Summary of changes

This topic summarizes changes to the IBM Spectrum Scale™ licensed program and the IBM Spectrum Scale library. Within each topic, double angle brackets ( ) surrounding text or illustrations indicate technical changes or additions made to the previous edition of the information.

Summary of changes
for IBM Spectrum Scale version 4 release 2.1
as updated, July 2016

This release of the IBM Spectrum Scale licensed program and the IBM Spectrum Scale library include the following improvements:
Auditing configuration changes
A syslog entry is automatically written whenever a GPFS™ command makes a configuration change. Adding the information to the syslog gives flexibility in mining, processing, and redirecting these events. Entries can also be written to the standard GPFS log. The commandAudit parameter of the mmchconfig command controls this option. For more information, see Audit messages for cluster configuration changes.

Automated configuration of sensors for performance monitoring
IBM Spectrum Scale now supports automated configuration of sensors for its performance monitoring tool. For more information on automated configuration of sensors, see Automated configuration.

Callback event for file system structure errors
A new user callback event fsstruct (file system structure error) is triggered when the file system detects an error in the metadata. Immediate notification enables the callback program to act to mitigate further errors. For more information, see mmaddcallback command

CES, NFS, and SMB protocols: Support for SLES V12 on x86 systems
Cluster Export Services (CES) partially supports SUSE Linux Enterprise Server (SLES) V12 on x86 systems. The SMB and NFS protocols are now supported via a manual installation process. For more information, see Manually installing IBM Spectrum Scale on SLES 12 systems.

Compression support for FPO environments
File compression is expanded to support the File Placement Optimizer (FPO) environment. For the FPO environment, you must set the block group factor to a multiple of 10 to avoid degrading file system performance. For more information, see File compression.

Deadlock management and debug data control
Deadlock management is extended with the following features:
  • The detection thresholds for deadlocks are automatically adjusted according to waiter length and cluster overloaded status.
  • New defaults more suitable for customer environments are established for the configuration variables deadlockDataCollectionDailyLimit, deadlockDataCollectionMinInterval, and others.

A new configuration variable debugDataControl controls the amount of debug data that is collected. The default setting is a minimal amount of debug information that is the most important for debugging issues.

For more information, see Managing deadlocks.

/dev/<fs_name> device for a file system on Linux
On Linux, GPFS no longer creates the /dev/<fs_name> device for a file system. Applications that relied on the file system device under /dev to be present, or that relied on "/dev" to be displayed in the output of the mount command, must find other ways to obtain the information. As a substitute, consider the information provided by the /etc/fstab file and /proc/mounts entries.

Documentation changes
The IBM Spectrum Scale documentation guides are changed as follows:
Table 1. Documentation guides in this release
Guides in release 4.2 and earlier Guides in release 4.2.1
IBM Spectrum Scale: Concepts, Planning, and Installation Guide IBM Spectrum Scale: Concepts, Planning, and Installation Guide
IBM Spectrum Scale: Administration and Programming Reference

Except commands, programming interfaces, and user exits

 IBM Spectrum Scale: Administration Guide
IBM Spectrum Scale: Advanced Administration Guide
IBM Spectrum Scale: Problem Determination Guide IBM Spectrum Scale: Problem Determination Guide
IBM Spectrum Scale: Data Management API Guide IBM Spectrum Scale: Command and Programming Reference
IBM Spectrum Scale: Administration and Programming Reference

(Commands, programming interfaces, and user exits)

For more information, see Product documentation in guide format.

The top-level structure of the IBM Spectrum Scale documentation in the IBM® Knowledge Center is changed as follows:

Table 2. Documentation structure in this release
Documentation structure in release 4.2 and earlier Documentation structure in release 4.2.1
  1. IBM Spectrum Scale: Concepts, Planning, and Installation Guide
  2. IBM Spectrum Scale: Administration and Programming Reference
  3. IBM Spectrum Scale: Advanced Administration Guide
  4. IBM Spectrum Scale: Data Management API Guide
  5. IBM Spectrum Scale: Problem Determination Guide
  1. Summary of changes
  2. Quick reference
  3. Product overview
  4. Planning
  5. Installing and upgrading
  6. Configuring
  7. Administering
  8. Monitoring
  9. Troubleshooting
  10. Command reference
  11. Programming reference
  12. Library and related publications
  13. Glossary
Encryption: Simplified setup and Vormetric DSM support
  • A new console command mmkeyserv greatly simplifies the setup of encryption both on the key server and the client node. IBM Security Key Lifecycle Manager (SKLM) V2.5.0.4 or later (including V2.6) is required.
  • Encryption support is added for key servers that run Vormetric Data Security Manager (DSM) V5.2.3 or later.
For more information, see the topic Preparation for encryption.

Federation in the performance monitoring tool
A performance monitoring tool installation with multiple collectors is called a federation. Federation is introduced in performance monitoring to increase the scalability or the fault-tolerance of the performance monitoring system. For more information on federation, see Configuring multiple collectors.

Guided installation
The spectrumscale installation toolkit now provides next step hints that are designed to help customers new to IBM Spectrum Scale with an easy workflow that helps customers to install and configure an IBM Spectrum Scale cluster.

Hadoop Support for IBM Spectrum Scale
HDFS transparency now supports running the Hadoop Map/Reduce workload inside the virtual machine container, Docker.

Federation is introduced in HDFS to solve the HDFS NameNode scaling problem.

Hadoop distcp is used for data migration from HDFS to the IBM Spectrum Scale file system and between two IBM Spectrum Scale file systems.

For more information, see the following topics:
HDFS transparency security has been introduced for the simple security mode and the Kerberos mode.

User authentication and authorization is weak in the simple mode. The data transfers and RPCs from the clients to the NameNode and DataNode are not encrypted. The Kerberos mode introduced in the Hadoop ecosystem provides a secure Hadoop environment.

For more information, see HDFS transparency security.

InfiniBand and RDMA performance
Performance is improved for clusters that use InfiniBand and RDMA for their intranode communications network.

Linux on z Systems™: Expanded features
The following IBM Spectrum Scale features are now available on Linux for z Systems:
  • Quality of Service (QoS) support.
  • Improved extended count key data (ECKD™) device handling: On different nodes, different bus IDs can refer to the same device.
  • IBM Spectrum Scale GUI now supported on Linux for z Systems.

--metadata-only parameter for mmrestripefs
A --metadata-only option for the mmrestripefs command allows the restripe to complete in less time than a full restripe of metadata and data. The savings in time is useful in situations where there is a concern about file system operations and you want to restripe. This operation is supported for migrating data off disks, rebalancing, restoring replication, and comparing replicas. For more information, see mmrestripefs command.

mmhealth: Monitoring services hosted on cluster nodes
A new command, mmhealth is added to monitor the health status of nodes and different services hosted on nodes. The mmhealth command also displays the event logs responsible for the unhealthy status of nodes and services, to analyze and determine the problem responsible for the service failure. For more information, see mmhealth command.

Object storage improvements

Quality of Service for I/O operations (QoS) improvements
Quality of Service for I/O operations is expanded to support the File Placement Optimizer (FPO) environment. For more information, see Setting the Quality of Service for I/O operations (QoS).

Re-create and restore options for protocols cluster failover
The failover procedure can choose between re-create and restore options. For more information on the failover options, see Performing failover for protocols cluster when primary cluster fails.

Re-create and restore options for failing back to an old primary for protocols cluster
When failing back to an old primary, the file protocol configuration can either be re-created or restored. For more information on failing back to an old primary, see Performing failback to old primary for protocols cluster.

Re-create and restore options for failing back to a new primary for protocols cluster
When failing back to a new primary, the file protocol configuration can either be re-created or restored. For more information on failing back to a new primary, see Performing failback to new primary for protocols cluster.

Support for Transparent Cloud Tiering
The Transparent Cloud Tiering feature leverages the existing ILM policy available in IBM Spectrum Scale, and administrators can define policies to migrate cold data to a cloud storage tier or recall data from the cloud storage tier on reaching certain threshold levels.

A new command, mmcloudgateway, is added to manage and configure the cloud storage tier. For information, see mmcloudgateway command.

Note: To enable Transparent Cloud Tiering nodes, you must first enable the Transparent Cloud Tiering feature. This feature provides a new level of storage tiering capability to IBM Spectrum Scale customers. Please contact your IBM Client Technical Specialist (or send an email to mailto:scale@us.ibm.com) to review your use case of the Transparent Cloud Tiering feature and to obtain the instructions to enable the feature in your environment.
workerThreads tunes file system performance
The workerThreads parameter of the mmchconfig command controls an integrated group of variables that tune file system performance. Use this variable to tune file systems in environments that are capable of high sequential or random read/write workloads or small-file activity. This variable can be used in any installation and is preferred over worker1Threads and prefetchThreads in new installations. For more information, see mmchconfig command.

IBM Spectrum Scale GUI changes
The following main changes are added in the IBM Spectrum Scale management GUI:
  • Renamed Monitoring > Topology page to NSDs. The NSDs page facilitates monitoring the status of Network Shared Disks (NSD) and nodes to NSD mapping in the system.
  • Added new Monitoring > Nodes page in the GUI. The Nodes page provides an easy way to monitor the performance, health status, and configuration aspects of all available nodes in the IBM Spectrum Scale cluster. The properties of a node display the status of various CES servicesrecr such as Object, NFS, and SMB as well as the authentication status of these services if they are enabled. It also displays other details such as network status, information on attached NSDs and file systems, and so on.
  • Monitoring performance of transparent cloud tiering services through Performance and Dashboard pages.
  • Renamed Monitoring > Performance page to Statistics.
  • Added capacity monitoring options in the Statistics page.
  • Added monitoring options for GPFS waiters in Monitoring > Statistics panel.
  • The following improvements are made in the Dashboards page:
    • You can assign a name to the dashboards and the user can switch between dashboards.
    • The dashboards are now stored on the server instead of the browser. Therefore, it can be shared among users and browsers.
    • Default dashboards are shipped with the GUI. When you open the IBM Spectrum Scale™ GUI after the installation or upgrade, you can see the default dashboards. You can further modify or delete the default dashboards to suit your requirements.
  • Renamed Download Logs page as Diagnostic Data. Now, the GUI can be used instead of the gpfs.snap command to collect the details of the issue. For more information on collecting diagnostic data through GUI, see Collecting diagnostic data through GUI.
  • The Files > Information Lifecycle page facilitates defining compression and deletion rules.
  • The new Settings > Object Service page facilitates start and stop feature for object services.
  • Up to 1000 nodes are supported.
  • The GUI can now be used an IBM Spectrum Scale cluster where sudo wrappers are used. For more information on how to configure IBM Spectrum Scale GUI to use sudo wrapper, see Configuring IBM Spectrum Scale GUI to use sudo wrapper.
  • IBM Spectrum Scale GUI support for System z® platform is available on RHEL7.2 and SLES12.
  • By default, GUI commands that change the configuration of the cluster cause an audit message to be sent to syslog. Optionally, an audit message can also be sent to the GPFS log. For more information, see Audit messages for cluster configuration changes.

NFS and SMB protocol troubleshooting information added
New AD Discovery tool to query and validate several AD settings.
New troubleshooting information for NFS issues:
  • NFS mount issues
  • NFS error events
  • NFS error scenarios
New troubleshooting information for SMB issues:
  • SMB client on Linux failures
  • SMB mount errors
  • SMB error events
  • SMB access issues

Documented commands, structures, and subroutines
The following lists the modifications to the documented commands, structures, and subroutines:
New commands
The following commands are new:
  • mmadquery
  • mmcloudgateway
  • mmhealth
  • mmkeyserv
New structures
There are no new structures.
New subroutines
There are no new subroutines.
Changed commands
The following commands were changed:
  • gpfs.snap
  • mmafmlocal
  • mmcallhome
  • mmchconfig
  • mmchnode
  • mmcesdr:
  • mmcrsnapshot: You can create multiple snapshots in the same command.
  • mmdelsnapshot: You can delete multiple snapshots in the same command.
  • mmlscluster
  • mmnfs
  • mmobj
  • mmrestripefs
  • mmsmb
  • mmprotocoltrace
  • mmsmb
  • mmuserauth
Changed structures
There are no changed structures.
Changed subroutines
gpfs_iopen() subroutine
gpfs_iopen64() subroutine
Deleted commands
There are no deleted commands.
Deleted structures
There are no deleted structures.
Deleted subroutines
There are no deleted subroutines.
Messages
The following lists the new, changed, and deleted messages:
New messages
6027-1826, 6027-2363, 6027-2364, 6027-2365, 6027-2366, 6027-2367, 6027-2368, 6027-2369, 6027-2370, 6027-2371, 6027-2372, 6027-2373, 6027-2374, 6027-2375, 6027-2376, 6027-2377, 6027-2378, 6027-3108, 6027-3720, 6027-3721, 6027-3722, 6027-3723, 6027-3724, 6027-3725, 6027-3726, 6027-3727, 6027-3728, 6027-3915, 6027-3916, 6027-3594, 6027-3595, 6027-3596
Changed messages
6027-1368, 6027-1235, 6027-1545, 6027-2271, 6027-2272, 6027-2273, 6027-2274, 6027-2951
Deleted messages
6027-1997