# Migrate to AlloyDB from PostgreSQL Using Database Migration Service - GSP1084

> Original task-by-task guide based on the **four-task Google Skills edition updated August 10, 2026**. This is an independently written practical walkthrough; see the [official lab](https://www.skills.google/games/7501/labs/45912) for the full original text and interface images.

## Overview

This exercise migrates a standalone PostgreSQL database hosted on a Compute Engine virtual machine into an existing AlloyDB for PostgreSQL cluster. Database Migration Service (DMS) first transfers existing data, then continuously replicates later changes using change data capture (CDC). Connectivity uses **VPC peering**. The source database's replication prerequisites, including `pglogical`, are already prepared in the lab environment.

### What you'll learn

- Check source PostgreSQL tables and reference row counts.
- Create a PostgreSQL source connection profile **inside** the DMS migration job wizard.
- Configure a continuous migration into the **existing** AlloyDB target and wait for CDC.
- Confirm the imported dataset from an AlloyDB client VM.
- Make a change to the source and check it replicates to the target.

## Setup and requirements

### Before you click the Start Lab button

This timed lab uses a temporary Google Cloud project and student credentials. Use a private/incognito browser window, start the lab, and sign in with **only** the account and password generated for this session; avoid using your own billing account. Don't click **End lab** until you're done, because it destroys the lab resources.

### How to start your lab and sign in to the Google Cloud console

1. Click **Start Lab**; the lab credentials and **Open Google Cloud console** link appear in the setup panel.
2. Open the console in an incognito window and sign in using the issued student account.
3. Accept the initial terms; skip account recovery, two-factor enrollment, and free-trial setup.

![Google Cloud navigation and search](https://cdn.qwiklabs.com/9Fk8NYFp3quE9mF%2FilWF6%2FlXY9OUBi3UWtb2Ne4uXNU%3D align="center")

### Activate Cloud Shell

Click **Activate Cloud Shell** in the console toolbar, then click **Continue** and **Authorize** as needed.

![Activate Cloud Shell icon](https://cdn.qwiklabs.com/ep8HmqYGdD%2FkUncAAYpV47OYoHwC8%2Bg0WK%2F8sidHquE%3D align="center")

Optionally inspect the active account and selected project:

```bash
gcloud auth list
gcloud config list project
```

The outputs will show the temporary project and account assigned to **your own** lab session. Never publish their credentials.

## Task 1. Verify the data in the source instance for migration

### Verify the data in the source instance

1. In the navigation menu, open **Compute Engine → VM instances**.
2. Locate `pg14-source` and click **SSH**.
3. Open the PostgreSQL shell:

```bash
sudo -u postgres psql
```

4. List available tables:

```sql
\dt
```

5. Count records in each of the six HR tables:

```sql
SELECT count(*) AS countries_row_count FROM countries;
SELECT count(*) AS departments_row_count FROM departments;
SELECT count(*) AS employees_row_count FROM employees;
SELECT count(*) AS jobs_row_count FROM jobs;
SELECT count(*) AS locations_row_count FROM locations;
SELECT count(*) AS regions_row_count FROM regions;
```

The reference dataset supplied with the lab has:

| Table | Rows |
| --- | --: |
| countries | 25 |
| departments | 27 |
| employees | 107 |
| jobs | 19 |
| locations | 23 |
| regions | 4 |

Leave the PostgreSQL shell, then close the source VM session:

```sql
\q
```

```bash
exit
```

## Task 2. Create and start a continuous migration job

**In this four-task edition, creating the source connection profile is part of Task 2**, not a standalone task.

### Get connectivity information for the PostgreSQL source instance

In **Compute Engine → VM instances**, locate `pg14-source` and copy its **Internal IP**. This address belongs to the temporary lab project; don't substitute an internet-facing IP.

### Create a new continuous migration job

1. Search for **Database Migration** in the Google Cloud console.
2. Open **Migration jobs → Create migration job**.

### Get started

Fill the wizard with these values and leave the other settings at their lab defaults:

| Field | Value |
| --- | --- |
| Migration job name | `postgres-to-alloydb` |
| Source database engine | PostgreSQL |
| Destination database engine | AlloyDB for PostgreSQL |
| Destination region | Region of the lab's existing target cluster |

Click **Save & continue**.

### Define a source

1. Under **Select source connection profile**, choose **Create a connection profile**.
2. Enter the source connection profile settings:

| Field | Value |
| --- | --- |
| Profile type | Source |
| Connection profile name | `pg14-source` |
| Source database engine | PostgreSQL |
| Destination database engine | AlloyDB for PostgreSQL |
| Region | Region assigned to the lab resources |
| Connectivity / host | Source VM **Internal IP** |
| Port | `5432` |
| Username | `postgres` |
| Password | Temporary lab database password (provided by the official lab) |

3. Click **Define** or expand the PostgreSQL connection settings as shown in the current console wizard, then **Save** and **Create** the connection profile.
4. Confirm that `pg14-source` is selected as the source, and click **Save & continue**.

### Test completed task

Click **Check my progress** for **Create a connection profile for the PostgreSQL source instance**. This is the **first graded checkpoint (50 points)** in this edition.

### Define the destination

1. Choose **Existing cluster** as the target type.
2. Select `alloydb-target-cluster`.
3. Choose **Select & continue**.
4. When prompted, enter `alloydb-target-cluster` to confirm the destination and continue.

**Do not create another AlloyDB cluster** or promote the target: the lab uses the provided target for continuous replication.

### Define the connectivity method

Select **VPC peering**, then click **Configure & continue**.

### Configure the migration databases

Select **All databases**, then **Save & continue**.

### Test and create the migration job

1. On the final wizard page, click **Test job**.
2. Once the test succeeds, click **Create & start job**, and confirm if prompted.
3. Do not select the create-only option: the migration must actually start.

### Review the status of the continuous migration job

Open the `postgres-to-alloydb` job details page. Its state should advance from **Starting** to **Running**; once the initial data copy finishes, the phase should display **CDC**. Do not proceed to data comparison before the continuous replication phase is available.

## Task 3. Confirm data load in the AlloyDB for PostgreSQL instance

### Check the AlloyDB for PostgreSQL instance

1. Open **AlloyDB for PostgreSQL → Clusters**.
2. Select `alloydb-target-cluster`, then open **Connectivity** and record its **Private IP**.
3. If the copied address contains `:5432`, remove the port suffix; the next command requires only the IP address.
4. Return to **Compute Engine → VM instances**, and click **SSH** for `alloydb-client`.
5. On the client VM, set the target IP using your observed private address:

```bash
export ALLOYDB="YOUR_ALLOYDB_PRIVATE_IP"
echo "$ALLOYDB" > alloydbip.txt
psql -h "$ALLOYDB" -U postgres
```

6. Enter the temporary database password from the official lab when prompted.
7. In `psql`, list the imported tables and compare their row counts to the source:

```sql
\dt
SELECT count(*) AS countries_row_count FROM countries;
SELECT count(*) AS departments_row_count FROM departments;
SELECT count(*) AS employees_row_count FROM employees;
SELECT count(*) AS jobs_row_count FROM jobs;
SELECT count(*) AS locations_row_count FROM locations;
SELECT count(*) AS regions_row_count FROM regions;
```

All six tables should be present, initially with the same counts as in Task 1.

8. Inspect the `regions` table:

```sql
SELECT region_id, region_name FROM regions;
```

At this stage there should be four original regions. **Leave the client terminal and psql session open** for the CDC check in Task 4.

## Task 4. Propagate a live update to the AlloyDB instance

This step checks that the continuous migration job also replicates updates made **after** the initial copy.

1. In Compute Engine, reopen SSH to the source VM `pg14-source`.
2. Launch `psql` on the source VM:

```bash
sudo -u postgres psql
```

3. Insert the new region requested by the lab:

```sql
INSERT INTO regions VALUES (5, 'Oceania');
SELECT region_id, region_name FROM regions;
```

The source should now list **five** rows, including `5 | Oceania`.

### Review data in the AlloyDB for PostgreSQL instance

1. Return to the still-open `alloydb-client` `psql` session.
2. Query the destination:

```sql
SELECT region_id, region_name FROM regions;
```

After CDC catches up, the destination should also show **five** rows and the new `Oceania` entry. If it has not appeared yet, wait for replication and repeat the query.

### Test completed task

Click **Check my progress** for **Test the continuous migration of data**. This is the **second graded checkpoint (50 points)**. The commands and reference outputs shown here do not represent a grader PASS from a live run.

## Congratulations!

You have configured a source profile, run a PostgreSQL-to-AlloyDB continuous migration, compared source and destination records, and checked an additional replicated update.

---

## Solution of Lab

### Quick

```bash
curl -LO raw.githubusercontent.com/ePlus-DEV/storage/refs/heads/main/labs/GSP1084/lab.sh
source lab.sh
```

---

### Manual

%[https://www.youtube.com/watch?v=Fvy6KeeplLY]

%[https://www.youtube.com/watch?v=64HnBEnN3jk]

## Sources and verification

- [Original Google Skills GSP1084 lab](https://www.skills.google/focuses/100854) — four-task edition, *Manual Last Updated August 10, 2026*.
- [Google Cloud documentation: Migrate to AlloyDB for PostgreSQL using DMS](https://docs.cloud.google.com/alloydb/docs/migrate-with-dms)

**Documentation note:** Task order, wizard fields, two checkpoints, example counts, and the two linked Google Skills interface images are matched to the August 2026 four-task edition. Text is paraphrased rather than copied verbatim. The image links reference Google's original CDN and remain subject to their availability and reuse terms. **Runtime/grader unverified in this session.**

Copyright (c) 2026 ePlus.DEV | Unlicense (applies to original guide and automation, not Google's images or documentation).