us...> - 2011-06-02 18:37:11
|
Revision: 4608 http://bigdata.svn.sourceforge.net/bigdata/?rev=4608&view=rev Author: thompsonbry Date: 2011-06-02 18:37:05 +0000 (Thu, 02 Jun 2011) Log Message: ----------- Documentation in build.xml on various deployment mechanisms. Modified Paths: -------------- branches/QUADS_QUERY_BRANCH/build.xml Modified: branches/QUADS_QUERY_BRANCH/build.xml =================================================================== --- branches/QUADS_QUERY_BRANCH/build.xml 2011-06-02 18:07:02 UTC (rev 4607) +++ branches/QUADS_QUERY_BRANCH/build.xml 2011-06-02 18:37:05 UTC (rev 4608) @@ -3,6 +3,50 @@ <!-- @todo change the release target to put release notes into the root of the archives. --> <!-- @todo maven2 setup so we can run and publish unit tests results. --> <!-- $Id$ --> + +<!-- Deployment models. + +Embedded. Bigdata can be run as an embedded database. You just need to bundle +the various jars. The jini and zookeeper dependencies are NOT required when +used as an embedded database. You can also use this approach to run bigdata +using the NanoSparqlServer with an embedded jetty servlet engine. + +See 'bundleJar'. + +Webapp. Bigdata can be deployed as a web application exposing a SPARQL end point +with a RESTful API. + +See 'war'. + +Cluster deployment. There are two ways to deploy the cluster. One is +using 'ant-install' from either the top-level directory (either as +checked out from SVN or as unpacked from the distribution artifact +DIST.${version}.tgz). The second is to use the deployment artifact +REL.${version}.tgz. + +(A) This provides deployment from the project source either as checked +out from SVN or form a DIST.${version}.tgz file. + +See 'ant-install' + +(B) This is the new deployment model, but it is not yet complete. +It is based on the "stage" target which provides unversioned jar +names in dist/bigdata/lib, lib-dl, and lib-ext. Basically, you +unpack the archive and install onto the node. A script can then +start the various services using pstart based on a desired node +configuration. See the btm branch and trac for details and code. + +See 'ant deploy-artifact' + +Source and binary releases. There are two targets which provide source and +binary releases. + +See 'ant-install-artifact' (source code artifact: DIST.${version}.tgz). + +See 'deploy-artifact' (binary artifact: REL.${version}.tgz). + +--> + <project name="bigdata" default="bundleJar" basedir="."> <property file="build.properties" /> @@ -1006,6 +1050,41 @@ <!-- --> <!-- RELEASE --> <!-- --> + <!-- + REL.${version}.tgz - This is basically a tarball wrapping the 'dist' + directory generated by "stage". It has everything you need to deploy + a bigdata instance to a node. The paradigm for a deployment based on + this artifact is to create an installer which will start the appropriate + services for each node of your cluster. For example, based on a config + file in a known location on each node. The installer can use pstart to + actually start the services as described by that configuration file. For + example, you might wrap the generated REL.${version}.tgz file with your + own installer. + + The structure of the archive is as follows. + + bigdata + bigdata/bin + bigdata/bin/disco-tool - command line utility for DiscoveryTool. + bigdata/bin/pstart - python service start utility. + bigdata/bin/config - jini configuration for use Jini ServiceStarter. + bigdata/bin/config/browser.config + bigdata/bin/config/disco.config + bigdata/bin/config/disco-logging.properties + bigdata/bin/config/reggie.config + bigdata/bin/config/serviceStarter.config - Jini ServiceStarter configuration for jini core services, zookeeper, and bigdata services. + bigdata/lib - unversioned jars. + bigdata/lib-dl - downloadable jars for jini ("*-dl.jar"). + bigdata/lib-ext - jsk-policy.jar + bigdata/var + bigdata/var/config - bigdata config files, + bigdata/var/config/build.properties - a copy of the top level project build.properties file. + bigdata/var/jini - jini core service configuration files, including startAll.config which is used to start the core jini services. + bigdata/var/logging - log4j and java.util log configuration files. + bigdata/var/policy - java policy files + bigdata/var/policy/policy.all - policy with all permissions granted. + bigdata/var/policy/service.policy + --> <target name="deploy-artifact" depends="clean, stage" description="Create compressed tar file for deployment."> @@ -1099,6 +1178,10 @@ </copy> </target> + <!-- + Generate DIST.${version}.tgz - This is basically a tarball containing the + entire source distribution. + --> <target name="ant-install-artifact" depends="deploy-artifact, ant-install-prepare" description="Create complete source tar file for ant based install."> This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site. |