How to migrate workloads on OpenStack using Hystax Acura

Overview

Hystax Acura offers a fully automated and reliable way to lift-and-shift all types of OpenStack workloads without experiencing downtime or complications arising from large-scale migration. Hystax supports all popular OpenStack distributions including UKCloud's Red Hat OpenStack.

Hystax offers a sustainable solution to achieve zero downtime during the OpenStack migration process. The instance keeps running throughout the migration, which is important when it's not an option to stop the applications running on the instance. Overcome common transfer challenges like high costs, long transfer times and security concerns – all in a single product. Lifting-and-shifting workloads to a chosen OpenStack distribution is more manageable than ever before.

This article provides instructions from the Hystax Acura Cloud Migration Step-by-Step Guide for how to migrate workloads using Hystax Acura.

Prerequisites

To utilise Hystax Acura cloud migration services, you'll first need a Hystax user account provisioned in the Hystax Acura Control Panel. You can request a Hystax user account by raising a ticket in the My Calls section of the UKCloud Portal.

Understanding concepts

Before starting to use Hystax Acura, it's helpful to understand a few commonly used terms and concepts.

Hystax Acura Control Panel

Hystax Acura Control Panel is a management portal deployed and managed by UKCloud. The Hystax Acura Control Panel enables an OpenStack user to create and manage all their workload migration requirements.

Note

All Hystax components are deployed and contained within UKCloud's infrastructure. No customer data is shared or transmitted to any third party.

Hystax Acura agents

These are the agents that are installed within a guest Linux or Windows operating system, which communicate with the Hystax Acura Control Panel to coordinate and manage workload migrations.

Migrating OpenStack workloads

To perform cloud migration you need to complete the following steps:

Replication flow

1. Login to the Hystax Acura Control Panel and navigate to a customer dashboard for a customer you target to perform migration.

   

2. On a customer dashboard you can find information about replicated machines, Migration Plans and Migrations running on a target cloud. By default, no resources are available for a new customer.

   

3. Click Download agents in the main menu, select a customer for which machines need to be replicated and download a replication agent.

   

   For OpenStack workload migrations there are two types of replication agents:

   • Windows agent - an internal agent that can be deployed to any number of customer Windows virtual or physical machines. Refer to http://docs.hystax.com/devices_protection.html#install-windows-and-linux-agents and the Operating Systems Compatibility Matrix.

   • Linux agent - an internal agent that can be deployed to any number of customer Linux virtual or physical machines. Refer to http://docs.hystax.com/devices_protection.html#install-windows-and-linux-agents and the Operating Systems Compatibility Matrix.

4. When any of the agents are deployed, the machine with the Hystax agent will appear in Hystax Acura Control Panel under customer dashboard in the Discovered state.

5. To start replication of machines, select the machines and click Start replication under Bulk actions or for any of the machines specifically. After Start replication is performed, machines will be replicated to a target cloud, first, Hystax Acura takes full replica and, later, incremental replicas according to a schedule set globally for all the machines or for any group of machines or for any machines specifically.

   

6. To change the replication schedule, navigate to Replication schedule under a machine settings or bulk actions. Refer to http://docs.hystax.com/dr_overview.html#edit-replication-schedule for more information on configuring the settings.

Run migration flow

1. Migration is based on Migration Plans, which are scenarios of what needs to be migrated and in what order. Refer to http://docs.hystax.com/dr_overview.html#disaster-recovery-plans for more information about Migration Plans. To create a Migration Plan click Add on Migration Plans region.

   

2. While creating a Migration Plan, orchestration is available under Basic and Expert modes. Basic mode helps to generate a plan with standard settings, expert mode is a JSON script with more powerful settings described under http://docs.hystax.com/dr_overview.html#dr-plan-syntax.

3. Add all the machines which need to be migrated, place them into ranks, set correct IP addresses, networks and boot delays.

   

4. When the Migration Plan is ready, you can run test or production failovers. To run a migration, select a Migration Plan and click Run Migration.

   

5. Select a restore point for the migration by selecting Restore point time. For all of the machines selected the closest snapshot in the past will be taken. Click Run Migration to start the process.

   

6. When a migration is running, you can see and control its status on the Cloud Site page available via customer dashboard.

   

7. When you are done with test migration you can delete it and start a new one. For the final cutover, just detach Hystax Acura from the migration and Hystax Acura will stop tracking any resources belonging to this migration.

   

Known issues

The Hystax Acura service is under constant development. The following are the known issues in the current release of the software, which should be resolved in the next release.

Creation of new persistent volumes on target instances

When the Hystax migration completes, the disk of the resulting new/target instances will use persistent volumes and snapshots. This may result in extra monthly storage costs.

Migrating Ubuntu workloads

Due to the way in which the Hystax software works, you may experience issues migrating an Ubuntu server/instance with the default disk mount settings. In Ubuntu, the root risk (/) is mounted by label (that is, in /etc/fstab there is a setting that states LABEL="…." for the "/" disk).

You need to modify this to use the UUID or device name (for example, /dev/vdb1).

Run the following steps on a default Ubuntu setup for a successful migration:

1. Ensure /etc/fstab refers to the boot and root disks using UUID="" or /dev/vda1, and so on, and not by LABEL="" (you can retrieve the device name using the lsblk command and return the UUID using the blkid command).

2. After changing /etc/fstab (if required), run update-grub as the root user.

3. Install the Hystax source agent (hlragent) as per instructions given on the Hystax Acura server (Web UI).

4. Reboot the Ubuntu instance.

5. At this point the Ubuntu instance is ready to be replicated and migrated.

Migrating Centos/RHEL workloads

You may encounter an issue where the instance fails to boot in the cloud site after migration, and instead enters a dracut debug state. In this the case, in the source instance, run dracut -fMa qemu. Replicate and re-run the migration and the issue should be resolved.

Migrating Microsoft Windows workloads

Error when creating and running a migration plan in the Acura user interface

When creating and running a migration plan in the Acura user interface, you may encounter an error relating to a malformed mac-address. To resolve this go to the Expert tab in the Edit Migration Plan interface and remove any lines that contain ""mac": "…..". After doing this save and re-run the migration plan and the issue will be resolved.

User management

When a Microsoft Windows instance is migrated and first created on the destination cloud, the administrator user password will be reset to a random different value. Before a migration is run, ensure the source server has a local user with sufficient administration privileges that enable you to login and administer the server (possibly change the administrator password back if required).

Further support

Should you experience any issues in using Hystax Acura or with your migrations, our support team is on hand to assist via a Support Request in the My Calls section of the UKCloud Portal.

Feedback

If you find a problem with this article, click Improve this Doc to make the change yourself or raise an issue in GitHub. If you have an idea for how we could improve any of our services, send an email to feedback@ukcloud.com.