> ## Documentation Index
> Fetch the complete documentation index at: https://dpos.xdr.ooo/llms.txt
> Use this file to discover all available pages before exploring further.

# 1. Upgrade Work Plan

> Understand the full maintenance flow, choose the correct path, and prepare the Mirror Server before touching any DP.

# 1. Upgrade Work Plan

Start here. This page gives the full operator sequence before you install or upgrade anything.

## Goal

For a Full upgrade, the end state is:

```text theme={null}
DP 6.2.x–6.5.0 on Ubuntu 16.04
  → Ubuntu 18.04
  → Ubuntu 20.04
  → Ubuntu 22.04
  → Ubuntu 24.04
  → DP 6.6.0 Phase 2 bringup
  → final validation
```

Every Ubuntu LTS hop is required.

## Choose your path

| Current DP state                             | Use this path                 | Final state             |
| -------------------------------------------- | ----------------------------- | ----------------------- |
| Supported DP 6.2.x–6.5.0 on Ubuntu 16.04     | **Full OS Upgrade + Phase 2** | Ubuntu 24.04 + DP 6.6.0 |
| Supported DP 6.2–6.5 already on Ubuntu 24.04 | **Phase 2 Only**              | DP 6.6.0                |
| Healthy DP 6.6.0 already on Ubuntu 24.04     | Normally no upgrade           | No change               |

## Plan the maintenance window

For the illustrated **3-node DL + 3-node DA** example, use the following as planning estimates rather than guaranteed durations:

| Stage                                |  Example duration | When it happens                                                  |
| ------------------------------------ | ----------------: | ---------------------------------------------------------------- |
| Prepare Mirror Server                |     **2–3 hours** | Before the production maintenance window                         |
| DP pre-checks and service pause      |  **\~30 minutes** | Start of maintenance                                             |
| Powered-off snapshot/checkpoint gate |  **\~30 minutes** | After DP pre-checks and service pause, before the Ubuntu upgrade |
| Ubuntu OS upgrade block on all nodes | **\~1.5–2 hours** | After the snapshot/checkpoint gate                               |
| Phase 2 staging on all nodes         |     **\~2 hours** | Maintenance window; allow network contingency                    |
| DL cluster bringup                   |     **\~2 hours** | After staging                                                    |
| DA cluster bringup                   |     **\~2 hours** | After DL is healthy                                              |
| Final validation and resume          | **15–30 minutes** | Before closing maintenance                                       |

The example diagram estimates a production maintenance window of roughly **8–9 hours**, with Mirror Server preparation completed beforehand. Actual time depends on node count, storage performance, WAN/LAN throughput, reboot time, package state, and cluster recovery time.

<Info>
  Do not compress the maintenance plan by skipping an Ubuntu LTS hop or by starting DA bringup before DL has passed its readiness gate.
</Info>

## Full maintenance sequence

<Steps>
  <Step title="Prepare the Mirror Server before downtime">
    Complete configuration, download/prepare, HTTP distribution, readiness validation, and generate fresh DP client commands. The required gate is **Upgrade Readiness = PASS**.
  </Step>

  <Step title="Pre-check the DPs and pause services">
    Confirm connectivity and current versions, and pause DP services when the current generated procedure instructs you to. Pre-checks and the service pause must be complete before entering the snapshot/checkpoint gate.
  </Step>

  <Step title="Satisfy the powered-off snapshot/checkpoint gate">
    After DP pre-checks and the service pause, if the current upgrade client instructs you to shut down and create a powered-off snapshot/checkpoint, complete that gate on every required node. Do not start the Ubuntu upgrade until this gate is complete.
  </Step>

  <Step title="Upgrade Ubuntu on every required node">
    After the snapshot/checkpoint gate, follow `16.04 → 18.04 → 20.04 → 22.04 → 24.04` in order. Each hop is a separate controlled transition.
  </Step>

  <Step title="Stage DP 6.6.0 Phase 2 on every required node">
    Complete staging/preflight on the DL master, DL workers, DA master, and DA workers before final cluster orchestration.
  </Step>

  <Step title="Bring up the DL cluster first">
    Run final orchestration from the DL master and wait until the DL cluster is healthy.
  </Step>

  <Step title="Bring up the DA cluster next">
    Run final orchestration from the DA master only after the DL gate has passed.
  </Step>

  <Step title="Resume and validate the complete system">
    Confirm DP version, node readiness, pods, host services, NTP, DNS, licensing, Web UI, and normal operations before closing the change.
  </Step>
</Steps>

<Frame caption="Example work plan for a 3-node DL cluster and 3-node DA cluster">
  <img src="https://mintcdn.com/dp-os-upgrade/Ut7S53Dv3egtSHOn/images/google-sites/01-work-plan/1.png?fit=max&auto=format&n=Ut7S53Dv3egtSHOn&q=85&s=f6d3035468d691b550dcb56434e800f2" alt="DP Ubuntu 24.04 offline upgrade work plan" width="682" height="1246" data-path="images/google-sites/01-work-plan/1.png" />
</Frame>

## Go / no-go gates

Do not move to the next major stage unless the current stage is complete:

| Gate                     | Required result                                                                                                  |
| ------------------------ | ---------------------------------------------------------------------------------------------------------------- |
| Mirror Server            | `Upgrade Readiness = PASS`                                                                                       |
| Generated commands       | Fresh Menu 7 output for the current Mirror Server IP                                                             |
| Snapshot/checkpoint gate | After DP pre-checks and service pause, any snapshot action explicitly required by the current client is complete |
| Phase 1                  | Ubuntu 24.04 validated on every required node after the snapshot/checkpoint gate                                 |
| Phase 2 staging          | Staging/preflight reports ready for bringup                                                                      |
| DL bringup               | DL bringup completes successfully before DA starts                                                               |
| Final validation         | DP 6.6.0 and all required cluster/services healthy                                                               |

## Safety rules

* Create the required **hypervisor snapshot(s) of every DP VM** before destructive DP work.
* The project does not provide a supported OS or DP-runtime rollback command. Snapshot restore is the recovery path for destructive failures that cannot be recovered in place.
* Do not reuse old generated commands after the Mirror Server IP or relevant configuration changes.
* Do not mix historical 6.5.0, `apt-cacher-ng`, or full-mirror runbooks with the current 6.6.0 Mirror Manager workflow.
* Menu 7 generated by the current installed Mirror Manager is the authoritative source for environment-specific DP commands.

## Before you continue

* [ ] Correct Full / Phase 2 Only path selected
* [ ] Maintenance window approved
* [ ] All target nodes and roles identified
* [ ] Snapshot/recovery method confirmed
* [ ] Mirror Server build or existing-server plan confirmed
* [ ] Required network paths and credentials available

## Next step

If you need to build a new Mirror Server, continue with **[2. Ubuntu 24.04 Server Installation](/runbook/ubuntu-24-04-server-installation)**.

If the Mirror Server already runs Ubuntu 24.04 and meets the current requirements, go directly to **[3. Mirror Server Setup](/runbook/mirror-server-setup)**.