3.7.61

New Features and Improvements

System

• Added the ability to store filesystem metadata on an SSD pool. This feature, known as "packed metadata", is useful for large filesystems in an HDD pool to speed up metadata operations.
  

  • Contact StorONE Support for assistance enabling this feature.
    

  • This feature is applicable only for new volumes created after the feature is enabled. You cannot migrate the filesystem metadata on an existing volume.
  • This feature is not currently supported with replication. Volumes created using this feature cannot be replicated, and the feature cannot be removed after volume creation. This means that if you want to replicate a volume that is using this feature, you will need to disable the feature, create a new volume, and replicate the new volume. You can migrate the data from the old volume to the new volume before or after configuring replication.
    
• Updated the garbage collection (GC) mechanism to reduce the frequency of full (blocking) garbage collections, and improve GC performance.
• Improved SCSI timeout handling to return BUSY to initiators instead of aborting tasks, allowing initiators to more quickly retry commands. This improves overall system performance and reliability, and reduces VMWare heartbeat timeouts that generate datastore disconnect/reconnect events.
  
• General performance improvements.
  

High Availability

• Improved handling of configuration and management operations when one node in an HA deployment is down.
  
• Set internal sync flag for NAS volume mounts to reduce HA transition time for the NAS server.

GUI

• Increased the size of text boxes to improve readability and usability.
• Added a link to the documentation site on the controller landing page.

CLI

• Successful asynchronous commands (commands run with the & operator) are now displayed in green.

Replication

• Added and improved messages when starting replication of an encrypted volume, including notifying user when the encrypted volume is locked.

Object store

• Updated internal concurrency and timeout settings to improve object store performance and reliability.
  
• Reduced the frequency of internal scanning processes.

Filesystem

• Increased the timeout for creating filesystems to 6 hours. This allows for creating larger filesystems on HDD-only volumes. StorONE recommends against creating HDD-only volumes when possible. For better performance, create volumes with an SSD upper tier.

Snapshots

• Improved snapshot deletion scheduler to always delete a single snapshot at a time, and to allow parallel snapshot deletions if the volumes reside on different controller nodes. This improves overall system performance and reduces resource contention.

Fixed Issues

System

• Fixed an issue with reporting the status of the support tunnel.
• Fixed a rare issue where different configuration changes had the same version ID.
• Fixed an issue with vRAID+ volumes (volumes striped across multiple drive pools) that would not load after a restart.
• Fixed an issue with modifying the tier size on vRAID+ volumes.
• Fixed an issue handling memory buffers that caused volume shutdown operations to hang (including HA transitions and volume delete operations).

High Availability

• Fixed error handling for unmounting filesystem volumes during an HA transition.

OmniUI (GUI, CLI, and REST API)

• Fixed issues with generating support packages.

• Fixed an issue where the 'Tier Data Degradation Histogram' was not displayed for vRAID+ volumes.
• Fixed name validation for floating IP to be case-insensitive.

GUI

• Fixed an issue preventing deleting host mappings using the web UI.

• Fixed an issue where clicking the green '+' icon to add a consistency group during the volume creation wizard opened the wrong modal.

CLI

• Fixed the misaligned CLI login animation.
• Fixed an issue where commands failed with the error Status: Failed To Connect To Service.

Replication

• Fixed an issue with replication transitions when a default gateway is defined on the replication floating IP.
• Fixed a race condition in HA failover on systems with replicated volumes.
• Fixed an issue where adding or modifying hosts (including host mappings) caused the system to lock up.

Network Attached Storage (NAS)

• Fixed issue with excessive memory consumption when using the VMware vSphere® Storage APIs – Array Integration (VAAI) plugin for NFS datastores.
  

Notifications

• Fixed an issue where snapshot retention notifications were sent for volumes without snapshots configured.

Behavior Changes

System

• The system now enforces a minimum of 40GB for the lower evacuation threshold.

OmniUI (GUI, CLI, and REST API)

• The application volumes list --details output now includes the encryption type for encrypted volumes.
• The applications filesystem list now includes the IsPackedMetadata attribute to display whether a filesystem is using the new "packed metadata" feature.