Primary Storage

 View Only

What’s New in Spectrum Scale 5.0.5?

By ISMAEL SOLIS MORENO posted Mon July 06, 2020 06:25 PM

  
Spectrum Scale 5.0.5 has been released, and there are a number of new features and improvements that are interesting to review.

One of the most important news is that IBM Spectrum Scale 5.0.5 is an Extended Update Support (EUS) release. An EUS release allows users to remain current on security and other fixes without having to perform a full functional upgrade. You can choose to take advantage of versions and releases while skipping mod packs and know even before GA, how long to expect support. This can help you to do better upgrade planning. For Spectrum Scale, every third release starting with 5.0.5, will have EUS with an additional 18 months of PTFs. The Spectrum Scale support lifetime is illustrated in the following diagram.



For more information on the IBM Maintenance Delivery Terminology such as “Version”, “Release”, “Mod Pack” and “Fix Pack”, please review the following page.

For more information on the IBM Spectrum Scale EUS and support lifetimes please refer to the following presentation.

The next sections of this blog entry, classify and summarize new functional capabilities and performance improvements of Spectrum Scale 5.0.5:

File system core

  • Windows Server 2019 NSD Server and NSD Client are supported. This enables you to have modern infrastructure on Windows OS such as an enhanced application platform.
  • IBM Security Key Lifecycle Manager (SKLM) 4.0 as well as Thales Vormetric Data Security Manager (DSM) 6.3 and 6.4 are supported as the Remote Key Management (RKM) server. This gives you a wider set of options to enhance your system data security.
  • Renewing expired client certificates without recreating a new client is now possible when using the IBM SKLM. This simplifies security management by reducing manual methods previously used.
  • Preserving IBM Spectrum Scale extended file attributes when copying files from one Spectrum Scale file system to another is now supported through the use of the –preserve option to the cp command. This makes file management easier and reduces the manual file attributes configuration across different file systems.
  • Thin provisioning improvements are added to space-reclaimable capabilities. Features such as online disk type change for space reclaim, autodetection on space reclaimable disk, and other methods to reclaim free space are now supported by IBM Spectrum Scale 5.0.5. This provides a wide selection of possibilities to improve your data allocation and make the most with your storage capacity.

File system performance

  • Performance is greatly improved for listing the files in a directory that contains a large number of fileset junctions. That means faster directory listing for file systems with many filesets.
  • Faster processing of quota checking for file systems with several independent filesets. Get faster info on space usage in a file system by user, group, and fileset.
  • Performance is improved for ESS file system restripe due to better self-tuning on the number of threads used. That means that for ESS you will be able to do faster rebalance or restore the replication of all files in a file system.
  • Performance is improved for mmap read workloads when multiple threads read the same file. This means faster concurrent in-memory reads.
  • Faster processing of snapshots by eliminating zero link count inodes. e.g. avoid moving unnecessary blocks when a snapshot is deleted. That means faster snapshots operations.

NFS and SMB

  • NFS monitoring improvements including the addition of client statistics, health monitoring to check the status of Ganesha, and DNS stats can be captured as part of authentication statistics collection. These tools can help you to collect a lot of data to improve your operation and performance.
  • SMB improvements include enhanced recovery lock to maintain cluster integrity, enhanced file handles for Mac clients, and better troubleshooting procedures for SMB shared modes. These enhancements look for better file handling and also to provide you the best practices to improve the operation of your SMB environments. 

Monitoring and health check

  • New commands to show the state of one or more nodes in the cluster and report the top processes. This provides you a better control per node.
  • New commands to verify the state of threshold rules for health monitoring. This will allow you a better way to monitor the system health and isolate problems caused by the violation of particular thresholds.
  • New health checking events added to sections including authentication, CES network, call home, and core GPFS. These events provide you fine-grained monitoring of your overall system.
  • Collect the top 10 processes per node and provide a report showing CPU and Memory. This will help you to identify in a detailed way, how your resources are being consumed.

AFM and AFM-DR

  • Improved performance during the data migration procedure. Thus, faster data movement.
  • Support for fast create for AFM and AFM-DR. This reduces the number of operations that are generated at the time of file creation at the cache. This improves the performance by saving many RPC exchanges between the application node and the gateway node and also saves memory and network bandwidth.

HDFS transparency

  • Improvements in HDFS for IBM Spectrum Scale 5.0.5 including the offline upgrade of HDFS Transparency for CES HDFS protocol, enhanced access to HDFS transparency logs, and improved replication configuration. Rely on the latest improvements and get the most of your Hadoop environment.

File audit logging

  • With fileset file audit logging, you can specify a list of up to 20 filesets to enable or skip audit logging from a file system. This provides you fine-grained and extensive control when audit logging your system.

Installation toolkit

  • Support for packages signed with a GPG (GNU Privacy Guard) key. This helps you to ensure the authenticity of distributed files. Therefore, to enhance your system security.
  • Support for both FQDN and short name in the /etc/hosts file. This makes even more transparent the installation process despite the type of names used in your system.
  • ECE Rolling Upgrade from 5.0.4.3 and 5.0.4.4 to 5.0.5. Making it easier to upgrade your ECE cluster and get the latest additions and improvements.
  • Offline upgrade of HDFS Transparency for CES HDFS protocol. Allowing you to have the most recent features to operate your Hadoop environment.

Management GUI

  • Enhanced interfaces to view details on clustered watch folder at the file system and fileset levels. This allows you better control over your file system components. For example, having the list of the clustered watch folders that are configured in the system.

Call home

  • New command to show configuration changes between two call home uploads. This can help you to track changes in your configuration and debug performance problems.
  • New option to add and remove child nodes to or from existing call home groups. This supports a more flexible management of call home groups.
  • New option to test the call home connectivity to the IBM infrastructure to validate information can be uploaded to IBM. This allows you to check the call home connectivity without the need to create call home groups for non-group members.

You can find more information about all these new features and more documentation on Spectrum Scale in the following links:








#Storage
#PrimaryStorage
#StorageManagementandReporting
0 comments
60 views

Permalink