资 源 简 介
Project description
Hotrepart is a demonstrator project to illustrate a process for online repartitioning of a relational database.
The ultimate goal is to demonstrate auto-scaling of an RDBMS under production load on a utility computing platform, probably Amazon EC2.
The current codebase does not approach production requirements, but nicely illustrates the principle of the process.
The project consists of a sample database schema and a load tester that puts the database under load, during which at any time the database can be repartitioned by calling a stored procedure via an SQL client. The database clients shouldn"t notice anything except (hopefully) an improvement in performance once the process is complete.
Future goals of this project include:
* Adapting the code to deploy new partitions on EC2 instances. Currently, the new partition runs on the same host as the existing database.
* Building a
文 件 列 表
connect.2.0.8.diff
lib
commons-beanutils-1.7.0.jar
commons-lang-2.3.jar
commons-logging-1.0.4.jar
postgresql-8.3-603.jdbc4.jar
spring-2.0.6.jar
loadtester.jar
loadtester.properties
safe.py
sql
items0.sql.unsafe
items_proxy.sql
modified.dblink.sql
wiki
.svn
DatabaseDesignCriteria.wiki
DatabaseInstallation.wiki
ImplementationDetails.wiki
LoadTesting.wiki
ProjectDescription.wiki
RepartitioningProcess.wiki
TableOfContents.wiki
TaskList.wiki
TermsUsed.wiki
wikigraphics
.svn
RepartitioningProcess.odg
RepartitioningProcessStep1.gif
RepartitioningProcessStep2.gif
RepartitioningProcessStep3.gif
RepartitioningProcessStep4.gif
