When an organization decides to replace a production server, a successful server migration is crucial to the business and must be planned carefully.
Before proceeding with the migration, there are some steps to follow and some requirements to meet to avoid nasty surprises such as data loss or service disruption.
The migration procedure involves the following steps:
- Understand what server migration is
- When to migrate
- Time required to complete a migration
- Collect data and applications to migrate
- New server installation
- Data transfer
- Test the new server
- Keep old server
What is server migration?
The server migration is the process of moving data from an old server to a new one and involves the selection of a new server to use and its configuration.
The reasons for a server migration can be summarized as follows:
- Obsolete hardware no longer supported by the vendor
- Insufficient space to store data and applications
- Insufficient resources in terms of CPUs and memory
- Need to improve performance
Because the server migration will cause some downtime for data accessibility and services, you need to plan carefully the entire procedure to ensure all data are properly transferred.
Using the Nakivo's monitoring tool, you can have an overview of your infrastructure health highlighting what servers are running out of resources and what server should be considered for a replacement.
When to migrate
The migration should be well planned to limit data unavailability preventing possible issues. You should consider the following scenarios before scheduling the server migration:
- Because the migration will interrupt data accessibility, you should schedule the procedure outside working hours.
- If the business is facing a critical time or is entering in a critical time (i.e. billing), perhaps it is not a good moment to migrate.
- Migration can be performed once business managers have been notified about the migration, what to expect, changes, and possibile services disruption.
- Documentation has been updated with all changes and the new configuration procedure.
- The source server backup has been taken and tested. For example, Nakivo Backup & Replication can test the backup ensuring the recoverability of the backup data.
Time required to complete a migration
Once the migration plan is complete, the effective time required for the data transfer process depends on several factors:
- The migration heavily depends on the amount of data to transfer and the available bandwidth. Hard to say exactly how long it takes, can be a matter of minutes or hours.
- If you also have applications to transfer, time required to finalize the procedure will take longer to complete.
Collect data and applications to migrate
Before proceeding with the migration, you should collect the list of data and applications to migrate. This avoid forgetting to migrate specific data or applications.
New server installation
For a successful migration, the new server must be properly sized in terms of space and resources. The new server must have enough room, CPU and memory to satisfy current resource demand and future grows. For the initial migration you should consider at least an extra 20% of space in the new server for application installation, updates and so on.
The new server must be also scalable to satisfy the possible increase of resource demand.
Before migrating, you should take a backup of the server to have a copy of your data in case something wrong happens during the migration. Regardless if the source server is physical or virtual, the backup can be taken using software like Nakivo Backup & Replication configuring the appropriate Backup Job.
The available bandwidth and installed NICs will heavily affect the data transfer speed. A large bandwidth with 10 GB NICs will perform better in terms of speed compared to a narrow bandwidth with 1 GB NICs.
A faster data transfer will reduce the server downtime ensuring a limited unavailability of data and services.
Test the new server
Once all the data have been migrated to the new server, you must test if everything works as expected before putting the new server on production.
The test is a crucial part of the migration process because you need to ensure that all data are accessible and services are working:
- Check if installed services are up and running.
- Both applications and dependencies work as expected.
- Clients can reach the server confirming that the network configuration is correct.
- Compare the previous operative scenario with the old server to discover potential misconfigurations.
- Check the Event Viewer for possible errors.
Keep old server
Once the migration procedure has been completed, keep the old server for at least one month before deleting. If any problem arises with the new server, you always have the chance to fallback providing a protection against unforeseen issues.
The backup of your replaced server can be stored in a supported cloud storage (i.e. Azure, AWS, Wasabi) for a long period without consuming your on-prem backup repository space. It can be deleted later when you are 100% sure that the new server works as expected without any issue.
Server migration can be tricky, time-consuming, and demands considerable effort from IT staff. A good analysis and planning make the difference between a successful migration or a global failure.