This page describes how to run a Python3 script that automates the migration of your group of individual instances to a stateful MIG.
If you want to perform the same steps manually, follow this tutorial.
The automated script performs the following steps to migrate your instances:
- Stop all instances
- Create disk image for boot disk if needed
- Create an instance template based on the properties of a chosen instance, except for attached data disks.
- Create an empty MIG.
- For each instance in the original group, perform the following steps:
- Clone all instance disks except the boot disk.
- Create an instance in the MIG based on the instance template from point 1, and include the cloned disks from the source instance.
- Print commands for cleaning up the source instances after you have verified that the stateful MIG serves your needs.
Note that the script leaves all standalone VMs stopped with their disks intact, for easy reverting if the MIG doesn't work as expected. This results in additional costs, for the following reasons:
- The script doesn't detatch or delete the original disks.
- The script creates images from existing disks.
- All source instances must have the same instance configuration.
- Boot disks must be stateless.
- Preservation of IP addresses is not supported.
- The script stops all the running standalone source instances.
- The script doesn't reuse the standalone VM names in the MIG.
To run this script you need to meet the following requirements:
- Python version >= 3.6.
pipenvor other virtual environment manager to install the script dependencies.- Compute Engine API enabled.
To authenticate your requests, download a service account JSON keyfile and point to it using an environment variable:
export GOOGLE_APPLICATION_CREDENTIALS="/path/to/keyfile.json"
Alternatively, use the following gcloud command:
gcloud auth application-default login
- Clone the repository and move to the folder with the script.
git clone https://github.com/GoogleCloudPlatform/compute-migration-scripts.git && cd compute-migration-scripts/instances-to-stateful-mig
- Install virtual environment and dependencies.
python3 -m venv env
source ./env/bin/activate
pip3 install -r requirements.txt
python3 migrate_script.py [-h] [-p PROJECT] [-s SOURCE_INSTANCES [SOURCE_INSTANCES ...]] [-b BASE_INSTANCE_NAME] -z SOURCE_INSTANCE_ZONE -m MIG_NAME [--regional]
optional arguments:
-h, --help show this help message and exit
-p PROJECT, --project PROJECT
-s SOURCE_INSTANCES [SOURCE_INSTANCES ...], --source_instances SOURCE_INSTANCES [SOURCE_INSTANCES ...]
-b BASE_INSTANCE_NAME, --base_instance_name BASE_INSTANCE_NAME
-z SOURCE_INSTANCE_ZONE, --source_instance_zone SOURCE_INSTANCE_ZONE
-m MIG_NAME, --mig_name MIG_NAME
--regional
| Short | Long | Default | Description |
|---|---|---|---|
-h |
--help |
show this help message and exit | |
-p |
--project |
project ID or project number of the GCP project you want to use. | |
-s |
--source_instances |
list of your single GCP instances you want to migrate. | |
-b |
--base_instance_name |
source_instances[0] | base GCP instance name (the template will be based on this instance) |
-z |
--source_zone |
zone name of the GCP instance you want to migrate. | |
-m |
--mig_name |
name of the stateful MIG you want to create. | |
--regional |
False | if provided, will create regional stateful MIG, which deploys instances to multiple zones across the same region | |
--image_for_boot_disk |
False | if provided, will create disk image for boot disk of base GCP instance |
Show the help text and exit.
Project name of the GCP instance you want to migrate. If project is not provided, it will be taken from credentials.
List of the individual GCP instances you want to migrate. You should provide at least 1 instance.
Base GCP instance name. The template will be based on this instance. If skipped
then the first instance from source_instances will be taken.
Zone name of the GCP instance you want to migrate.
Name of the stateful MIG you want to create.
If this flag is set, then stateful MIG will be
regional,
which deploys instances to multiple zones across the same region. If the flag
isn't set, then stateful MIG will be zonal, which deploys instances to a single
zone. Instance redistribution type will be set to NONE. You cannot change
instance redistribution for stateful MIGs. See Limitations
If this flag is set, then script will create disk image for boot disk of base GCP instance.
python3 migrate_script.py -s instance-1 instance-2 instance-3 -z us-central1-a -m my-mig --image_for_boot_disk
Instance instance-1 is not stopped. Stopping ...
Instance instance-1 stopped
==========
Instance instance-2 is not stopped. Stopping ...
Instance instance-2 stopped
==========
Instance instance-3 is not stopped. Stopping ...
Instance instance-3 stopped
==========
Creating disk image for boot image instance-1 ...
Disk image instance-1-image-ed0d24 created
==========
Creating base instance template ...
Instance template instance-1-template-004cb8 created
==========
Creating empty MIG my-mig...
MIG my-mig created
==========
Creating disk a-disk-1-bb2a30 from disk a-disk-1
==========
Creating disk a-disk-2-f7344b from disk a-disk-2
==========
Adding instance instance-1-b7273d to my-mig MIG
==========
Creating disk a-disk-3-7498a2 from disk a-disk-3
==========
Adding instance instance-2-6475c6 to my-mig MIG
==========
Adding instance instance-3-17d7c7 to my-mig MIG
==========
Migration successfully finished. Time spent: 457 seconds.
Use the following command to delete the individual source instances:
* gcloud compute instances delete instance-1 instance-2 instance-3
To revert all changes, use this clean up commands:
* gcloud compute instance-groups managed delete my-mig
* gcloud compute instance-templates delete instance-1-template-004cb8
* gcloud compute disks delete a-disk-1-bb2a30
* gcloud compute disks delete a-disk-2-f7344b
* gcloud compute disks delete a-disk-3-7498a2
* gcloud compute images delete instance-1-image-ed0d24
Futher steps:
- Configuring autohealing: https://cloud.google.com/compute/docs/tutorials/migrate-workload-to-stateful-mig#configuring_autohealing
- Using a stateful policy instead of per-instance configurations:
https://cloud.google.com/compute/docs/tutorials/migrate-workload-to-stateful-mig#using_a_stateful_policy_instead_of_per-instance_configurations
- Adding more VMs:
https://cloud.google.com/compute/docs/tutorials/migrate-workload-to-stateful-mig#adding_more_vms