IBM Support

Known Issues and Limitations: IBM Spectrum Protect Plus V10.1.6.x

Preventive Service Planning


Abstract

This document details the known Issues & limitations for IBM Spectrum Protect Plus V10.1.6 and later patches.

Content

This document is divided into linked sections for ease of navigation. Use the following links to jump to the section of the document that you require:


IBM Spectrum Protect Plus common

IBM Spectrum Protect Plus server can fail when file indexing database exceeds maximum size (internal reference #SPPSUP-1685)
Problem: When file indexing is running by IBM Spectrum Protect Plus, entries are created in an internal database, which are used for file level restores. If this database grows beyond the maximum size (~2.1 billion entries), it might cause the IBM Spectrum Protect Plus server to fail and not be able to be restarted. This limitation can be identified in the Virgo log by:
"Too many documents, composite IndexReaders cannot exceed 2147483647"
Workaround: Complete the following procedure:
Important: File restore for all completed backups do not work after running this procedure. VMDK and full VM restores continue to work. Backup jobs begin file indexing and future file level restores will work based on the new backup jobs.

  1. Open an SSH session to the IBM Spectrum Protect Plus appliance with the serveradmin account. At the command line, enter:
    $ sudo rm /data3/lucene/lucidx/*
  2. After the removal process completes, start Virgo:
    $ sudo systemctl start virgo
  3. Exclude VMs that have file catalog enabled, which are not required to have file level restores, so that the maximum limit is not reached.

Limitation: Since V10.1.6.
 

IBM Spectrum Protect Plus does not use TLS encryption for SMTP (internal reference #SPPSUP-1514)
Problem: The IBM Spectrum Protect Plus server can be set to send email notification for reports. IBM Spectrum Protect Plus does not use TLS encryption for SMTP. As a result, when IBM Spectrum Protect Plus sends a report notification email, it might fail to be sent if the email provider requires TLS encryption.
Workaround: None.
Limitation:  Since V10.1.5.
 

"Offload" terminology appears after upgrading from IBM Spectrum Protect Plus V10.1.4 to V10.1.5 (internal reference #SPP-10526)
Problem: IBM Spectrum Protect Plus underwent a terminology change in version 10.1.5 where the word "offload" was changed to "copy." When upgrading from an IBM Spectrum Protect Plus version 10.1.4 appliance to version 10.1.5, the incorrect term "offload" appears in the Start Job Options drop down menu.
Workaround: None.
Limitation:  Since V10.1.5.
 

Restore type cannot be selected in the restore wizard in IBM Spectrum Protect Plus when using Firefox 60.9.0esr (64 Bit) on Windows (internal reference #SPP-10211)
Problem: When using Mozilla Firefox 60.9.0esr (64 Bit) on Microsoft Windows, the restore type cannot be selected in the restore wizard for IBM Spectrum Protect Plus.  Previous and more recent versions of Mozilla Firefox do not exhibit this behavior.
Workaround: Use a more recent version of Mozilla Firefox or another supported web-browser.
Limitation:  Since V10.1.5.
 

SQL and Exchange log backups fail due to vSnap with multiple network interface cards (NICs) (internal reference #SPPSUP-1253)
Problem: IBM Spectrum Protect Plus mail fail during Exchange and SQL log backups if the backup destination is a vSnap server that has multiple network interface cards (NICs) configured.
Workaround: To resolve this issue, the server message block (SMB) service needs to be disabled on the NICs on which Exchange or SQL server has no connection by issuing the following command on the vSnap server:

$ sudo vsnap network update –id --services “mgmt,repl,nfs,iscsi”

Re-attempt the failed backup job.
Limitation:  Since V10.1.4.
 

Jobs cannot be saved when scheduling a report with a frequency less than 5 minutes when using Microsoft Edge (internal reference #SPP-9353)
Problem: By default, a warning message displays in IBM Spectrum Protect Plus when attempting to schedule a report with a frequency that is less than 5 minutes. However, when defining a report schedule with a frequency that is less than 5 minutes with Microsoft Edge, the warning message does not display, and the job cannot be saved.
Workaround: Select frequencies that are 5 minutes or higher, or use a different supported browser when creating schedules.
Limitation: Solved with Microsoft Edge 41 or Microsoft EdgeHTML 16.16299 and later. Permanent restriction for earlier Microsoft Edge and Microsoft EdgeHTML browsers.
Limitation: Since V10.1.4.
 

In IBM Spectrum Protect Plus, after upgrading to version 10.1.4 and attempting to log in, the user might receive the message, "Username has expired and must be changed" (internal reference #SPP-9339 & SPP-6327)
Problem:  When upgrading to IBM Spectrum Protect Plus version 10.1.4, the browser might attempt to re-use previously cached credentials.  Upon the first log in for version 10.1.4, the user might receive the message "Username has expired and must be changed."
Workaround:  After upgrade to IBM Spectrum Protect Plus version 10.1.4 and before logging in for the first time, clear the browser cache.  For Chrome, Edge, or Firefox, press SHIFT+CTRL+DELETE to open the appropriate screen for clearing the browser cache.
Limitation:  Since V10.1.4.
 

Pointing an IBM Spectrum Protect Plus to another IBM Spectrum Protect object client node without unregistering and reregeristing the IBM Spectrum Protect endpoint can produce errors when running backup jobs (internal reference #SPP-8747)
Problem: In IBM Spectrum Protect Plus, when a IBM Spectrum Protect (SP) server is registered, it is recognized based on hostname and port as a unique endpoint, and the credentials are associated with that endpoint. If a new object client node is created on the same IBM Spectrum Protect server, and the credentials previously registered in IBM Spectrum Protect Plus are updated to point to the new node, then IBM Spectrum Protect Plus does not recognize that the object client node has changed. IBM Spectrum Protect Plus continues to treat it as the same endpoint as before. As a result, IBM Spectrum Protect Plus might try to refer to data that existed in the older node that is no longer present in the new node.
Workaround:  When a IBM Spectrum Protect repository server has been registered in IBM Spectrum Protect Plus using a particular object client node, the registration cannot be updated to point to another node. If the access key / secret key registered into IBM Spectrum Protect Plus are updated, the new keys must continue to be associated with the same node as before. To stop using the existing node and start using another node on the same IBM Spectrum Protect Plus server, remove any SLAs that point to the repository server and then unregister the repository server from IBM Spectrum Protect Plus.  Next, re-register the repository server using the keys associated with the new node and re-create the necessary SLAs to point to the new server.
Limitation:  Since V10.1.4.
 

The IBM Spectrum Protect Plus Administrator Console list of time zones is not translated to reflect local language selection (internal reference #SPP-7982 & SPP-8631)
Problem: If a user (administrator) of the IBM Spectrum Protect Plus enters the Admin Console and attempts to set the time zone, the list of time zones presented in the menu is not translated and not reflect that user's local language choice.  The time zone is not translated because it uses the timedatectl utility on the appliance to obtain this information.
Workaround: None.
Limitation:  Since V10.1.4.
 

IBM Spectrum Protect Plus does not support Unicode normalization (internal reference #SPP-6852 & SPP-8705)
Problem: When using the search fields in the IBM Spectrum Protect Plus product, glyphs that are visually equivalent but created with different code point representation might not be returned in the search results.
For example, a user from France might create a backup: sauvegarde_incrémentielle_220419. Depending on the “é” used in the string, the results might not be returned when searching because IBM Spectrum Protect Plus does not support Unicode normalization.
That is, (U+0065 U+0301) versus é (U+00E9) are two different representations of the same character in the string. Depending on the one being used in search, results might not be returned because the target of the search contained a character that was created using an alternate code point representation.
Workaround: When using any search box in IBM Spectrum Protect Plus, if it is suspected that characters such as the one above was used, always remember to include different variations of the code point representation in the search.
Limitation: Since V10.1.4. Permanent restriction.
 

IBM Spectrum Protect Plus uses third-party libraries that might not reflect language selection (internal reference #SPP-8301 & SPP-8633 & SPP-8517)
Problem: The IBM Spectrum Protect Plus product uses third-party libraries that might not reflect the user’s local language selection.  Therefore, errors or exceptions from these libraries might not adhere to the language selection.
Workaround: None.
Limitation: Since V10.1.4. Permanent restriction.
 

The IBM Spectrum Protect Plus Administrator Console doesn’t display the time and date in cultural format (internal reference #SPP-7984 & SPP-8632)
Problem: If a user (administrator) of the IBM Spectrum Protect Plus enters the Admin Console and that user has the web-browser configured to use a region setting other than that of the United States, the time and date might not be displayed in the proper cultural format. The time and date uses the timedatectl utility on the appliance to obtain this information.
Workaround: None.
Limitation: Since V10.1.4. Permanent restriction.
 

When a report is downloaded to PDF format, some non-English characters might not be displayed properly (internal reference #SPP-1634)
Problem: IBM Spectrum Protect Plus provides a mechanism whereby job history can be downloaded in PDF format.  If elements in the report contain non-English characters, it might be the case that these strings are not displayed in the PDF correctly.  This occurs due to the version of a third-party framework that is used by IBM Spectrum Protect Plus for PDF rendering.
Workaround: None.
Limitation: Since V10.1.4. Permanent Restriction.
 

IBM Spectrum Protect Plus sorting order is not in ICU collation order (internal reference #SPP-6851 & SPP-8634)
Problem: The IBM Spectrum Protect Plus product does not follow ICU collation sorting for menus, therefore, the ordering of menus will appear in code point order.  For example, some languages sort letters differently from code point order. As such, the sorted order of characters and words as they appear in menus when using these languages will appear out of expected order.
Workaround: None.
Limitation: Since V10.1.4. Permanent restriction.
 

vSnap hangs due to connectivity failure
Problem:  Offload to the Cloud or a Repository server might fail due to network connectivity issues. This might prevent success in further offloading for the same volume.
Workaround:  Restart the vSnap to discard the suspended pool and resume routine operations.
Limitation:  Since V10.1.3.
 

Admin Console user interface still displays trial license after updating to perpetual license (internal reference #SPP-7043)
Problem: The Admin Console user interface displays the trial license information after updating to the perpetual license.
Workaround: Clear your browser cache and reload the Admin console to show the correct information.
Limitation:  Since V10.1.3.
 

Application Server can be deleted even during an active restore job (internal reference #SPP-6796)
Problem: Application servers defined within IBM Spectrum Protect Plus can be deleted by a user even during an active restore job. This might result in subsequent cleanup failures.
Workaround: Check for active restore jobs pointing to the Application Server before deletion.
Limitation:  Since V10.1.3.
 

Backups of a failover cluster with SQL DBs residing on Cluster Shared Volume always runs as base backup (internal reference #SPP-6580)
Problem: SQL database backups residing on Cluster Shared Volume (CSV) always runs as base backup since change block tracking is not supported on CSV file system.
Workaround: Failover clusters with DBs running on CSV file system is not supported.
Limitation:  Since V10.1.3.
 

SLA policy and backup option selections for a folder are not retained if the folder is moved (internal reference #SPP-4388)
Problem: If a folder is moved through the vCenter interface after SLA policies or backup options are assigned to the folder and an Inventory job is run, the selections are not retained in IBM Spectrum Protect Plus. Any virtual machines within this folder that were previously backed up is no longer backed up.
Workaround: After a folder is moved and its new location is logged by the Inventory job, re-associate the folder with the SLA policy and reassign the backup options.
Limitation:  Since V10.1.2.
 

If the latest restore point of a resource is manually expired, older versions of the restore point do not display on Restore screens (internal reference #SPP-4220)
Problem: If the latest restore point of a resource is manually expired through the Restore Point Retention screen, older versions of the restore point do not display on associated Restore screens.
Workaround: Run the backup job again to capture the latest restore point. When complete, older versions of the restore point do display on associated Restore screens.
Limitation:  Since V10.1.2.
 

Recurring VM restore fails at second restore time (internal reference #SPP-10076)
Problem: Configure a reoccurring restore in clone mode can't be combined with specifying a dedicated VM name. The second and all following restore operations fail due to naming and UUID conflicts in the vCenter.
Workaround: Do not specify a dedicated VM name when configuring a reoccurring restore in clone mode. If a name is specified, delete or rename the previously created VM, and re-run the job.
Limitation: Since V10.1.2, Permanent restriction.
 

After a catalog restore completes on a target appliance, the associated catalog backup job returns an ABORT state on the destination appliance (internal reference #SPP-994)
Problem: After a catalog restore completes on a target IBM Spectrum Protect Plus appliance, the catalog backup job that is carried over to the new appliance initially shows that the job last ran in an ABORT state.
Workaround: This is expected due to the snapshot transition. When the catalog backup is run on the source, the snapshot is taken while the catalog backup is running.
Limitation: Since V10.1.1. Permanent restriction.
 

Restored VMware VM does not have local permissions set (internal reference #SPPSUP-1772)
Problem: After doing a production restore of a VMware VM, the local user permissions are not set.
Workaround: Manually set the permissions through vCenter.
Limitation: Since V10.1.0.
 

Certificates in the .p12 format cannot be imported through the administrative console (internal reference #SPP-9355)
Problem: The administrative console supports importing external public key certificates, which allows the appliance to make outbound connections to resources, such as secured LDAP and secured HTTPS servers. The certificate import feature supports .cer and .crt certificate types. Note that .p12 certificates can only be imported into the IBM Spectrum Protect Plus keystore.
Workaround: To import the .p12 certificates into the keystore, follow the procedure:
Log in as root user on the IBM Spectrum Protect Plus appliance and at the command line execute the following command:

“/usr/java/latest/bin/keytool -importkeystore -deststorepass ecx-beta -destkeystore /opt/virgo/configuration/keystore -srckeystore cs.p12 -srcstoretype PKCS12"

Then restart the appliance.
Limitation: Since V10.1.0.
 

In IBM Spectrum Protect Plus, permission groups specify the ability to edit a certificate, but no edit button is displayed on the Keys and Certificates screen (internal reference #SPP-8611)
Problem: Certain pre-defined roles have assigned permission groups that allow for creating, viewing, editing, or removing certain features of IBM Spectrum Protect Plus.  In particular, the Application Admin, Backup Only, Restore Only and SYSADMIN groups have the Edit ability for Certificates.  However, if a user that is assigned to one of these groups views certifications on the Keys and Certificates screen, there is no Edit button.  Currently, there is no way to edit a certificate from the Keys and Certificates screen.
Workaround: None.
Limitation: Permanent restriction.
 

IBM Spectrum Protect Plus system log sizes grow in some situations (internal reference #SPPSUP-374)
Problem:  IBM Spectrum Protect Plus does not offer a mechanism to limit log sizes to a specific size. If the system log file is too large the file system might be filled and the download of this log file is difficult. Logs that contribute to large deviations are however regularly reviewed for its usefulness to troubleshooting.
Workaround: Make sure that logs are reviewed and removed regularly.
Limitation: Permanent restriction.
 


 


Hypervisor VMware related

IBM Spectrum Protect Plus backup of Windows Server 2019 fails with error: “Error: com.vmware.vim25.ApplicationQuiesceFault” (internal reference #SPP-9873)
Problem: Running a backup job of Windows Server 2019 without enabling the option “Fall back to unquiesced snapshot if quiesced snapshot fails” can fail with the error message:
“Error: com.vmware.vim25.ApplicationQuiesceFault An error occurred while quiescing the virtual machine.”
Workaround: Disable the VMware Snapshot Provider service in the guest operating system and allow the Microsoft Volume Shadow Service to run automatically.  When this has been done, re-attempt the backup job that failed with the error.
Disable the VMware Snapshot Provider service in the guest operating system:

  1. Log in to the guest operating system using administrator credentials.
  2. On the keyboard, click the Windows key +R to open the Run dialog box.
  3. Type services.msc and press the Enter key.
  4. Locate and right-click on the VMware Snapshot Provider entry and left-click on Properties.
  5. Set Startup type to Disabled and click Apply and then OK.

Ensure that the Microsoft Volume Shadow Service is set to automatic in the guest operating system:

  1. Log in to the guest operating system using administrator credentials.
  2. On the keyboard, click the Windows key +R to open the Run dialog box.
  3. Type services.msc and press the Enter key.
  4. Locate and right-click on the Volume Shadow Service entry and left-click on Properties.
  5. Set Startup type to Automatic and click Apply and then OK.

Re-attempt any backup jobs that failed as a result of the error. For more information see https://kb.vmware.com/s/article/60395
Limitation: This issue is resolved in Windows Server version 1809 and all versions of Windows Server 2019 OS Build 17763.1012. For more information, see January 23, 2020—KB4534321 (OS Build 17763.1012)
Limitation:  Since V10.1.3. Permanent restriction.
 

An RBAC user assigned a VMWare Tag and Categories along with either Hosts and Clusters or VMs and Templates results in data not being displayed (internal reference #SPP-8913)
Problem: Users that have been assigned to the Tag and Categories resource group and either Hosts and Clusters or virtual machines and Templates will result in data not being displayed for the Hosts and Clusters view or virtual machines and Templates view.  Only information for Tags and Categories will be displayed when that is selected as a view when running backup operations.
Workaround: Do not assign a user to more than one VMware resource group.
Limitation: Permanent restriction.
Limitation:  Since V10.1.4. Permanent restriction.
 


 


Hypervisor Hyper-V related

Backup and restore operations for Hyper-V shared virtual hard disks fails (internal reference #SPP-1108)
Problem: When attempting to run a backup of a Hyper-V shared virtual hard disk (shared VHDX), the backup fails with the following error:
“Failed to create Snapshot .ava.lang.NullPointerException.”
Shared virtual hard disks in a Hyper-V environment are not supported in IBM Spectrum Protect Plus.
Workaround: None.
Limitation: Since V10.1.1. Permanent restriction.
 


 


SQL application server related

IBM Spectrum Protect Plus limits the log backup of a database from only one SLA (internal reference #SPPSUP-703)
Problem: The log backup of a database is limited to only one SLA. If there are two SLAs, one to protect an SQL Server Always-On availability groups (AG) and another to protect databases on a stand-alone instance, if a database is moved from AG to stand-alone, or stand-alone to AG, log backup for that database might fail from the SLA that protects the target of the move.
Workaround: After moving the database from the stand-alone to AG or from AG  to stand-alone, run the SLA that was protecting the database before the move once more. Doing this will remove the log backup for that database. When that is completed, run the second SLA that will protect the database at the new location. That will create the log backup for that database.
Limitation: Since V10.1.1. Permanent restriction.
 


 


ORACLE application server related

An entry in ‘srvctl’ is not created after a successful Oracle Database restore to a RAC destination (internal reference #SPP-843)
Problem: After successfully restoring an Oracle Database to a RAC destination, IBM Spectrum Protect Plus does not automatically create an entry in 'srvctl' for the restored database.
Workaround: An entry in ‘srvctl’ must be manually created if wanted.
Limitation:  Since V10.1.1.
 


 


Db2 application server related

Database is subtly excluded from scheduled backup when volume group is running out of free space (internal reference #SPP-db2-179999 & #SPP-9108)
Problem: During inventory, volume groups holding data of databases is checked for free space to ensure that temporary snapshots can be successfully created during the backup operation. In case of there is not sufficient free space available, the database is declared to be ineligible for backup with the according reason message. However, the free space can be sufficient initially and the database has been added to a SLA policy. When the free space declines successively it might fall below the minimum limit. A warning message will indicate the situation, however, it is not rated as a failed backup - the database will just be silently skipped. The summary of the scheduled backup run will just show a decreased number of protected databases.
Workaround: Create backup reports frequently and verify that all important databases have recent backups. Monitor the free space of volume groups and ensure to expand it timely when necessary.
Limitation:  Since V10.1.2.
 

In IBM Spectrum Protect Plus, the log backup volume is not unmounted when log backup enablement fails (internal reference #SPP-8833)
Problem:  When log backup is not enabled in IBM Spectrum Protect Plus and then it gets enabled, the log backup enablement command is executed so that IBM Spectrum Protect Plus will mount a log backup volume.  Should the agent fail in enabling the log backup, even though it gets reported, the host agent does not unmount the log backup volume and it remains in a mounted state. The mounted volume will continue to be used for future log backup attempts.
Workaround: None
Limitation:  Since V10.1.4.
 

Application cataloging fails when using fully qualified hostnames in db2nodes.cfg (internal reference #SPP-7173)
Problem:  When hosts in db2nodes.cfg are specified by its fully qualified names (including dots), cataloging of the database instance fails on inventory with the following error message:
Cataloging failed for server <servername>:org.springframework.data.mapping.model.MappingException: Map key <hostname> contains dots but no replacement was configured! Make sure map keys don't contain dots in the first place or configure an appropriate replacement!
Workaround: Ensure to have only short hostnames in db2nodes.cfg
Limitation:  Since V10.1.4.
 

Archive log backups are limited to a single SLA policy per database (internal reference #SPP-db2-175439)
Problem: When archive log backups are enabled on the Db2 backup options panel, it applies to all SLA policies the database is associated with.
For each SLA policy, an extra archive log backup target is created on the vSnap repository. However, Db2 can only handle a maximum of two log archive paths in parallel, provided there are no other archive log destinations in use (e.g. Tivoli Storage Manager). While there might be two SLA policies associated were applicable, it is highly recommended to stay limited to a single SLA policy per database when log backup is enabled. Otherwise, disabling log backup or disassociating SLA policies can lead to situations requiring manual intervention. See below for details.   
Workaround: When the log backup option is enabled for a distinct database, ensure that this database is only associated with a single SLA policy.
If two SLA policies have been associated nonetheless, you will encounter issues when either the log backup is disabled or when one of the SLA policies is disassociated. The result is the Db2 configuration parameters logarchmeth1/2 cannot be reset to its original values leading to the following error message:
"Log backup for database <DB name> is not disabled because otherwise the database will no more be configured for rollforward recovery. As a consequence, a restart of the Db2 database would be required."
If you encounter this situation, reset the logarchmeth1/2 parameters manually to its original settings or to LOGRETAIN.
Example:

db2 UPDATE DB CFG USING logarchmeth1 LOGRETAIN logarchmeth2 OFF

Limitation:  Since V10.1.4.
 

Redirected restore operations to alternative locations, require path names to include the instance and database names (internal reference #SPP-db2-175655)
Problem: When a database is restored to an alternate instance name or database name, the disk paths must include the instance name and the database name.
This detail is required so that the path names can be altered during the relocation procedure. The condition applies to all types of paths: Database paths, container paths, storage paths, log and mirror log paths.
It is best to include both the instance name and the database name in all path names to enable the full set of restore capabilities, including restoring the same database to the same machine multiple times with different names and instances. If the path names do not include the instance name and database name, the following restrictions apply:

  1. If one or more of the path names includes the database name but not the database instance name.
    Restriction: To restore multiple copies of the database to a single target machine at the same time, each copy needs an individual database name assigned even when each restore operation targets another database instance.

  2. If one or more of the relevant path names includes the database instance name but not the database name.
    Restriction: To restore multiple copies of the database to a single target machine at the same time, each copy needs to be restored to a separate database instance even when an individual database name has been assigned to each copy.

  3. If one or more of the path names does not include either the database name or the database instance name.
    Restriction: Only a single copy of the database can be restored to a single target machine at the same time. Both database and instance name need to match the original names.

Workaround: Use a dedicated target machine for each restore.
Limitation:  Since V10.1.3.
 


 


MongoDB application server related

In rare cases the arbiter node in a distributed MongoDB Replica Set crash during IBM Spectrum Protect Plus inventory (internal reference #SPP-mongo-182265)
Problem: If a MongoDB Replica Set is distributed across multiple hosts and contains an arbiter node that resides on a SLES 12 server, the arbiter node can crash during the IBM Spectrum Protect Plus inventory operation. This problem is caused by the external Python library pymongo that is leveraged by the IBM Spectrum Protect Plus MongoDB agent. The arbiter node can be restarted manually.
Workaround:

Limitation:  Since V10.1.4.
 


Exchange application server related

Exchange Mailbox Restore Configuration Wizard display the Media Check entry. (internal reference #SPP-exchange-1168)
Problem: As running IBM Spectrum Protect Plus - Exchange Mailbox Restore Configuration Wizard to check the requirements items, the Media Check entry is displayed as failed. The media check is obsolete for SPP Exchange thus creating a file 'SetupFCM.exe' is needed to get past the check.
Workaround: Find any valid Windows executable file(.exe ) and rename it to SetupFCM.exe. After Media Check failed, Click the “Failed” link on the configuration wizard panel. Specify the SetupFCM.exe above in popped up window. The media check succeeds.
Limitation:  Since V10.1.6.
 

The ExchSchd directory and Windows Scheduler do not sync after enablelogbakup or disablelogbackup commands (internal reference #SPP-exchange-1137)
Problem: In normal runtime condition, the C:\ProgramData\SPP\ExchScheds directory entries keep synchronized with Windows Scheduler entries. They are all required by IBM Spectrum Protect Plus for Exchange agent to run log back up correctly.
In some rare conditions, such as the user removes or modify them partially (this should never be done manually by user, they should do it on Spectrum Protect Plus console GUI). Some invalid ExchScheds directory or Windows Scheduler entries are left. This might cause confusion to the user.
Workaround: There is no workaround. But it's safe to leave all the entries there. There is no data lose and extra system resource consumption.
Limitation:  Since V10.1.6.
 

During log backup on the exchange machine, if the log backup scheduler is trying to be updated, then the scheduler might be corrupted. (internal reference #SPP-exchange-1162)
Problem: Logbackpup process will open and hold log files under logbackup scheduler folder. If an disablelogbackup command comes when logbackpup running at backend (same scheduler), and if IBM Spectrum Protect Plus Exchange agent try to delete the scheduler folder in handling the disablelogbackup command, the disablelogbackup will fail by the reason of "failed to delete the scheduler folder".

This problem might happen in this condition:

  • Launch the scheduled task at agent machine to run logbackup of DB1
  • Disable logbackpup of DB1 and run related SLA to trigger the disablelogbackup command.
    If DB1 is the last DB in the logbackpup scheduler, then Exchange agent will try to delete the scheduler folder.
    If the logbackpup of this scheduler is still running at backend, the disablelogbackup fails.

When this problem happens, the disablelogbackup command will fail and the overall status of backup job will be PARTIAL.
Workaround: Since the disablelogbackup command failed, IBM Spectrum Protect Plus server will send another disablelogbackup command in next backup job. So if the logbackpup is not running at that time, the next disablelogbackup will remove the folder successfully.
Limitation:  Since V10.1.6.
 

Deploy granular restore package to a remote server that has Microsoft Outlook installed does not work as described in readme file (internal reference ##SPP-exchange-1170)
Problem: It is not possible to deploy granular restore by following the deployment steps in the readme file. The reason is that the SetupFCM.exe does not exist in the granular restore deployment package.
Workaround:  to be able to deploy granular restore follow the instructions:

  1. Copy granular restore package from Exchange Application Server
    C:\Program Files\IBM\IBM Spectrum Protect Plus\tools\exchange\imr
  2. Paste to remote server where you will manage the granular restore operation from.
    C:\temp
  3. On the remote server, create deployment diagnostics folder.
    mkdir C:\temp\diag
  4. On the remote server, run the following to deploy MMC GUI and granular restore components.
    c:\temp\imr\install_imr.bat <version_nr>-TIV-TSMEXC-Win.exe 10.1.6 c:\temp

Limitation:  Since V10.1.6.
 

Unsupported characters are displayed as '?' in MMC GUI mailbox restore browser (internal reference #SPP-6054)
Problem: If mailbox is created without specifying mailbox alias, Exchange server uses the value of 'DisplayName' parameter for the 'Alias' property value. If 'DisplayName' contains characters outside of Unicode range from U+00A1 to U+00FF, then Alias will contain unsupported characters.
Exchange removes unsupported characters from 'Alias' by converting them to question marks. As a result, MMC GUI mailbox restore browser shows mailboxes with some characters displayed as '?' question marks.
For more information, see Microsoft documentation: https://docs.microsoft.com/en-us/powershell/module/exchange/mailboxes/new-mailbox?view=exchange-ps
Workaround: When creating public folder mailbox or user mailbox, explicitly set mailbox alias value and use supported characters in mailbox alias.
Limitation: Since V10.1.3. Permanent restriction.
 


 


Office 365 application server related

IBM has tested Office 365 support in IBM Spectrum Protect Plus with 1000 Exchange Online & OneDrive users.
 

Restoring the calendar event will fail if the event doesn't have correct time zone (internal reference #SPP365-100)
Problem: When creating a calendar event using e.g. PowerShell, it is possible to create an event without a correct time zone. Such an event can be protected but cannot be restored.
Workaround: None.
Limitation:  Since V10.1.5.
 

The backup job will hang until all data to be protected are processed if there is no free space left at the backup destination (internal reference #SPP365-68)
Problem: When there is no free space left on the backup destination, the agent will anyway try to process all objects that should be protected.
Workaround: Add additional space to the backup destination.
Limitation:  Since V10.1.5.
 


 


Kubernetes application server related

No limitations for Kubernetes application server.
 


 


File System application server related

After deployment of the File System Restore Service a local installed Python version cannot be used anymore (internal reference #SPP-13033)
Problem: The File System Restore Service changes the global settings for all users. After that, an installed Python version cannot be found anymore. That is happening for all user on the system. All Python using programs on the System will fail to run properly.
Workaround: To solve this behavior:

  • Uninstall the global Python version and install for all user that needs Python a personal version, or
  • Change for every user that needs Python the local path settings.
    Make sure that you do not change the global path settings. If the global path settings are changed a restore is not possible anymore.
    Also make sure not to change the settings of the administrative user that is used for the File System Agent.

Limitation:  Since V10.1.6.
 

File system restore job might remain in state 'running' (internal reference #SPP-13032)
Problem: Running restore it might happen that a restore task remains in state 'Running' although it has finished successfully. This is a display error only.
Workaround: Check the file system if the files and folders subject to restore have been restored successfully. Then cancel the restore operation in the GUI.
Limitation:  Since V10.1.6.
 

First cancelled restore job makes the second job on the same system unusable (internal reference #SPP-13071)
Problem: When two or more file system restore jobs are created for the same system and one of the jobs is canceled, then the restore GUI for the other jobs might no longer be usable.
Workaround: Cancel **restore jobs not before all objects subject to restore from all file systems are finished.
Limitation:  Since V10.1.6.
 

Protection of stub files is not supported (internal reference #SPP-13082)
Problem: Stub files (e.g. IBM Spectrum Protect HSM for Windows migrated files) are not supported for backup and restore. When trying to restore such a file, the operation will fail. 
Workaround: Do not backup volumes, which include stub files or exclude all folders containing such files.
Limitation:  Since V10.1.6.
 

Inventory not detecting all file systems when VM was cloned or created from a template (internal reference #SPP-13143)
Problem: The Windows Filesystem inventory in IBM Spectrum Protect Plus is affected by a commonly known issue when the file server instance was created from a template or cloned from other VMs. In this case the Universally Unique Identifier (UUID) for the Virtual Machine Disks (VMDK) could be the same for original and the cloned VMs. Since the IBM Spectrum Protect Plus UI is using the VMDK UUID to relate the volume to a File Server instance, it only displays one instance of the volume in the UI. All other volumes with identical UUID are not displayed even though the inventory detected these.
Workaround: Microsoft and VMware documented procedures how to solve the problem with the duplicated GUID for the volumes. The basic idea is to change the UUID for the VMDK after creating the VM from template or cloning. Refer to: https://kb.vmware.com/s/article/2006865
Limitation:  Since V10.1.6.
 


 


Limitations solved with V10.1.6

Restore multiple DBs failed to mount the databases due to files removed (internal reference #SPP-10523)
Problem: Normally Exchange will replay the log files after restore completed. But with the “No Recovery” option selected, we don’t want Exchange to replay the logs after restore. So we must call VSS SetAdditionalRestores() method. The SetAdditionalRestores() will prevent the log files to be replayed. With a wrong logic, the SetAdditionalRestores() will only be called in the “restore to alternate location” case. So in the “restore to original location” case, the log files will always be replayed even the “No Recovery” selected.
Problem Verification: Restore to original location (with empty “New Database Name” or “New Database Name” is same as original database name) and select the “No Recovery” option. After restore, the restored log files will still be replayed and deleted.
Workaround: None.
Limitation: Since V10.1.5. Solved with V10.1.6.
 

Granular restore operation fails with missing required Exchange Roles for the current user (internal reference #SPP-exchange-946)
Problem: During granular restore, in the granular restore UI, a configuration error message pops up. If the user does not have the user rights listed, the protection operations fail. The user will not have the ability to run granular restore until the user rights are updated.
Workaround:
1. Verify that the user domain used to register the server
2. Click OK and add the indicated Exchange roles for the specified user
Limitation: Solved with V10.1.6.
 

The organization name used for the registration of a Microsoft Office 365 tenant must not include dots (internal reference #SPP365-59)
Problem: When the organization name used for the registration of a Microsoft Office 365 tenant includes dots in the name, the restore will not work.
Example: Using the full organization name "my_name.onmicrosoft.com" the backup will work, but the restore will fail.
Workaround: Use an organization name that has no dots in it.
Example: Using the organization name "my_name".
Limitation: Since V10.1.5. Solved with V10.1.6.
 

Not able to restore an on-demand backup immediately. (internal reference #SPP-Kubernetes-597)
Problem: "kubectl describe" of an on-demand backup shows the backup is done but it is not restorable until the backup is reflected in the schedule backup request.
Workaround: Wait until the on-demand backup is reflected in the schedule backup request before trying to restore the on-demand backup.
Limitation: Solved with V10.1.6.
 

SQL backup failed with a confusing error message when backup LUN was disconnected at the end of the backup. (internal reference #SPP-10623)
Problem: The volume mount status was not validated before access the backup LUN.
Workaround: No workaround, job failed.
Limitation: Since V10.1.5. Solved with V10.1.6.
 

The "Unregister" option for a VADP Proxy appears as an empty string (internal reference #SPP-10802)
Problem: In the IBM Spectrum Protect Plus GUI the VAPD proxy page under the system configuration, select a VADP proxy from the list and then chose the 3 dots in the right top corner to bring up the action for VADP proxy. The Unregister option for a VADP Proxy appears as an empty string.
Workaround: Clicking on the empty string in the drop down of options initiates the unregister action.
Limitation: Since V10.1.5. Solved with V10.1.6.
 

Backup target LUNs on vSnap storage do not automatically expand for Hyper-V backups (internal reference #SPP-2719)
Problem: For Hyper-V backups, if the source data store is expanded, the backup target LUN on vSnap storage does not automatically expand.
Workaround: None.
Limitation: Since V10.1.1. Solved with V10.1.6.


 


Limitations solved with V10.1.5

In IBM Spectrum Protect Plus, when a data center is associated with an SLA, the number of backed up VMs might be more than the number of VMs in the data center (internal reference #SPP-8509)
Problem: An NFS volume can be mounted to any number of data centers in the same vCenter. When an NFS volume is mounted on more than one data center on a vCenter, even though vCenter treats the same volume as two different data stores, IBM Spectrum Protect Plus treats this single data store and combines all the VMs and VMDKs residing on this data store from all the data centers this data store is mounted on. Therefore, any SLA selection/resolution against this data store will cause all the VMs from the different data centers to be backed up.
Workaround: None.
Limitation: Since V10.1.1. Solved with V10.1.5, see APAR IT30102.
 

User Interface allows cross-platform restore operations (internal reference #SPP-7292)
Problem: In environments with machines of different OS platforms (e.g. AIX and Linux x86) the user interface allows the target instance to be of a different OS platform. Hence, you can trigger the restore of a Linux database to a database instance on AIX and vice versa. Even though the user interface does not prevent you from doing so it will lead to unexpected results.
Workaround: Ensure that the selected target instance matches the original OS platform.
Limitation: Since V10.1.4. Solved with V10.1.5.
 

Data Protection for Exchange 8.1.8 and IBM Spectrum Protect Plus 10.1.4 Granular Restore GUI cannot coexist on the same application server (internal reference #SPP-exchange-978)
Problem: When IBM Spectrum Protect Plus is used to protect a Microsoft Exchange Server where 8.1.8 Data Protection for Exchange is already installed, if the user attempts to run a granular restore operation, the operation will have unexpected behavior. The job log will not log a failure. When the user remotely connects to the application server, the granular restore UI will not display correctly and some functions will not work.
Workaround:
1. Cancel the granular restore operation, if it is in the Resource Active state
2. On the Exchange server, close the granular restore or MMC UI
3. Completely uninstall 8.1.8 Data Protection for Exchange
4. Rerun granular restore
Limitation: Since V10.1.4. Solved with V10.1.5.
 

When upgrading from IBM Spectrum Protect Plus version 10.1.3 to version 10.1.4, manually running a previously created policy (10.1.3) that has an associated sub-policy added will result in a null-pointer exception (internal reference #SPPSUP-1115)<
Problem: In IBM Spectrum Protect Plus, after upgrading from version 10.1.3 to version 10.1.4, manually running a previously created SLA policy that also has an associated sub-policy will result in a null-pointer exception. The automatic running of the SLA policy will work correctly.
Workaround: Edit the original SLA policy, but do not change anything with the policy and then click Save. Next, manually start the policy. The user will then be prompted to select the sub-policy.
Limitation: Since V10.1.3. Solved with V10.1.5.
 

Inventory fails for Db2 databases using non-ASCII characters in path or file names (internal reference #SPP-5965)
Problem: Databases with non-ASCII characters in path or file names (Database paths, container paths, storage paths, log and mirror log paths) cannot be inventoried due to the following error:

Command line '/usr/bin/python /tmp/tmpWzEDYB' has ended with exit code 1. stdout: ===SPPAgent-output===
stderr: File "/tmp/tmpWzEDYB", line 2 SyntaxError: Non-ASCII character '\xc7' in file /tmp/tmpWzEDYB on line 2, but no encoding declared; see http://www.python.org/peps/pep-0263.html for details

Workaround: Ensure that all path and file names of the database are built from ASCII characters only.
Limitation: Since V10.1.3. Solved with V10.1.5.
 

Redirected restore operations to alternative locations will change ownership and permissions for all database paths up to the root file system (internal reference #SPP-db2-179927)
Problem: When a database is restored to an alternate instance name or database name, all disk paths of the database will be altered to the new names during the relocation procedure that is running on the target machine following the actual restore and recovery.
To ensure the database management system has access to the relocated paths, the ownership and permissions of the paths are also updated. This applies to the complete path chain up to but not including the root file system ("/"). When the operation is completed, only the database instance user and other members of its group still have write access to the full chain of these paths. These conditions can lead to issues given that the top-level directories of the altered paths are representing directories that other users need to have write access to, too.
The condition applies to all types of paths: Database paths, container paths, storage paths, log and mirror log paths.
Workaround: Ensure that all database paths of the database are dedicated to the database the full chain up to the root file system so that exclusive write permissions is acceptable.
Limitation: Since V10.1.3. Solved with V10.1.5.
 

Unused objects are not all being deleted after cloud offload (internal reference #SPP-5849)
Problem: When an SLA policy is deleted and all the recovery points are purged out from vSnap, some of the objects that are not being used might still exist in cloud buckets.
Workaround: None
Limitation: Since V10.1.3. Solved with V10.1.5.
 

Error messages returned from Oracle servers running non-English languages might not be displayed correctly (internal reference #SPP-6193)
Problem: When IBM Spectrum Protect Plus experiences errors while attempting to backup or restore an Oracle Database, it attempts to capture the error messages returned by Oracle and display them in the backup/restore job log. If the Oracle Database runs with the NLS_LANGUAGE set to a non-English language by default, then the error messages might contain characters that cannot be formatted or displayed correctly in the job log. As a result, those specific error messages might not be readable when the job log is viewed or downloaded.
Workaround: While the message itself might not be readable, the message prefix will contain a standard error code starting with "ORA" or "RMAN". Use Oracle support/documentation resources to look up the message associated with the error code.
Limitation: Since V10.1.3. Solved with V10.1.5.
 

Restore of a Db2 database with renaming might break ssh communication (internal reference #SPP-db2-185334)
Problem: When a database is restored to an alternate instance name and / or database name, all disk paths of the database will be altered to the new names during the relocation procedure that is running on the target machine following the actual restore and recovery. To ensure the database management system has access to the relocated paths, the ownership and permissions of the paths are also updated. This applies to the complete path chain up to but not including the root file system ("/"). The permission mask being used for this operation is 0775.  
For the case that data or log paths are subdirectories of the Db2 instance user home directory those permissions are updated, too. This can cause issues with key based ssh authentication given that ssh is configured in strict mode. In strict mode key files and the directories that they are stored in must not be writable by anyone but the owner so the mask must be 0755 or even less for group and others.  The broken ssh connectivity will cause the database instance to be nonoperational given that it is multi-partitioned (DPF) and configured for ssh.
Workaround: 
1.
Set "StrictModes" to "no" in sshd_config and restart sshd service.
or
2. Correct permissions on Db2 instance user home directory after the restore operation is completed.
Limitation: Since V10.1.2. Solved with V10.1.5.
 

vSnap password fails when attempting to access CIFS share with the associated account credentials (internal reference #SPPSUP-561)
Problem: The vSnap credentials registered with IBM Spectrum Protect Plus do not work when trying to access the CIFS share when using the same account credentials. The credentials will need to be re-synchronized to ensure continued access.
Workaround: The vSnap password should be changed to the same value to manually force synchronization. In this example, the username is vsnapuser Credentials can be manually synchronized by entering the following commands on the vSnap Server:

  • Update the operating system credentials, which are used for management communication through the vSnap API:
    sudo passwd vsnapuser 
  • Update the CIFS credentials, which are used for communication between the application server and vSnap:
    sudo smbpasswd vsnapuser 
    Finally, update the vSnap registration in IBM Spectrum Protect Plus to ensure that the updated password is entered into the system. At this time, you might re-run the job that failed as a result of the problem.

Limitation: Since V10.1.2. Added in IBM Knowledge Center
 

First entries in lists automatically display in menus when using Microsoft Edge (internal reference #SPP-9354)
Problem: When selecting options from drop-down menus in IBM Spectrum Protect Plus, the first entry in a list automatically displays in the menu when using the Microsoft Edge browser; however, the item is not selected. In addition, if there is one entry in a drop-down menu, the item cannot be selected when using the Microsoft Edge browser.
Workaround: None.
Limitation: Since V10.1.4. Solved with V10.1.5.
 

[{"Business Unit":{"code":"BU054","label":"Systems w\/TPS"},"Product":{"code":"SSNQFQ","label":"IBM Spectrum Protect Plus"},"ARM Category":[],"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"All Version(s)"}]

Document Information

Modified date:
03 July 2020

UID

ibm16209657