Skip to main content

​
Prerequisites

Complete these checks before preparing or executing a production DP upgrade.

​
Mirror Server

Use a clean Ubuntu 24.04 LTS amd64 server with:
  • 2 vCPU minimum; 4 vCPU recommended
  • 4 GB RAM minimum; 8 GB recommended
  • 100 GB total disk for the current artifact set
  • a stable IPv4 address reachable by all target DPs
  • outbound HTTPS to the current GitHub, R2, ACPS, and Ubuntu bootstrap destinations
The Mirror Server must be able to serve DP hosts over TCP 80.

​
Credentials

Have the following available before Menu 2:
  • ACPS username
  • ACPS password
  • common aella worker SSH password when DL/DA workers are configured
Do not place passwords or tokens into documentation, public tickets, or shell history unnecessarily.

​
DP snapshots

Create a full hypervisor snapshot of every DP VM before the destructive upgrade. The project does not provide supported OS or DP-runtime rollback commands. Snapshot restore is the recovery method when a destructive failure cannot be safely recovered in place.

​
Choose the correct mode

Confirm the current DP OS state and select:
  • Full OS Upgrade + Phase 2 for supported DPs on Ubuntu 16.04
  • Phase 2 Only for supported DPs already on Ubuntu 24.04
The current target is DP 6.6.0.

​
Cluster information

For DL/DA clusters, collect:
  • DL worker management IPs
  • DA worker management IPs
  • common aella worker password
  • normal master-to-worker connectivity
Do not include master IPs in the worker lists.

​
Final pre-execution gate

Do not start DP-side commands until:
  1. Configuration is saved
  2. Download and Prepare completed
  3. HTTP Distribution is enabled
  4. Verify Upgrade Readiness reports PASS
  5. Menu 7 commands were generated from the current configuration