Summary of changes

Snapshot and data management

• Added support for writable snapshots on AFM filesets for flexibility in data operations. For more information, see Writable snapshots in the AFM local update mode.
• Introduced compound remove operations to improve efficiency in file deletions within AFM filesets. For more information, see AFM fast remove for large directories.
• Enabled out-of-band cache updates (outbandupdate) for NFS backends to keep cache data synchronized. For more information, see AFM cache synchronization with outband updates and Synchronizing AFM cache by using outband updates.

Prefetch and performance improvements

• Enhanced multi-prefetch capabilities by using the --multi-prefetch parameter for faster and more efficient data downloads. For more information, see mmafmctl.
• Added support for prefetch with the --wait parameter to improve execution control. For more information, see mmafmctl command.
• Supported the prefetch operations for files and directories by using a single list to simplify bulk operations.
• Enabled asynchronous prefetch (async prefetch) for AFM NFS and GPFS backend target filesets to support CAST. For more information, see Asynchronous auto prefetch for AFM NFS-based filesets.

Resiliency and availability

• Added failover mapping to handle home NFS outages for system resiliency and continuity improvement. For more information, see The failover mapping for AFM NFS targets.

Configuration enhancements

• Enabled configuration of AFM with nonstandard ports for NFS targets for deployment flexibility. For more information, see Nonstandard NFS ports for AFM.

Usability and management

• Improved migration precheck and overall usability enhancements by enabling the afmMigrate parameter. For more information, see Data migration by using AFM migration enhancements.
• Expanded coverage for new parameters for the system tuning. For more information about, afmAllowReadBypass, afmReadBypassMapSize, afmMaxAsyncReadCount, afmMaxPMounts, afmPrefetchFileSize, and afmPrefetchSize, see mmchconfig command.

• Review mode for call home uploads
• Only for IBM Storage Scale Systems, new option to automatically collect gpfs.snap for automatically created service tickets.

• The cloudkit command can deploy an IBM Storage Scale cluster on IBM® Cloud.
• Protocol support for Google Cloud Platform clusters.

• Starting with IBM Storage Scale 6.0.1.0, the no value is set as default for the mmapRangeLock attribute of the mmchconfig command. This default setting helps to avoid possible deadlocks when the file is concurrently accessed by using mixed mmap operations (mmap and non-mmap operations). For more information, see mmchconfig command.
• Starting with IBM Storage Scale 6.0.1.0, the mmchconfig command supports the restripePauseEnabled and restripePauseSetTimeWindow parameters, which can pause a restripe process when direct input/output (DIO) applications are detected. This new capability prevents potential starvation for DIO threads when a command like mmchdisk, mmdeldisk, mmadddisk, or mmrestripefs is used. For more information, see mmchconfig command.
• Starting with IBM Storage Scale 6.0.1.0, to prevent applications from inadvertently using the fcntl(F_SETFL, O_DIRECT) call without receiving an indication that it is not supported, the code now enforces this limitation and rejects the call. fcntl(F_SETFL, O_DIRECT) has never been supported on a GPFS file system. For more information, see Considerations for the use of direct I/O (O_DIRECT).
• For new clusters with a minimum release level of 6.0.1.0 or later, the default value of the maxMBpS parameter is 25000. For existing clusters, the default value of this parameter is 2048 and the system administrator can change it to 25000 if the minimum release level of the cluster is at least 6.0.1.0. For more information, see mmchconfig command.
• A new parameter is introduced with IBM Storage Scale 6.0.1.0 for the mmbackup command: --ignore-migration-state. When this parameter is specified, if the only change of a file is the migration state, that file is not selected as an update candidate. For more information, see mmbackup command.
• Through improvements to the mmfsckx command, IBM Storage Scale 6.0.1 introduces enhanced checks in mmfsckx. The mmfsckx command can now detect and report inconsistencies with directory entries. This function only detects and reports the directory entries inconsistencies but does not repair them. For more information, see mmfsckx command.
• Cluster communication from remote clusters is strictly inspected on the home cluster. Invalid remote procedure calls (RPCs) are rejected and warnings are logged in the home cluster. For more information, see Accessing a remote GPFS file system.
• The mmauth command now supports the --force option. For more information, see mmauth command.
• Thanks to scaling changes made to the IBM Storage Scale feature known as remote fileset access control (RFAC), the limit of 31 clients per cluster is eliminated. For more information, see Fileset access control for remote clusters.

Only for IBM Storage Scale Systems and tailored for specific hardware, more targeted diagnostics data is collected.

• Supports AFM and AFM Object Storage operations through the scalectl commands and the corresponding native REST API endpoints. For more information, see scalectl afm command, scalectl afmcos command, AFM endpoints, and AFMCOS endpoints.
• Supports applying file system policies through the scalectl policy apply command and the corresponding native REST API with support for RBAC domain filtering and GRPC communication. For more information, see scalectl policy command and gpfs_igetfilesetdomainid() subroutine.
• Full coexistence support for the mmaddnode command within native REST API environments by removing all prior restrictions at release level 6.0.1.0 or later. This support includes auto-migration of newly added nodes for concurrent management through classic IBM Storage Scale mm-commands and the native REST API.
• Enables file system rebalance and restripe operations through the scalectl filesystem rebalance and scalectl filesystem restripe commands and the corresponding native REST API endpoints. For more information, see scalectl filesystem command, /scalemgmt/v3/filesystems/{filesystem}:rebalance: POST and /scalemgmt/v3/filesystems/{filesystem}:restripe: POST.
• Supports parallel copy and file synchronization operations through the scalectl utils xcp command and the corresponding native REST API endpoints. For more information, see scalectl utils command and XCP operation endpoints.
• Added support for managing default quotas and per-fileset quotas through the IBM Storage Scale native REST API and corresponding scalectl filesystem quota commands. It includes support for listing and updating quotas, enabling or disabling default quotas and quota enforcement, checking quota usage, reporting quota discrepancies, and filtering quota entries by quota name. For more information, see scalectl filesystem command, scalectl fileset command, and Quota endpoints.
• Supports creating, listing, retrieving, and deleting directories within a file system through the scalectl filesystem directory commands and the corresponding native REST API endpoints. For more information, see scalectl filesystem command and File system endpoints.
• Includes the --fast-tier option in the scalectl filesystem disk add command for configuring fast tier storage pools. For more information, see scalectl filesystem command.
• Accepts the unique identifier (UID) value in addition to the file system name for the scalectl filesystem get command and the corresponding native REST API file system GET endpoint. For more information, see scalectl filesystem command and /scalemgmt/v3/filesystems/{filesystem}: GET.
• Retrieves time zone, operating system, architecture, and IBM Storage Scale version information in the JSON output of the scalectl node status command and the corresponding native REST API endpoints. For more information, see /scalemgmt/v3/nodes/status: GET.
• Added support for installing the scalectl binary as a separate gpfs.scalectl package, enabling use of scalectl on remote systems outside the cluster without requiring the full gpfs.scaleapi package dependencies.
  Note: The gpfs.scaleapi package continues to provide the core REST API components and depends on gpfs.scalectl, so scalectl remains available on IBM Storage Scale nodes.

For more information, see IBM Storage Scale management API endpoints.

• Modernization of the GUI toward IBM’s Carbon standard, with enhancements to the Nodes and Events pages.
• Fileset-level multitenancy.
• AFM overview implementation.
• S3 implementation.
• Enhancements for creating and deleting snapshots.
• Improvements to support the creation of many filesets through the REST API.
• Support for Azure Linux.

For more information, see Introduction to IBM Storage Scale GUI.

• Starting with 6.0.1.0, you can run IBM Storage Scale on Red Hat Enterprise Linux® 9.8 (RHEL 9.8).
• Starting with 6.0.1.0, you can run IBM Storage Scale on RHEL 10.2.
• Starting with 6.0.1.0, you can run IBM Storage Scale on SUSE Linux Enterprise Server 16 (SLES 16).

For more information, see 2.1 What is supported on IBM Storage Scale for AIX, Linux, and Windows?.

• New sensor GPFSSnapshot to better analyze every phase of the snapshot creation or deletion.
• New sensor CesS3 to monitor S3 protocol performance.
• New CLI options --legend and --top for the mmperfmon command.
• New default measurements: nsdds_lat_disk_write_sum and nsdds_lat_disk_read_sum.
• More descriptive error messages for mmperfmon, mmpstat, mmptop commands.
• New default threshold added: Neterr_Rule.
• Only for IBM Storage Scale Systems, added anomaly detection to monitor and report probable anomalies in cluster behavior.
• Support for the NVIDIA Nsight plug-in.
• For easier deployment of the Grafana bridge, new RPM and DEB packages: gpfs.grafana-bridge.
• Multi-tenancy observability improvements to allow tenants to query their own storage metrics.

First-time data capture

• Enhanced this feature for periodic statistics collection, hang diagnostics, core dump generation, and tracing tools for server and client-side analysis. Also, supported data capture for default Samba config, CTDB config.
• Added new scripts such as ganeshatracectl and nfs_client_debug_script.py to capture metrics such as NFS operations, memory usage, stack traces, network state, and packet. For more information, see NFS client and server runtime diagnostic.

For more information, see First-time data capture for CES NFS.

Samba support

• Updated Samba on IBM Storage Scale to the upstream Samba 4.22.4.
• Added The SMB multichannel feature in Samba.
• Improved domain controller failover time for multiple DCs in Samba.

Support for S3 over RDMA

Started S3 over RDMA support. This feature enables high‑performance object access to data in GPFS file systems through RDMA transport for S3 data traffic. For more information, see S3 over RDMA overview and deployment.

Support for S3 IAM user management

Started S3 Identity Access Management (IAM) support for S3 user management. For more information, see IAM users for S3 access control.

CES S3 sensor for performance

Added support to collect performance metrics for S3 operations. For more information, see CES S3 metrics.

Support for LDAP

AUTH by using LDAP is supported.

• Extended AFM replication monitoring, which adds dedicated events with tailored user actions.
• Added ARP cache monitoring.
• mmhealth cluster show <component> -v now shows events like mmhealth node show.
• The mmhealth command introduces the alias --not-ok option for the --unhealthy option.
• For the distinction between administrator and daemon names in the mmhealth command, the human-readable output is now switched to administrator names only. The -Y output stays consistent but adds the new field adminname.
• If more than three events are active, the output for mmhealth node show displays ellipsis after the third event.
• Autoexpel now allows to specify a custom exclusion list.

QoS is a storage management feature that helps to ensure that critical applications consistently receive the performance they require, while maximizing the efficiency of the storage infrastructure. For more information, see Quality of service.

New commands

• scalectl afm
• scalectl afmcos
• scalectl utils

Changed commands

• gpfs.snap
• mmadddisk
• mmaddnode
• mmafmctl
• mmafmconfig
• mmauth
• mmbackup
• mmcachectl
• mmcallhome
• mmchattr
• mmchconfig
• mmchfileset
• mmchfirmware
• mmchfs
• mmchpolicy
• mmchvdisk server
• mmcrfileset
• mmcrfs
• mmcrsnapshot
• mmdelfs
• mmdumpperfdata
• mmfsckx
• mmhealth
• mmlsfs
• mmlslicensem
• mmperfmon
• mmrestorefs
• mms3
• mmsmb
• mmvdisk pdisk
• mmvdisk vdisket
• scalectl
• scalectl filesystem
• scalectl fileset
• scalectl policy

New messages

Changed messages

Deleted messages