Upgrading TeamForge Orchestrate

Upgrade Overview

Proxy configuration steps after upgrading from 1.4.1 (or earlier)

Starting with version 8.1, Orchestrate must be configured to proxy through TeamForge's web server. This provides users with a single hostname/port combination for accessing TeamForge and Orchestrate. It is required since some new features (like the JIRA adapter) rely on the proxy.

See Proxy upgrade configuration after upgrading to set up proxying.

Data loss potential during upgrades

While we have designed the upgrade process to minimize data loss, there is still a chance that Orchestrate misses data from sources during an upgrade.

During normal operation, adapters communicate source data (such as builds and commits) directly with RabbitMQ, the message queue layer between your data sources and Orchestrate. Orchestrate then consumes these messages from the queue as fast as it can. Disruption to RabbitMQ or the database layer (MongoDB) may lead to data loss as follows:

• RabbitMQ cannot receive messages. If RabbitMQ is down, adapters are unable to deliver messages and Orchestrate never receives those "lost" messages.
• MongoDB cannot store received messages. If MongoDB is down, the Orchestrate queue reading processes unsuccessfully attempt to write the messages to the database and the message is lost.

Risk from these two scenarios can be mitigated by upgrading MongoDB and RabbitMQ components at times of low traffic (i.e., after work hours).

Backup your data

Before upgrading, it is a good idea to backup TeamForge Orchestrate system and data, see Backup and Restore.