Legacy Block Storage Migration FAQ

Updated on January 24, 2024
Legacy Block Storage Migration FAQ header image

What is a "Legacy" block storage volume and how do I know if I have one of these subscriptions ?

Some Vultr customers have Block Storage volumes labeled as type "NVME Legacy" in the Block Storage section of the Vultr customer portal. These will be at our New Jersey location, and they represent data belonging to the original block storage product before it was redesigned and reimplemented.

What action is required for these legacy block volumes?

The data in all of these Legacy block storage volumes must be migrated to the new block storage product, preferably by the customer themselves as this causes far less disruption.

We strongly advise all customers with these subscriptions to migrate their data out of these volumes as soon as they are able to do so. The deadline for the completion of all customer-led data migrations is June 1, 2023.

How is the replacement NVME New Jersey block storage product better and does the new product cost more?

The modern NVMe cluster is superior in the following three areas:

  1. Higher performance
  2. Greater total capacity
  3. Better stability

Note that there is a large measurable difference on all three of these elements; this is an improvement you will notice.

The cost of the replacement NVME block storage is exactly the same as what you are already paying for the legacy subscription(s), $1 per 10 GB.

Why should I migrate this data as soon as feasible ?

There are at least four key reasons :

  1. If you do not migrate this data, the VPS instance which uses the legacy block subscription will have downtime when Vultr inevitably has to perform the migration later.
  2. Stability, availability, and reliability are greatly improved in the new product.
  3. To take advantage of the higher performance of new block storage.
  4. It is necessary to expand the size of legacy block volumes.

What advantages does a customer-executed manual migration have over a Vultr-executed migration path?

There are four substantial advantages:

  1. It allows you to develop and execute a migration plan. This leads to minimizing the downtime of the VPS instance which uses each legacy block volume.
  2. You have control over the migration date and time window.
  3. Only data which truly needs to be copied can be transferred, instead of literally all data blocks composing the volume.
  4. We offer SLA credit eligibility for redundant block subscriptions used in the migration process (a maximum of one month's credit per legacy block subscription).

Please also read the next section to better understand why you almost certainly do not want to wait for Vultr to migrate your data.

For customers who do not migrate their legacy data on time, what happens?

If you do not complete this data migration process by our third and final deadline of June 1 2023 , Vultr reserves the right to migrate this data during any time period of its own selection. Note that if Vultr conducts migrations of this data, it will result not only in the temporary unavailability of the legacy block volume(s), but any VPS instance(s) which are clients of those volume(s) will also be shut down for the migration window!

Due to this being announced far in advance, there will be no SLA credits of any kind issued for any downtime caused by legacy block migrations after this final deadline.

Furthermore, after the final deadline has passed, Vultr will be prioritizing the migrations of data as well as data preservation far higher than the ordinary client user I/O operations. In effect, this means a combination of lower performance on legacy volumes and a greater amount of scheduled maintenance affecting legacy data online availability .

Please regard this matter with high concern. Some customers still have a very considerable amount of data stored in the legacy product, and the required downtime for a Vultr-led migration in that case could take many hours or even days .

How would I manually migrate each legacy block volume?

Here is an example process workflow :

  1. If you do not have any off-site backups of your block storage volume's data, first create one. This is worthwhile standard practice in any situation.
  2. Locate the VPS which is using your legacy block storage volume in the customer portal, under the Products --> Block Storage tab.
  3. Login to that VPS and determine what is using the block storage volume. Shutdown any processes that are accessing the volume, and then unmount/disassociate it from the operating system so that no further accesses can occur.
  4. Create a new block storage volume of the same size in New Jersey. The NVME block product is the same price you are currently paying and has substantially superior performance, so we recommend that over HDD block storage.
  5. Attach the new block storage volume to the same VPS as the old one. It will appear as an additional disk device.
  6. Verify that the old block device is not being accessed by the VPS, then copy the data from the old block storage device to the new block storage device. This can be done in a variety of different mutually-exclusive ways. One method would be to do a raw block-level copy directly between the two devices. Alternatively, creating a filesystem on the new device manually allows for either a metadata-preserving recursive file copy, or using specialized file copy tools like rsync.
  7. Activate and/or mount the new block device, and validate all the data was transferred properly.
  8. Detach the old block storage volume in the Vultr control panel. The old subscription can be closed, but we highly recommend validating your off-site backup and new live volume copy first.

The exact commands required will differ by VPS operating system (Linux, BSD, Windows), as well as what type of backup(s) and copies you intend to create. For example, some relevant commands for various Linux approaches include lsblk, umount, mount, tar, cp, dd, sftp, rsync, sync.

Whichever method and tools you use, make sure to check the relevant documentation and test some example commands first. There are typically specific command flags that are needed to generate an exact copy (including preserving metadata like timestamps and permissions).

Can you provide an ordered timeline of relevant events for reference?

  • 2022 Q1: New block storage product is deployed and tested.
  • 2022 Q2: All newly made block storage subscriptions set to use only the newer product.
  • 2022 Sept: Customers notified about the conversion of the old product to a legacy line, including the need to migrate this data.
  • 2022 Oct: First version of the Customer Legacy Block Storage FAQ published.
  • 2022 Nov: Resizing of legacy block volumes is disabled.
  • 2022 Dec: Initially requested completion target for customer-led legacy data migrations.
  • 2023 Feb 21: Further notifications to customers after legacy product performance issues.
  • 2023 Apr 1: Second requested target for customer-led data migrations.
  • 2023 Jun 1: Third and final deadline for all customer-led data migrations.
  • 2023 Q3: Vultr migrates all remaining data with any required downtime.
  • 2023 Q4: Legacy block storage will be fully decommissioned.