?

Pure FlashBlade Replication Base Setup

5:57
87
Plays

This video is part one of a video series where WWT resources will walk through the base network configuration and array connection establishment for two FlashBlade systems in the Advanced Technology Center (ATC).

Pure Storage and WWT are working together to help showcase the latest technology advancements in the Advanced Technology Center.  

We are excited to have worked with Pure Storage to upgrade our FlashBlade systems to Purity //FB 3.0. This software release has opened up important new features that our joint customers have been waiting for. The most significant feature in this release is File and Object Replication.

New Features


File Replication 
Purity//FB 3.0 now offers FlashBlade file replication. FlashBlade file replication enables the asynchronous snapshot-based replication of data from a source FlashBlade array to a target FlashBlade array. The primary use case for file replication is disaster recovery, which enables ongoing protection against site or data center failures. 

File replication is configured and managed using the Purity//FB GUI, CLI, and REST API. Note the following: 

  • File systems created on arrays running a pre-Purity//FB 3.0.0 version are not eligible for replication. 
  • Only snapshots created as part of the replication process are transferred.
  • Encryption over the wire is not yet supported.

Note that when connecting arrays: 

  • The management IP address required during configuration cannot be a loopback IP address.
  • An IP address must be specified for the replication address; specifying the hostname/FQDN is not allowed.
 
Object Replication 
Purity//FB 3.0.0 introduces object replication. Two FlashBlade arrays can be connected, or a FlashBlade and an AWS S3 target can be connected, allowing asynchronous object replication from a bucket on the array to a bucket on the second array or S3 target. With object replication, object are automatically replicated when they are written to the source bucket. 

S3 targets that are not AWS are not supported. Object replication is configured and managed using the Purity//FB GUI, CLI, and REST API. Note the following: 

  • Objects that were created before the creation of the object replica link are not eligible for replication.
  • Object versioning on the target bucket must be enabled.

Note that when connecting arrays, the management IP address required during configuration cannot be a loopback IP address. Cascading replication topology is not supported. 
 
File System Rollback 
Active file systems can now be restored to a previous point in time from a snapshot. When restoring a file system, data is "rolled back" to a snapshot – the data from the snapshot overwrites the contents of the file system. Any snapshots newer than the snapshot to be used as the restore point must be destroyed and eradicated prior to performing the restoration. Any data lost since the last snapshot is also lost. 

Restoring a file system from a snapshot will overwrite the contents of the file system with data from the snapshot, and may disrupt workloads running against the file system. Stopping all I/O, including reads, writes and all metadata operations against the file system is recommended. 

Note that when file system is rolled back to an older snapshot, the quota limits remain unchanged (i.e. the quota limits are not part of the snapshot). 

File system rollback is available for snapshots created on a FlashBlade array running Purity//FB 3.0.0 or later.