Presentation on how to chat with PDF using ChatGPT code interpreter
Alfresco Large Scale Enterprise Deployments
1. Large Scale Deployments
Scalability and Reliability
Paul Hampton
Director of Product Marketing
Alfresco
2. Introduction
● Alfresco provide enterprise scale Content
Management
● Many moving parts
o Application server
o Database
o Index stores
o Content stores
o Application
● Multiple points for potential failure
2
3. Single System Install
● Simple install
Database
Content store
● Installs everything on App Server
one system
Share
Tomcat
EHCache
DataBase
Content
Index
Store
3
4. What Needs to be Clustered?
● Application Server
Share
● Content Stores Tomcat
EHCache
● Databases
● The Application
DataBase
Content
Index
Store
● Note:
o Each server must have its own index
o EHCache must be synchronised between all servers
4
5. Clustered Alfresco
Load Balancer
● Clustered Alfresco servers
o Share database and content Share Share
store
Tomcat Tomcat
● Load balancer distributes EHCache EHCache
load between servers
● Point of failure
Index
Index
o Application Server
o Database
o Content store
DataBase
Content
Store
Can make use of high
availability disk (i.e. SAN
storage)
5
6. Replicated Content Stores
Load Balancer
● Used to synchronise content
between content stores Share Share
● Supports two modes Tomcat Tomcat
o Inbound EHCache EHCache
o Outbound
Content
Content
Index
Index
Store
Store
● Point of failure
o Application Server
o Database Database
o Content store Replicating Secondary Replicating
Content Content Content
Store Store Store
6
7. Example – Inbound
Load Balancer
● Content is copied on 1 5
read ?
Share Share
● Pulls content from
Tomcat Tomcat
EHCache EHCache
secondary content store 7
when needed 2
Content
Content
Index
Index
Store
Store
4 Database
6
Replicating Secondary Replicating
Content
Store
3 Content
Store
Content
Store
7
8. Example – Outbound
Load Balancer
● Content is copied on 1 6
write ?
Share Share
● Push content from
Tomcat Tomcat
EHCache EHCache
primary store to 7
secondary store 2
Content
Content
Index
Index
Store
Store
● Synchronous or
Asynchronous
4 Database
5
Replicating Secondary Replicating
Content
Store
3 Content
Store
Content
Store
8
9. Database Clustering
Load Balancer
● Use native DB
clustering Share Share
● This is specific to the
Tomcat Tomcat
EHCache EHCache
DB and OS
● Two options:
Content
Content
Index
Index
Store
Store
o Master / Slave
o Master / Master
Database
Database Database
(Master) (Slave)
Replicating Secondary Replicating
Content
Fallover Content Content
Store Store Store
Database Clustering
9
10. Application Clustering
● Point of failure Load Balancer
o Application Server Share Share
o Database
Tomcat Tomcat
o Content store
● How to address
scalability
10
12. JMX Administration
● Use standard JMX tools
to access configuration
settings
o Stop, Configure and Start
Subsystems
● Monitor
o Database connections
o Memory usage
o Alfresco Subsystems
12
13. Configuration Synchronisation
Load Balancer
● Use JMX to change
configuration settings Share Share
● Settings are saved to
Tomcat Tomcat
EHCache EHCache
Database
● Settings are
Content
Content
Index
Index
Store
Store
synchronised to other
Alfresco servers
Data base
Replicating Secondary Replicating
Content Content Content
Store Store Store
13
14. Summary
● Alfresco has been designed to be...
o Scalable
o Reliable
● Multiple options
● Enterprise Edition configuration tools simplify
configuration
● More Information
o WIKI – Cluster Configuration
o Docs – Administering an Alfresco ECM Production
Environment (page 158)
14