How EZswitch works on standalone instances

This topic describes how the Zesty Disk EZswitch feature works in the standalone environment:

• EZswitch on Linux

• EZswitch on Windows

EZswitch on Linux

EZswitch on Linux automatically creates a new Zesty Disk filesystem and transfers existing files to it.

While EZswitch is doing its magic, the mount point name will change temporarily. The original mount point name will remain active so your applications can read and write throughout. When EZswitch is done copying data, the original name will return, and it and the data will be on the Zesty Disk filesystem.

These changes take place seamlessly throughout the process, without your need to do anything. We will point out when the changes take place in the following description.

After the first restart, the Zesty Disk starts at a size that may be much smaller than the size of the original filesystem. As more data is copied, the Zesty Disk expands to accommodate the data as well as to ensure that you are always covered for new writes.

EZswitch requires two machine reboots:

After starting EZswitch, preparations are done (like creating a snapshot and the new Zesty Disk volume), and then:

• After the first restart, the data is transferred to the Zesty Disk.

• After the second restart, cleanup is done.

  You can schedule the restarts to be done automatically (during a maintenance window if you prefer) or you can do it manually, from the platform, the instance, or the AWS console.

  If you schedule automatic restarts, you can override that by manually restarting, and vice-versa.

As part of the migration, EZswitch creates a data snapshot to enable a restore from the snapshot.

For an in-depth look on how this works during EZswitch, see Behind the scenes at EZswitch.

Best practice recommendation: After the existing data is fully transferred to the Zesty Disk, we recommend that you remove the old AWS volume and the EZswitch-initiated snapshots so that you no longer pay for them. You can do this automatically during the EZswitch setup.

Limitations

In addition to the general System requirements, EZswitch does not support the following:

• Parallel migration of multiple filesystems on the same instance

• Instances with a Linux filesystem based on Logical Volume Manager (LVM)

• Instances running RHEL v8.0 or v9.0 or Rocky v9.3 or v9.4

• Instances running applications whose filesystem uses very large files that need to be opened all together, such as MSSQL

This list is not exhaustive and is subject to change without notice.

You can run EZswitch in either of the following ways:

• Run EZswitch for Linux from the Zesty platform

• Run EZswitch for Linux from the zestyctl CLI

Behind the scenes at EZswitch on Linux

The EZswitch feature creates a new Zesty Disk filesystem on a clean volume and then migrates data from an existing filesystem to it.

EZswitch requires two restarts. Other than during the actual restart periods, your application can continue performing reads and writes to the original mount point.

To make this work, the following goes on behind the scenes:

1. After beginning EZswitch, and before the first restart, the following take place:

   1. A snapshot of the existing volume is created so you can restore, if necessary.
      If snapshot creation takes longer than 5 minutes, EZswitch will continue, while the snapshot is being created in the background.

   2. The existing filesystem mount point is being prepared for the next reboot.

2. After the first restart, the following takes place:

   • Two new mount points are created:

     • New: This is the new Zesty Disk mount point.

     • Old: This is simply a different name for the original mount point.

     Note: The original mount point still works for the application using overlay technology.

   • All writes take place to New.

   • Data from the original mount point is copied in the background from Old to New. New starts with a very small capacity. It expands as more data is copied or written to it.

   • Reads first look to New. If the data isn't there (it hasn’t been copied yet), the reads look to Old.

   • After all the data is copied to New, all reads/writes will go to New, and the system is ready for the next restart.

3. After the second restart, the following takes place:

   • Old is no longer shown in the Managed Filesystems tab and its original volume is detached from the instance (shown as “Available” in AWS).

     Note: It is recommended that you delete this volume from AWS so as not to incur further costs. You can do this automatically (when starting EZswitch) or manually.

   • New disappears from the Managed Filesystems tab. Its mount point name returns to the name of the original mount point (now as a Zesty Disk).

EZswitch on Windows

EZswitch on Windows converts the filesystem on a current disk to a Zesty Disk-managed filesystem. The data remains in place; just the filesystem type is changed.

You run EZswitch with a single PowerShell command (not from the Zesty platform), and it does not require any instance downtime.

You can run EZswitch on any disk with a basic volume in the default Windows filesystem (NTFS).

After you run EZswitch, the filesystem will expand and shrink like any other Zesty Disk filesystem. The size of the filesystem base volume, however, cannot be changed. For this reason, if the base volume is very large, before running EZswitch, you may want to copy the data to a smaller volume and then run EZswitch on the smaller volume.

To run EZswitch on Windows, see Run EZswitch on Windows.