What's New?
The Life cycle process now accepts a new Storage-Optimizer action designed to perform storage optimization of the target object storage. This new capability frees up object storage space caused by incomplete or failed Model9 processes that is not freed during normal processing.
- Fixed a
NullPointerExceptionissue when opening the “More Details” pop-up of a data set backup that was already deleted from the object storage. - Fixed a
NullPointerExceptionissue while running a backup or archive for a data set created by BMC IAM. - Fixed an unexpected error during discovery when a backup of an uncataloged data set by volume was performed, while another data set by the same name was cataloged on another volume.
- Fixed an ABEND
S613-20that can occur while writing Cloud Data Sets in an environment in which some Cloud Data Set jobs useVOL=REF(tape stacking) - Cloud Data Sets now blocks unsupported
EXCPcalls to prevent unexpected behavior when using unsupported applications.
- Improved the dashboard page loading time after initial log-in to the web UI by implementing lazy loading techniques for each presented element.
- Fixed Cloud Data Set (CDS) VDEV component issue where
S0C4-38abends occur and lead to common storage leaks in the CSA/ECSA. - Fixed a letter case issue when restoring a data set from backup, where the new name contained mixed case letters which results in a failed restore run.
- Fixed management class values not shown in the UI for CLI-originated data set backups.
- Fixed the Mode9 server s390x installation artifact which was missing the PostgreSQL container.
The Model9 management server was enhanced with two API calls to allow retrieving all the run ids of a specified policy and to retrieve the activity log of a specific activity. These new APIs enhance the ability of automation systems to integrate with the Model9 platform.
Automatic retry capability for object storage actions was enhanced to tolerate some more network-related issues which are intermediate. This new enhancement reduces the failure ratio as all Model9 components retry each action before it notifies of a failure to tolerate intermediate issues.
Cloud Data Sets support is enhanced to support generational data set (GDS) expiration where the generation data group is defined with PURGE and SCRATCH. This allows automatic handling of the expiration process of a given generation in case it is no longer needed.
Reporting capability which allows creating scheduled reports that present the storage utilization. This new capability allows creating storage usage reports grouped by either the policy name, the originating CLI command or by high-level qualifier and type.
- Improved login failure messages in cases where the user that performs the login action doesn’t have an OMVS segment defined. The new behavior will now explicitly notify the user of exactly why the login failed.
- Fixed an issue with the server API (SAPI) that in some rare cases caused a
OptimisticLockingFailureExceptionwhen using the same API key for all SAPI jobs. This fix now allows using the same key for all SAPI jobs. - Fixed a memory leak that prevented the “Data set cleanup” policy to run when using large storage groups or volumes. This fix enables the full use of the new policy type.
The life cycle process is changed such that projection index objects are created in a date folders-like structure, reducing the large number of objects in a single folder. This new behavior improves the scalability of the life cycle process and reduces the risk of list timeouts when communicating with on-premises object storage systems.
Cloud Data Sets support is enhanced to tolerate JCL jobs stacking multiple data sets on a single tape volume. The stacking tolerance support allows creation of several data sets using the same VOLSER. Cloud Data Sets created in stacking tolerance mode receive the same VOLSER but are cataloged with the same file sequence 1. A known limitation in this release is that JCL jobs referring to a Cloud Data Set using its specified file sequence number will fail as all Cloud Data Sets are cataloged with FSEQN=1. To override this limitation, either specify a sequence number of 1 to read the data set or use the catalog entry to refer to the data set.
- To avoid incorrect deletion of archived GDS generation data set in cases when a recall operation is running in parallel to a Life Cycle process handling the same GDS, the Life cycle process now uses an additional locking mechanism while deleting expired generational data set. This change only applies to archived generational data sets with catalog-controlled expiration dates (i.e. EXPDT=99000).
- Added the ability to run the “Data set cleanup” policy type (introduced in the previous release) via the Management Server API (SAPI).
- Fixed an issue where Cloud Data Sets virtual tape drives with a sub-channel number of 0000 will not vary online after restarting the VDEV started task.
- Fixed the Restore panel on the Resources page in the UI to explicitly specify “Null storage class” and “Null management class” when selecting a null SMS storage class or null SMS management class.
- Fixed an issue on the Resources page in the UI where the toggle bar may overlay the resources table when a resource menu is open. The toggle bar is now fixed to the top section of the page.
Data set organization is commonly used to filter in or out resources when performing certain actions in z/OS like archive and backup. Model9 Manager only supported the following filters for archive and backup: Data Set Name Volume Name Storage Group Name If the data set is ZFS
Starting with version 2.4.0, the storage administrator will now have the ability to filter out based on the following additional data set organization types: Sequential (All sequential type data sets) Partitioned (PDS & PDSE) VSAM (All types)
Create new policy “Filters” section:
Known limitations: The data set organization types are grouped into three groups and not allowed to exclude, for example, only data sets with organization PDSE, and not PDS
Data sets that have passed their expiration date, are to be deleted by the data management product. Model9 Cloud data manager platform only provided an external JCL to handle the deletion of expired data sets.
Starting with version 2.4.0, a new policy by the name of “Data Set Cleanup” is introduced. This policy type will handle the deletion of expired data sets from a specified storage group or volume name
Create a new “Data Set Cleanup” policy:
Known limitations: Only cataloged SMS data sets will be handled by the new policy type in this release. Uncatalogued and non-SMS managed data sets will not be deleted by this policy
Audit trail is an essential capability for every product to have a clear picture which actions were performed and by whom.
Starting with version 2.4.0, new audit trail capabilities are introduced to collect information about CLI-initiated actions.
The Audit log file output example:
Known limitations: UI-initiated actions are not yet logged
The new user interface that was introduced in version 2.3.0 only provided a preview version of the resources page, without allowing to perform any actions. In version 2.3.5 the ability to restore a single data set was introduced.
Starting with version 2.4.0, the resources page will now allow restoring a single full volume and will allow recalling a single archive
Restore volume action:
Known limitations: Only single resource actions are supported. Performing export for an imported cloud data set is not yet available.
QSAM is one of the most commonly used access methods for tape data sets. There are a large number of z/OS utilities and applications which utilize it to read & write data.
This support will allow more applications to read & write directly to cloud object storage without the need for interim storage like tape or DASD.
Known limitations: Even though both read and write are supported, not all access method calls were implemented. Thus only limited applications will be officially supported by Model9.
QSAM write/read for RECFM=U is not supported
QSAM LBI is not supported
Until this version, the CDS read buffers weren’t utilized properly thus limiting the read throughput.
Starting with v2.4.0 the CDS read buffers will be used fully, meaning that the read-ahead buffer, which by default is set to 5MB, can now read several blocks into one buffer instead of one block per buffer.
Model9 Gravity (formerly known as the Model9 Transformation Service), makes mainframe data available for cloud-native AI/ML and analytics tools. As part of this release, mainframe Db2 Image Copies and archive logs, VSAM data sets, and sequential files are easily transformed to CSV and other open file formats stored directly in any cloud object storage of your choice. The current Transform Services are available for our current customers. This capability is currently available for selected customers only. Please contact Model9 Customer Success for more information.
- Model9’s discovery process generates “CONSTRUCT ACCSS-SMS CNS” entries in the LOGREC data set, even though there is no actual issue during the discovery process due to the SMS buffer size being too small. New: Beginning with version 2.4.0, there is a new parameter to control the buffer size used to query the SMS settings named:
sms.ssi.answer_area.size.bytesThe default value is set to 1, which is the same value as today - Model9’s discovery elapsed times in large policies could have taken a long time to complete, in some cases even hours. New: Beginning with version 2.4.0, the discovery mechanism introduced internal changes so it would perform the initial resources scan more efficiently. This change is expected to shorten the discovery times to a matter of minutes
- Model9’s credentials encryption tool was missing from the s390x server artifact. New: Starting from version 2.4.0, the credentials encryption tool will now be delivered with both the x86 and the s390x server artifacts
- Model9’s Dashboard page sometimes takes a long time to load after login. New: Starting from Version 2.4.0, the dashboard page will perform lazy loading for the screen components in order to improve the user experience.
- New Resources page - To continue the preview offered in the previous release, the new resources page now allows restoring data sets from backup originating from the server. Furthermore, the filtering mechanism now allows selecting custom dates & times for a more granular search.
New resources page
- API keys creation blocked for non-administrators - The creation of API keys that are used by the Server API (SAPI) to run policies from z/OS, will be blocked for non-admins. Only Model9 admin users will be able to generate new keys. All previously created keys will remain as is.
- The removal time of full dumps associated with a large number of indirect resources was improved by parallelizing the delete of the indirect resources.
- Archive policy which ended in status “ended warning” due to Model9 database being out of sync will now end with status “end ok”.
- CLI restore volume command (RESTVOL) now supports restoring z/VM CP volumes.
- The discovery network buffers are now configurable to improve performance.
Apache Log4j Vulnerabilities To address the recently disclosed Apache Log4j vulnerabilities
(CVE-2021-44228, CVE-2021-45046, and CVE-2021-45105), Model9 components do not make use of Apache Log4j thus there's no need for any additional actions.
- New Resources page - A sneak-peek of Model9’s upcoming new User Interface (UI) is now available! Head to the resources page of the current UI application and you can see new sort, search, and enhanced filtering capabilities. Your feedback counts! Help us improve our application by answering this short survey
New resources page
- Record Descriptor Word (RDW) support - The transformation service now includes the RDW field for raw output format supported for record format V and VB.
- RACF TRUSTED attribute support - The RACF TRUSTED attribute can now be used instead of the OPERATIONS attribute to run the Model9 agent and the life cycle procedures. This will allow logging all the actions the Model9 components are performing. For more details see the Installation and Upgrade Guide.
- Fixed an expiration issue for rolled off archive Generation Data Sets (GDS), when the base Generation Data Group (GDG) was defined with the PURGE attribute. Prior to this release the expiration date wasn’t set correctly and the life cycle was unable to delete the data set after it was rolled off.
- Fixed Command Line Interface (CLI) LISTDSN returning "no results found" when resource name contains special characters.
- Fixed CDS parallelization issue on slow systems, by using a locking mechanism.
- Fixed CDS jobs failures in slow systems, when jobs run one after another.
- Fixed automatic recall hook collision with 3rd party products invoking nested SVC 26 (LOCATE).
- Allow searching resources based on qualified names and with wildcards. This allows filtering more easily for a specific resource.
- Run log will now indicate if a full dump has a dump conditional name to allow identifying volumes with temporary names.
- Reduce the agent’s General purpose CP (GCP) by setting the default value of _EDC_PTHREAD_YIELD to 0.
Apache Log4j Vulnerabilities To address the recently disclosed Apache Log4j vulnerabilities
(CVE-2021-44228, CVE-2021-45046, and CVE-2021-45105), Model9 components do not make use of Apache Log4j thus there's no need for any additional actions.
- New Resources page - A sneak-peek of Model9’s upcoming new User Interface (UI) is now available! Head to the resources page of the current UI application and you can see new sort, search, and enhanced filtering capabilities. Your feedback counts! Help us improve our application by answering this short survey
New resources page
- Record Descriptor Word (RDW) support - The transformation service now includes the RDW field for raw output format supported for record format V and VB.
- RACF TRUSTED attribute support - The RACF TRUSTED attribute can now be used instead of the OPERATIONS attribute to run the Model9 agent and the life cycle procedures. This will allow logging all the actions the Model9 components are performing. For more details see the Installation and Upgrade Guide.
- Fixed an expiration issue for rolled off archive Generation Data Sets (GDS), when the base Generation Data Group (GDG) was defined with the PURGE attribute. Prior to this release the expiration date wasn’t set correctly and the life cycle was unable to delete the data set after it was rolled off.
- Fixed Command Line Interface (CLI) LISTDSN returning "no results found" when resource name contains special characters.
- Allow searching resources based on qualified names and with wildcards. This allows filtering more easily for a specific resource.
- Run log will now indicate if a full dump has a dump conditional name to allow identifying volumes with temporary names.
- Reduce the agent’s General purpose CP (GCP) by setting the default value of _EDC_PTHREAD_YIELD to 0.
- A new setting was added to control the DFDSS OPTIMIZED used during data set backup and archive.
- MinIO artifact upgraded to a release 2021-07-12.
- Introduced a fix where the life cycle fails with
ZM9L014Edue to internal data retrieval issue - Fixed an unclear message when a data set restore through the UI fails.
- Model9 UI fix to correct license expiration notice.
- Fixed UI restore being unresponsive when no active agents are available.
- UI Resources page table width is now customizable.
- UI Resources page can now filter based on date.
- Obfuscation capabilities for the object storage credentials allowing for obfuscation of the secret key to prevent it from being in clear text.
- Transformation service support for FBA and VSAM with Extended Addressability file types is now available.
- Multi delete functionality from the status report page allows clearing the entire content of a specific run.
- Reintroduce the CMD origin column for the LISTDSN CLI command to indicate whether the backup originated from CLI or the UI.
- Model9 UI CSS selectors were fixed in several places which caused some glitches.
- Model9’s lifecycle process should now take into consideration the previous run results when it initiates to verify if previous dates need to be reprocessed.
- Fixed support issues for special characters in the SMS Management Class name.
- Fixed CLI DELBACK command to support more than a 1000 backup versions for a single data set.
- Transformation service infinite loop prevention fix for DB2 image copy processing.
- Policy Finalization process will now be a parallelized process which will reduce the server time for policy runs.
- UI API Keys page internal infrastructure upgrade to speed up its loading times and fix keys removal issue.
- UI Resources page would now allow filtering based on the origin of a data set backup.
- Management server resources search page: removed the 350 version limitation. The search page now allows searching through all versions in the system. Maximum page has has been increased to 500 versions. As part of this change, the '**' wildcard has been superseded with '*' which now implies zero or more characters.
- AWS S3 Infrequent Access and Azure Blob Cool tier support which allow a considerable reduction in storage costs for newly written blobs.
- Archive policy retention override support.
- Backup versions originating from CLI can now be deleted and restored from the management server's resources search page.
- Transformation services now support converting DB2 Image Copies to open system formats such as JSON and CSV.
- In some cases, archive errors were reported incorrectly in the management server's backup logs.
- Restore data sets did not retain the original backed up data set's creation and reference dates.
- In some cases, full volume dumps ended in error due to DCOLLECT warnings.
- Incorrect backup and archive handling of data sets with the special EXPTD date of '99000'.
- Full dumps of volumes with short VOLSERs completed with a warning.
- Archive policies edit page incorrectly showed a generations field.
- LISTVOL and LISTDSN CLI commands did not show a warning in the case of failed object storage connectivity.
- In some cases, a restore operation's error was not propagated to the backup log on the management server.
- Fixed several stability issues in M9SAPI.
- Improved z/OS below the line memory management.
- Improved memory management in M9SAPI.
- Azure Active Directory (AD) authentication support for Microsoft Azure cloud users. Model9 introduces Azure Active Directory (AD) authentication and proxy support, providing simplified access to Azure’s public cloud. Model9 users can now use their enterprise-provided AD credentials to gain access to the Azure cloud. See Microsoft Azure configuration for more details.
- Ease migration from CA-Disk to cloud data management. Gain better control of backup policies with the ability to specify per-policy retention periods and deletion dates. This capability helps smooth the migration from legacy backup software such as CA-Disk to modern, cloud-oriented solutions. See Backing up data sets for more details.
- Enhanced system health visibility. Version 1.8 provides at-a-glance system health and connectivity data for the Model9 management server and database, as well as the object storage service. See Server status for more details.
- Improved access to log data. Mainframe users can download server logs for advanced analysis - without having to set up an SSH session on the Linux server. See Where to look for error messages for more details.In addition, log summary reports can be viewed on the Model9 browser page without having to open and scan individual log files. See Viewing resources that were created by an activity for more details.
- You can use the new BACKDSN CLI command to perform on-demand, data-set level backup operations directly from the mainframe to any cloud/on-premise, hybrid or public. This functionality enables the execution of backup operations within scripts and JCLs, and simplifies storage administration.
- The BACKDSN CLI command also includes NEWNAME, NEWDATE and NEWTIME parameters intended for simplifying migration from legacy backup software by allowing you to specify the same backup attributes as were stored by the original backup software.
- In addition, the BACKDSN CLI command supports specifying a backup retention period. The life cycle management process is enhanced to delete backups whose retention period has expired.
- A new DELBACK CLI command is used to delete data set backups created using the BACKDSN CLI command. Moreover, the command enables management of backups within scripts and JCLs, and simplifies storage administration using native Mainframe tools only.
- The agent is enhanced to manage the number of concurrent system utilities, such as ADRDSSU (DFDSS), active at the same time and queue additional requests when no more instances can be run in parallel due to address space virtual storage constraints. This enhancement enables running multiple server policies in parallel and avoiding errors due to agent resources being unavailable.
- Obsolete activity logs can now be deleted to free up space on the management server file system and remove unneeded information from the management interface. New server parameters control how long activity logs shall be retained after all resources have expired.
- The life cycle management process has been enhanced to parallelize backup and archive deletes providing a performance improvement of up to 90% in processing time for large data sets.
- The restore and recall actions, whether invoked from CLI or from the management UI, have been enhanced to better utilize network throughput, providing up to 35% reduction in restore or recall times for large data sets.
- The output of the LISTDSN CLI command has been changed to include an additional column indicating whether a data set backup was created using a CLI command or a policy run. Please refer to the CLI User Guide for more information.
- The management graphical user interface is updated to correctly handle data set delete actions performed on Write-Once-Read-Many (WORM) protected storage, such as Hitachi HCP WORM or AWS S3 ObjectLock, and avoid inconsistencies between the management server database and the object storage. Note that data immutability is only maintained by the object storage platform..
- Blocked the erroneous ability to restore archived data sets, which led to pending ENQs being left standing in the Model9 agent address space.
- Fixed discovery process not to skip storage groups and volume serial names containing the ‘$’ sign.
- Fixed discovery process not to skip data sets residing on EAV volumes during an archive policy run, which marked the entire run as ‘Ended Error’.
- Fixed an issue with listing objects using certain API calls when using Hitachi HCP object storage as the target storage.
- The Policies page in the user interface has been enhanced to remember the page number, sort order and search filter for the duration of the session, whether editing or creating a new policy.
Release 1.6 features an enriched set of mainframe-focused operations that eliminate dependence on external systems
- Use the new CLI ARCHIVE (Command Line Interface) command to perform on-demand, data-set level archiving operations directly from the mainframe to any cloud - on-prem, hybrid or public. This functionality enables the execution of archive operations within scripts and JCLs, and simplifies storage administration.
- Run the Model9 management server in a z/OS container extension (zCX), and eliminate dependence on external systems. The Model9 docker container can be deployed on zCX, Linux on Z, or any other docker, while supporting the familiar graphical user interface. To learn more, join the Model9 speaking session at SHARE Virtual 2020, on September 30, 2020, 4:15 PM-5:00 PM.
- Simplify monitoring by extracting the policy’s run logs from the management server to the SAPI batch job in z/OS. No need to access the logs via the server.
- Convert the “discovery” phase to a zIIP eligible process for improved policy performance and reduced CPU consumption. See the installation guide for further details.
- The Archive policy now skips data sets with expiration date in the past, instead of archiving them. The policy run log reports these data sets to notify the user that an action may be required outside of the archive policy in order to delete these data sets.
- The Archive policy now uses the correct scope of SYSPLEX instead of the previously used SYSTEM when checking whether a data set is eligible for archive. The false scope would cause certain data sets to be marked as eligible for archive in the discovery phase of the policy, but to fail during the archive action itself.
- When backing up a symbolic link to a file, the policy will only backup the symbolic link itself and not the file that the link is pointing to. If the policy's filter matches the actual file itself, it will be backed up (just not via the link). When backing up a symbolic link to a directory, the policy will backup the directory's contents and the link pointing to the directory, but the directory itself (which might have permissions) will not be backed up.
IBM Java 8 SR6 FP15 is not supported for the z/OS agent.
- Added full integration with Cohesity's Data Platform and marketplace, consolidating onto Cohesity for unified management, simpler protection, and deep security.
- Added detection and notification on recall and restore of objects that were moved to Amazon S3 Glacier Deep Archive.
- Added support for archive and recall of GDG Extended format data sets.
- Added detection and handling of write inconsistencies in the underlying IBM JVM. The fix detects an incomplete write attempt, reports it to the log and performs a retry. Additionally, a change to the JVM interface was implemented to significantly decrease the probability of this situation to happen.
- Fixed the Export action to support SMF files format.
- Fixed policy execution to prevent it from skipping the rest of a volume or of a storage group when encountering an unexpected error.
- Reduced automatic recall retry attempts, to avoid long waits and excessive messages to the agent’s log and to the SYSLOG.
- Fixed policy scheduling to ignore disabled policies.
- Increase the CLI LISTDSN output size to accommodate very large queries.
- Provide a configurable UI session timeout.
- Support multiple licenses in a single license file to better facilitate SYSPLEX spanning over different machines and to avoid manual license updates in disaster recovery scenarios.
- Simplify installation by improving the configuration file structure.
- The server logs by default now rotate every 24 hours or 500MB.
- The location of the server logs has been changed and is now under MODEL9_HOME/logs to allow easier retrieval of the logs when needed.
- Improved jclouds log visibility to allow faster diagnosis of network-related issues.
- Automatic recall does not perform as expected in situations such as IDCAMS DELETE MASK and LISTDSI NORECALL. These situations are documented internally and will be handled in a future release. The suggested workaround is to use the CLI DELARC command to delete archived data sets and the ZM9$NORC DD to exclude automatic recall from specific jobs.