IBM Support

Version 7.1: Updates for IBM Tivoli Storage Manager for UNIX and Linux Backup-Archive Clients Installation and User's Guide

Product Documentation


Abstract

This document contains updates for "IBM® Tivoli® Storage Manager for UNIX and Linux Backup-Archive Clients Installation and User's Guide V7.1". The updates are for the Version 7.1 product information at http://www.ibm.com/support/knowledgecenter/SSGSG7/landing/welcome_ssgsg7.html.

Content

Documentation updates

Updates that apply to V7.1.8
Updates that apply to V7.1.6
Updates that apply to V7.1.4 and later versions
Updates that apply to V7.1.3 and later versions
Updates that apply to V7.1.2
Updates that apply to V7.1.1 and later versions
Updates that apply to all modifications of V7.1


 

Updates that apply to V7.1.8

Per reference #197071 
Messages, return codes, and error codes > ANS 0000-9999 messages > ANS messages list

The following message has been added:

ANS9419W
VMware Changed Block Tracking (CBT) cannot be enabled or reset for virtual machine 'virtual machine name' due to an existing snapshot. Full VM backup continues, and includes both used and unused areas of the disk.
 
Explanation
VMware does not allow changing state of Changed Block Tracking(CBT), when a snapshot exists. The backup continues, but instead of backing up just the used blocks (in the full VM backup case), or just the changed blocks (in the incremental VM backup case), the entire virtual machine is backed up. This backup includes both the used and unused blocks of the disk.
 
System action
The backup continues.
 
User response
Delete snapshot on the target vm to allow CBT to be enabled or reset.
Per APAR IT23560
Clients > Configuring backup-archive clients > Configure the Tivoli Storage Manager client > Configuring Tivoli Storage Manager client/server communication with Secure Sockets Layer > Creating a symbolic link to access the latest GSKit library

The following text has been added:

Before you begin
  • • An IBM Spectrum Protect™ client, V7.1.8 and later V7 levels, and V8.1.2 and later levels, requires GSKit version 8.0.50.78.
    • An IBM Spectrum Protect client, V7.1.7 and earlier levels, and V8.1.1 and earlier V8 levels, requires a version of GSKit earlier than version 8.0.50.78.


Per APAR IT23086
Clients > Configuring backup-archive clients > Configure the Tivoli Storage Manager client > Automated client failover configuration and use > Configuring the client for automated failover

The following text has been added:

Note
If the replication server is V7.1.7 or earlier, and SSL is enabled, you must manually install the SSL certificate on the client with the following command:
  • gsk8capicmd_64 -cert -add -db dsmcert.kdb -stashed -label "TSM server STSM01 self-signed key" -file <certificate_file> -format ascii

Where <certificate_file> is the path to the corresponding certificate.

 

Updates that apply to V7.1.6



Per APAR IT20791
Clients > Backup-archive client options and commands > Processing options > Client options reference

In the Enableinstrumentation, Instrlogmax, and Instrologname options, the following statement is removed:

"The option can be set in the client option set on IBM Spectrum Protect server."

Per APAR IT20793
Clients > Configure the IBM Spectrum Protect client > Configuring NetApp and IBM Spectrum Protect for snapshot difference incremental backups > Protecting clustered-data ONTAP NetApp file server volumes

In Step 2, the order of the arguments in the set netappsvm command is incorrect:
dsmc set netappsvm management_filer_hostname storage_virtual_machine_hostname storage_virtual_machine_name

The correct order of the arguments is updated to the following text:
dsmc set netappsvm storage_virtual_machine_hostnamemanagement_filer_hostname storage_virtual_machine_name

Topic: Clients > Backup-archive client options and commands > Processing options > Client options reference > Vmtagdatamover ( http://www.ibm.com/support/knowledgecenter/SSGSG7_7.1.6/client/r_opt_vmtagdatamover.html)

The following requirements are updated:

  • In order for the Data Protection for VMware vSphere GUI to function correctly with tagging support, ensure that the following requirements are met during the installation of the GUI:

    If you use other data movers, running on virtual machines or physical machines as additional data movers, you can install them on other servers. For tagging support, all these data movers must also be configured with the vmtagdatamover=yes option. These additional data movers do not require the Data Protection for VMware vSphere GUI to be installed on the same server in order for them to work correctly as tag-based data movers.

    • At least one data mover and the Data Protection for VMware vSphere GUI must be installed on the same server. This data mover node must be configured so that the vCenter server credentials are saved. You can save the credentials by running the configuration wizard to save the data mover node password, or by using the dsmc set password command in the data mover command line.
    • Linux operating systemsFor Linux data movers, ensure that you specify the data mover installation directory and the Java™ shared library libjvm.so in the LD_LIBRARY_PATH environment variable. The path to libjvm.so is used for tagging support when you enable the vmtagdatamover option on the data mover. For instructions, see Setting up the data mover nodes in a vSphere environment.
    • Linux operating systemsOn Linux operating systems, the Data Protection for VMware vSphere GUI must be installed by using the default user name (tdpvmware).
    • Linux operating systemsOn Linux data mover nodes, the default password file (/etc/adsm/TSM.PWD) must be used.

Topic: Clients > Backup-archive client options and commands > Processing options > Client options reference > Vmtagdatamover ( http://www.ibm.com/support/knowledgecenter/SSGSG7_7.1.6/client/r_opt_vmtagdatamover.html)

The following requirement has been removed:

 
  • The VMware Platform Services Controller (PSC) and the vCenter server must be installed on the same host.
 

Per APAR IT16144


Topic: Backup-archive client options and commands > Processing options > Client options reference > Instrlogmax (http://www.ibm.com/support/knowledgecenter/SSGSG7_7.1.6/client/r_opt_instrlogmax.html)

The following paragraph is incorrect and has been removed:

"If you do not specify the instrlogmax option, the instrumentation log can grow without any limit on its size. You must manually manage the log contents to prevent the log from depleting disk resources."


 

Updates that apply to V7.1.4 and later versions



Per APAR IT22875
Clients > Backup-archive client options and commands> Processing options > Client options reference > Domain.vmfull


The following statement is added for the vmname parameter under Domain.vmfull for VMware virtual machines:

"The names are case-sensitive."


 

Updates that apply to V7.1.3 and later versions



Per APAR IT20379 (up to V7.1.6)

Clients > Backup-archive client options and commands > Processing options > Client options reference >Snapdiff

The following text is added to the first example:

"Important: For Linux clients, when you export the path of a volume on NetApp 7-mode filers, the export path must equal the volume name (for example, /vol_name), or the export path must begin with "/vol" and include one subdirectory level (for example, /vol/vol_name). Do not add a slash ("/") to the end of the export path."

Per APAR IT15064



In the topic "Configuring NetApp and Tivoli Storage Manager for snapshot difference incremental backups" (http://www.ibm.com/support/knowledgecenter/SSGSG7_7.1.3/client/t_netap_file_server_configure.html), the following text is updated:

In step 1b:
  • For 7-mode NetApp filers: Allowed Capabilities: login-http-admin,api-*
    For clustered-data ONTAP NetApp filers, the only capability that is required is ontapapi with the admin role.

The topic "Protection for clustered-data ONTAP NetApp file server volumes" (http://www.ibm.com/support/knowledgecenter/SSGSG7_7.1.3/client/c_cfg_netapp_cmode.html) is replaced by the following text:

Protecting clustered-data ONTAP NetApp file server volumes

You can create a snapshot differential incremental backup of a volume on a NetApp file server that is part of a clustered-data ONTAP configuration (c-mode file server).

Before you begin

- Complete the procedure in "Configuring NetApp and Tivoli Storage Manager for snapshot difference incremental backups."
- Ensure that the clustered-data ONTAP environment is correctly set up by the NetApp storage virtual machine administrator.

Restriction: Tivoli® Storage Manager support for snapshot differential incremental backups of clustered-data ONTAP volumes is supported only on NetApp ONTAP 8.2.1 and later versions.

About this task

In a clustered-data ONTAP environment, storage virtual machines (also referred to as data vServers) contain data volumes that can be protected by the backup-archive client.

A storage virtual machine consists of a single infinite volume or one or more flex volumes. Volumes are accessed remotely using file sharing (CIFS on Windows operating systems, NFS on AIX® and Linux operating systems).

The storage virtual machines are managed by the cluster management filer, which is the physical filer (the c-mode filer) on which the storage virtual machines reside. The backup client is installed on the remote machine that accesses the volumes.

The backup-archive client must be configured with credentials for the NetApp c-mode filers that are being accessed for backup operations.

Requirements:

The following information is required for this procedure:
- The host name or IP address of the cluster management filer.
- The host name or IP address of the storage virtual machine.
- The storage virtual machine name.
- The cluster management filer credentials (user name and password).
The cluster management filer user that is configured by the client must be assigned the ontapapi capability with the role of admin.

The ontapapi capability does not allow interactive access to the filer with methods such as telnet, ssh, or http/https. No other user capabilities are required to run snapshot differential incremental backups.

Procedure

Complete the following steps on the remote machine where the backup-archive client is installed:

1. Configure the backup-archive client with the cluster management filer credentials. Use the dsmc set password command to store the credentials of the management filer that is associated with the storage virtual machine. For example, enter the following command:

dsmc set password –type=filer management_filer_hostname
management_filer_username management_filer_password

Where:

management_filer_hostname
The host name or IP address of the cluster management filer.
management_filer_username
The user name of the cluster management filer.
management_filer_password
The password for user of the management filer.

Tip: The cluster management filer password is encrypted when it is stored by the backup-archive client.

2. Associate each storage virtual machine with the management filer with the dsmc set netappsvm command. For example, enter the following command:

dsmc set netappsvm management_filer_hostname
storage_virtual_machine_hostname storage_virtual_machine_name

Where:

management_filer_hostname
The host name or IP address of the cluster management filer.
storage_virutal_machine_hostname
The host name or IP address of the storage virtual machine that is used to mount volumes to back up.
storage_virtual_machine_name
The name of the storage virtual machine.

Note: The host name or IP address of the storage virtual machine that is used to mount volumes must be consistent with what is specified in the dsmc set commands. For example, if the volumes are mounted with a storage virtual machine IP address, the IP address (not the host name) must be used in the dsmc set commands. Otherwise, client authentication with the cluster management filer fails.

You need only to specify the dsmc set netappsvm command once for each storage virtual machine. If the storage virtual machine is moved to a different cluster management filer, you must use the command to update the associated cluster management filer host name.

3. (Windows) Map the volumes to drive letters. For example, enter the following command for each storage virtual machine:

net use y: \\storage_virtual_machine_hostname domain_name\CIFS_share_name

Where:

y:
The drive to map the volume to.
storage_virtual_machine_hostname
The host name or IP address of the storage virtual machine.
domain_name\CIFS_share_name
The CIFS share that is defined on the filer on the volume being backed up.

4. (AIX, Linux) Mount the remote storage virtual machine to a local file system. For example, enter the following command for each storage virtual machine:

mount storage_virtual_machine_hostname /tmp/fs1

Where:

storage_virtual_machine_hostname
The host name or IP address of the storage virtual machine.
/tmp/fs1
An example of a file system to mount the storage virtual machine volume to.

5. Start a full progressive incremental backup of a flex or infinite volume.

By default, HTTP access to the NetApp file server is not enabled. If you did not configure your file server to allow access by using HTTP, use the backup-archive client snapdiffhttps option to enable access to the cluster management server with the HTTPS protocol.

For example, on Windows clients, enter the following command:

dsmc incr y: -snapdiff -snapdiffhttps

For example, on AIX or Linux clients, enter the following command:

dsmc incr /tmp/fs1 -snapdiff -snapdiffhttps

Tip: You need only to run the full progressive incremental backup once. After this backup is successfully completed, run differential backups in future backup operations.

6. Start a snapshot differential backup of the flex or infinite volume.
For example, on Windows clients, enter the following command:

dsmc incr y: -snapdiff -snapdiffhttps

For example, on AIX or Linux clients, enter the following command:

dsmc incr /tmp/fs1 -snapdiff -snapdiffhttps

Example

A backup-archive client user wants to complete a snapshot differential incremental backup of the volumes on a c-mode file server. The user is using a Windows backup-archive client to complete the backup and the volumes are mounted as CIFS shares. The c-mode filer configuration is as follows:

ONTAP 8.31 management filer

Hostname: netapp1mgmt.example.com
User: netapp1mgmt_user
Password: pass4netapp1mgmt
CIFS Domain Controller: WINDC
Domain User: domainuser

Flex volume storage virtual machine

Hostname: netapp1-v1.example.com
Storage virtual machine name: netapp1-client1
CIFS share: demovol
Volume name: demovol

Infinite volume storage virtual machine

Hostname: netapp1-v4.example.com
Storage virtual machine name: netapp1-infiniteVolume1
CIFS Share: InfiniteVol

The user completes the following steps on the backup-archive client:

1. Configure the client with the management filer credentials by issuing the following command:

dsmc set password –type=filer netapp1mgmt.example.com netapp1mgmt_user pass4netapp1mgmt

2. Define storage virtual machine associations for each storage virtual machine with the following commands:

dsmc set netappsvm netapp1mgmt.example.com netapp1-v1.example.com netapp1-client1

dsmc set netappsvm netapp1mgmt.example.com netapp1-v4.example.com netapp1-infiniteVolume1

3. Map remote volumes to drive letters for each storage virtual machine:

net use y: \\netapp1-v1.example.com\demovol WINDC\domainuser

net use z: \\netapp1-v4.example.com\InfiniteVol WINDC\domainuser

4. Run a full progressive incremental backup of the flex volume and infinite volume:

dsmc incr y: -snapdiff -snapdiffhttps

dsmc incr z: -snapdiff -snapdiffhttps

You need only to run the full progressive incremental backup once. After this backup is successfully completed, run differential backups in future backup operations.

5. Run a snapshot differential backup of the flex volume and infinite volume:

dsmc incr y: -snapdiff -snapdiffhttps

dsmc incr z: -snapdiff -snapdiffhttps


 

Updates that apply to V7.1.2

Per APAR IT03221



The backup-archive client has two new options that will allow all pRDM or vRDM disks with missing LUNs to be converted into thin-provisioned VMFS disks: vmnoprdmdisks and vmnovrdmdisks.


Linux operating systemsWindows operating systems
Vmnoprdmdisks

This option enables Tivoli Storage Manager to restore configuration information for the pRDM volumes that are associated with a VMware virtual machine, even if the LUNs that were associated with the volumes cannot be found. Because pRDM volumes are not included in virtual machine snapshot, only the configuration information can be restored, and not the data that was on the volumes.

This option does not apply to backups of Microsoft Hyper-V virtual machines.

Supported Clients

This option is valid for Windows and Linux clients that are installed on a vStorage backup server.

Options File

Windows operating systemsPlace this option in the client options file (dsm.opt), or specify it as a command-line parameter on the restore vm command.

Linux operating systemsPlace this option in the client options file (dsm.opt), in the client system options file (dsm.sys), or specify it as a command-line parameter on the restore vm command.

Syntax


Read syntax diagramSkip visual syntax diagram
                  .-NO--.   
>>-VMNOPRDmdisks--+-----+--------------------------------------><
                  '-YES-'   



Parameters

YES
Specify this value if you must restore a virtual machine that you backed up with -vmprocesswithprdm=yes, and the original LUNs that were mapped by the raw device mappings file cannot be located. This setting causes the client to skip attempts to locate the missing LUNs used by the pRDM volumes, and restore the configuration information (disk labels) that were associated with them. The pRDM volumes are restored as thin-provisioned VMFS VMDKs. You can then use the vSphere client to create the necessary pRDM mappings.

NO
Setting -vmnoprdmdisk=no causes restore operations for virtual machines that were backed up with -processvmwithprdm=yes to fail if the original LUNs that were mapped to by the raw device mappings file cannot be located. This value is the default value.

Examples

Option file:
VMNOPRDMDISKS YES

Command line:
dsmc restore vm vm123 -vmnoprdmdisks=yes


Linux operating systemsWindows operating systems
Vmnovrdmdisks

This option enables Tivoli Storage Manager to restore configuration information and data for vRDM volumes that are associated with a VMware virtual machine, even if the LUNs that were associated with the volumes cannot be found.

This option does not apply to backups of Microsoft Hyper-V virtual machines.

Supported Clients

This option is valid for Windows and Linux clients that are installed on a vStorage backup server.

Options File

Windows operating systemsPlace this option in the client options file (dsm.opt), or specify it as a command-line parameter on the restore vm command.

Linux operating systemsPlace this option in the client options file (dsm.opt), in the client system options file (dsm.sys), or specify it as a command-line parameter on the restore vm command.

Syntax


Read syntax diagramSkip visual syntax diagram
                  .-NO--.   
>>-VMNOVRDmdisks--+-----+--------------------------------------><
                  '-YES-'   



Parameters

YES
Specify this value if you must restore a virtual machine that you backed up, and the original LUNs that were mapped by the raw device mappings file cannot be located. This setting causes the client to skip attempts to locate the missing LUNs used by the vRDM volumes, and restore the configuration information (disk labels) and the data that was backed up. The vRDM volumes are restored as thin-provisioned VMFS VMDKs.

NO
Setting -vmnovrdmdisk=no causes restore operations for virtual machines that had vRDM volume to fail, if the original LUNs that were mapped to by the raw device mappings file cannot be located. This value is the default value.

Examples

Option file:
VMNOVRDMDISKS YES

Command line:
dsmc restore vm vm123 -vmnovrdmdisks=yes


 

Updates that apply to V7.1.1 and later versions


Per APAR IT16782
Topic: Backup-archive client options and commands > Processing options > Client options reference > Domain > Automounted file systems

The following paragraph is added:

Important: On some Linux distributions, automounted file system mount points or maps of file system type (AutoFS) might not be listed in the current mount table. As a result, the automounted file systems, which are unmounted during backup or archive processing, might be incorrectly processed and stored as part of a wrong domain (for example, as part of domain all-local, all-nfs, or all-lofs, depending on the actual file system type). Therefore, in such Linux distribution environments, you must specify the appropriate automount option setting to correctly process your domain option setting at all points in time.


 

Updates that apply to all modifications of V7.1



Per APAR IT23865
Clients > Backup-archive client options and commands > Processing options > Client options reference > Queryschedperiod

The following text has been added:

Tip: If the period set by the queryschedperiod option is much smaller than the randomization window of a schedule that is set by the server administrator, the start of the schedule can be delayed. To avoid such a delay, adjust the following values:
  • The client action duration (with the SET CLIENTACTDURATION server command)
  • The randomization of scheduled start times (with the SET RANDOMIZE server command)
  • The value of the queryschedperiod option

Given the settings for the client action duration and the randomization window of a schedule, the following examples show how to calculate the query schedule period.

Example 1:


  Client Action Duration: 1 Days
 Schedule Randomization Percentage: 25%
 Query Schedule Period: 6 hours

 Client Action Duration of 1 day = 24 hours
   24 hours x .25 = 6 hours
 Use a query schedule period of 6 hours or higher.


Example 2:
  Client Action Duration: 3 Days
 Schedule Randomization Percentage: 10%
 Query Schedule Period: 8 hours

 Client Action Duration of 3 days = 72 hours
   72 x .10 = 7.2
 Use a query schedule period of 8 hours or higher.



Per APAR IT23581
Clients > Back up and restore data with backup-archive clients > Backing up your data > Deleting backup data

The "Results" section has been updated as follows:

Results
Note:
  • If you specify Delete Active Objects or Delete Inactive Objects, only the files are considered for removal.
  • If you specify Delete Active Objects or Delete Inactive Objects and select a directory that contains no files for removal, the following message is displayed during the delete backup operation:

    ANS5030E No objects on server match query.

    The last parent inactive directory is removed based on retention policy settings on the server.
  • A directory is deleted only if you select Delete All Objects.
  • To delete file spaces, click Utilities > Delete Filespaces from the main window.
  • To delete backup copies using the command-line client, use the delete backup command.

Per APAR IT23279
Clients > Back up and restore data with backup-archive clients > Backing up your data > Display backup processing status

In Table 1. Client command line informational messages, the descriptions of "Data transfer time", "Network data transfer rate", "Aggregate data transfer rate", and "Total number of bytes transferred" are incorrect. These descriptions have been updated as follows:
 
Data transfer time: The sum of the times that each backup, archive, restore, or retrieve session takes to send data across the network. This number does not include the time for the client to read the data from disk before the date is sent, nor the time to wait for server transactions to complete.

This number can be greater than the elapsed processing time if the operation uses multiple concurrent sessions to move data, such as multi-session backup and restore operations.

This number includes the time that it takes to send data more than once due to retries, such as when a file changes during a backup operation.
Network data transfer rate: The average rate at which the network transfers data between the client and the server. This statistic is calculated by dividing the total number of bytes transferred by the time to transfer the data over the network. This statistic does not include the time for the client to read the data from disk before the data is sent, nor the time to wait for server transactions to complete.
Aggregate data transfer rate: The total number of bytes transferred during a backup, archive, restore, or retrieve operation, divided by the total elapsed time of the operation.
Total number of bytes transferred: The total number of bytes transferred during the backup, archive, restore, or retrieve operation. This value includes data that is sent more than once due to retries, such as when a file changes during a backup operation.



Per APAR IT22672
Clients > Backup-archive client options and commands> Backup Image

The following statement is added:

"The image backup operation is not supported on any partition that resides on a multipath device."

In addition, the explanation for message ANS1068E has been modified as follows:

"The selected path is not a valid object for image operations. The selected path is either a remote
device or a partition that resides on a multipath device."


Per APAR IT22157
Clients > Backup-archive client options and commands > Processing options > Client options reference > Snapdiff

The following statement is incorrect:

(Windows) The snapdiff (snapshot difference) option is for backing up NAS/N-Series file server volumes that are NFS or CIFS attached.

It is replaced by the following two statements:

(Windows) The snapdiff option is for backing up NAS/N-Series file server volumes that are CIFS attached.
(Linux) The snapdiff option is for backing up NAS/N-Series file server volumes that are NFS attached.

Per APAR IT19974


Clients > Back up and restore data with backup-archive clients > Backing up your data > Performing an incremental, selective, or incremental-by-date backup
(UNIX and Linux) > Full and partial incremental backup

The last bulleted item in the following passage is inaccurate. The statement only applies to GPFS file systems and backups are triggered by updates to the ctime attribute.
...
Back up directories
A directory is backed up in any of the following circumstances:
-The directory was not previously backed up
-The directory permissions changed since the last backup
-The directory Access Control List changed since the last backup
-The directory Extended Attributes changed since the last backup
-The directory modification time stamp changed since the last backup
...

The last bulleted item in this passage is changed to the following text to indicate that directory backups are triggered by updates to the ctime attribute for GPFS file systems:
...
Back up directories
A directory is backed up in any of the following circumstances:
-The directory was not previously backed up.
-The directory permissions changed since the last backup.
-The directory Access Control List changed since the last backup.
-The directory Extended Attributes changed since the last backup.
-The change time (ctime) attribute is updated since the last backup (for GPFS file systems only). For more details, see the updatectime option.
...

Per APAR IT19424


Clients > Back up and restore data with backup-archive clients > Backing up your data > Performing an incremental, selective, or incremental-by-date backup
(UNIX and Linux) > Full and partial incremental backup

The following text is updated to indicate that the ctime attribute statement applies only to objects in GPFS file systems in the following client topic:

...
If only the following attributes change, the attributes are updated on the IBM Spectrum Protect server, but the file is not backed up:
-File owner
-File permissions
-Inode
-Group ID
-Change time (ctime) attribute (for objects in GPFS file systems only and if the updatectime option is set to yes). For more details, see the updatectime option.
...

Per APAR IT18953


Clients > Backup-archive client options and commands > Processing options > Client options reference > Updatectime

The following text is updated:

Use the updatectime option to check the change time (ctime) attribute during an incremental backup operation.

Parameters

no
The backup-archive client does not check the change time (ctime attribute) during a backup operation. This value is the default.
yes
The backup-archive client checks the change time (ctime attribute) during a backup operation. If the ctime attribute changed since the last backup operation, the ctime attribute is updated on the IBM Spectrum Protect™ server. The object is not backed up unless it has either ACLs or extended attributes. The client checks files and directories."

The following text is removed: "Use this option with the incremental, selective, or archive commands."

Per APAR IT17292



Clients > Configure the IBM Spectrum Protect client > Configure your system for journal-based backup > Journal daemon configuration

The following text is added:

"To stop the journal daemon on AIX, issue the kill nnnn command, where nnnn is the process ID of tsmjbbd. Before the journal daemon process (tsmjbbd) shuts down, it notifies the filepath kernel extension to stop buffering file changes.

Important: Do not use the kill -9 nnnn command, because the kill -9 command immediately ends the process without notifying filepath to stop buffering file changes."
 

Per APAR IT17557



Clients > Back up and restore data with backup-archive clients > Restoring your data > Restore AIX encrypted files

The following paragraph is updated:

"After restoring a file that was backed up in raw format, you might find that the file cannot be decrypted. The encryption key originally used for the file might no longer be available in the keystore of the user. In this case, you must restore the keystore used at the time of backup."

Per APAR IT17094



Clients > Back up and restore data with backup-archive clients > Backing up your data > Image backup

The following text is added:

"For Linux clients: Image backup of DASD devices with raw-track access mode on Linux on z Systems™ is not supported."

Clients > Backup-archive client options and commands > Using commands > Backup Image

The following text is added:

"For the Linux client, image backup of DASD devices with raw-track access mode on Linux on z Systems™ is not supported."

Installation


Per APAR IC98225

Installing the Tivoli Storage Manager Solaris client
The installation steps have been revised to describe differences in the pkgadd command behavior on Solaris 10 and 11. Steps 3, 4, 5, and 6 contain notes to indicate the following behavior:

Note: On Solaris 10, this command installs the 64-bit <component name> in the global zone and in all running non-global zones. If you want to install them in the global zone only, use the -G parameter of the pkgadd command. On Solaris 11, the client components are installed only in the zone where this command is run.


Per Defect 114656

Getting started > Ending a session
  • The following text is added:
    "Do not press Ctrl-C or use the UNIX kill -15 command because it can lead to unexpected results."


Protecting files and workstations

Per APAR IT02775

Steps 2a and 2b have been revised to more clearly describe the requirements for the NetApp HTTP server:

Configuring NetApp and Tivoli Storage Manager for snapshot difference incremental backups

You must configure the NetApp file server connection information to run the snapshot difference incremental backup command on the Tivoli® Storage Manager client. You must also use the set password command to specify the file server host name, and the password and user name that is used to access the file server.


1. Establish a console session on the NetApp filer and define a new user on the file server by using the following steps:
a. Add the user ID to a group that permits users to log in to the file server with http and running API commands.
b. From the file server, enter the following command to list the user ID to verify the settings and verify that the output is similar:
useradmin user list snapdiff_user
Name: snapdiff_user
Info:
Rid: 131077
Groups: snapdiff_group
Full Name:
Allowed Capabilities: login-http-admin,api-*

c. If the security.passwd.firstlogin.enable option for the user ID on the NetApp server is set to on, ensure that all groups have the login-telnet and cli–passwd* capabilities.
Tip: When security.passwd.firstlogin.enable option is enabled, the user ID is set to expired when created. The user cannot run any commands, including snapshot difference incremental, until their password is changed. Users in groups that do not have these capabilities cannot log in to the storage system. Refer to the NetApp documentation for details on defining a user ID and a password on the NetApp file server.
2. Configure the NetApp Data ONTAP built-in HTTP server to allow remote administrative sessions to the NetApp filer.
a. Optional: If you plan to use a plain HTTP connection for snapshot differential backups, turn on the httpd.admin.enable option on the NetApp filer to allow HTTP administrative access.
b. If you plan to use a secure HTTPS connection for snapshot differential backups (by specifying the -snapdiffhttps option), turn on the httpd.admin.ssl.enable option on the NetApp filer.
c. From the Tivoli Storage Manager client node, test the connection between the Tivoli Storage Manager client computer and the NetApp ONTAP server to ensure that firewalls or other NetApp configuration options do not prevent you from connecting to the NetApp server.
Tip: See the NetApp ONTAP documentation for instructions on how to test the connection.
3. Windows operating systemsExport the NetApp volumes and consider the following settings:
Tip: See the NetApp documentation for details on exporting the NetApp volumes for use with Windows.
  • Map the NetApp volumes by using CIFS.
  • Ensure the NetApp volumes have the NTFS security setting.
4. AIX operating systemsLinux operating systemsExport the NetApp volumes and consider the following settings:
AIX operating systemsLinux operating systemsTip: See the NetApp documentation for details on exporting the NetApp volumes for use with AIX®, or Linux hosts.
  • Map the NetApp volumes by using an NFS mount.
  • Ensure the NetApp volumes have the UNIX security setting
5. Set the user ID, and password on Tivoli Storage Manager client for the user ID that you created in step 1 using the following steps:
a. AIX operating systemsLinux operating systemsLog in as the root user ID.
b. Windows operating systemsLog on as the user with read/write access to the CIFS share.
c. From the Tivoli Storage Manager client command line, enter the following command:
dsmc set password –type=filer my_file_server snapdiff_user newPassword
Substitute the following values:
my_file_server
This value is the fully qualified host name of your NetApp file server.
snapdiff_user
This value is the user ID that you created in step 1.
newPassword
This value is the password for the user ID that you created in step 1.

File system and ACL support

This topic has been updated to include the following information.
  • The table row for Linux_x86_64 clients was missing information. The row content has been updated to read as follows:
Platform File system ACL support
Linux operating systems
Linux x86_64

Btrfs
XFS
EXT2
EXT3
EXT4
ReiserFS
GPFS
JFS
VxFS
NSS

yes
yes
yes
yes
yes
yes
yes
no
no
yes
Platform File system ACL support
Linux operating systems
Linux on Power Systems™ Servers

XFS
EXT2
EXT3
EXT4
ReiserFS
JFS
GPFS

yes
yes
yes
yes
yes
no
yes
  • The table row for Linux on Power Systems Servers contain an error in the ACL support column for GPFS file systems. In the table, the ACL support column entry for GPFS should state"yes".
  • The paragraphs that describe support for GPFS and mixed node cluster environments has been updated to read as follows:

AIX operating systemsLinux operating systemsImportant: If you are running GPFS for AIX or GPFS for Linux x86_64 in a multinode cluster, and all nodes share a mounted GPFS file system, Tivoli Storage Manager processes this file system as a local file system. Tivoli Storage Manager backs up the file system on each node during an incremental backup. To avoid this, you can do one of the following things:
  • Explicitly configure the domain statement in the client user-options file (dsm.opt) to list the file systems you want that node to back up.
  • Set the exclude.fs option in the dsm.sys file to exclude the GPFS file system from backup services.

Restriction: If the GPFS cluster contains mixed nodes (some AIX®, some Linux, and some Windows), you must use the AIX or Linux clients to protect data in the cluster.

Per APAR IT03613

Backup network file systems

The topic that was titled "Back up NFS file systems" has been rewritten and given a new title. The following text replaces the old topic:

AIX operating systemsHP-UX operating systemsLinux operating systemsMac OS X operating systemsOracle Solaris operating systems
Backup network file systems

The Tivoli® Storage Manager backup-archive client can be configured to protect files that are accessed with either Network File System (NFS) or Common Internet File System (CIFS) protocols.

Backup performance is better when you install the backup-archive client where the file system physically resides, but sometimes it is necessary to access file systems using NFS or CIFS to back up or recover data on remote shares. The Tivoli Storage Manager UNIX and Linux backup-archive client can back up, archive, restore, and retrieve file data on an NFS or CIFS-mounted share. For NFS, this includes all versions of the NFS protocol, including NFS version 2, NFS version 3, and NFS version 4.

Tivoli Storage Manager can back up and restore access control lists when configured to use NFS version 4.

The following restrictions apply when the backup-archive client protects data on network file system volumes:

  • Tivoli Storage Manager backup-archive clients cannot perform image backups of network file system volumes.
  • Tivoli Storage Manager AIX clients cannot perform snapshot-based file backups or archive files on network file system volumes.
  • Tivoli Storage Manager backup-archive clients cannot perform journal-based backups of network file system volumes.
  • Tivoli Storage Manager backup-archive clients might not be able to back up NetApp volume snapshots if they are accessed using NFS protocol. If the NetApp filer provides different device identifiers for its volume snapshots, these snapshots might be excluded from backups; the behavior depends on the OS version, the NetApp Filer version, and its settings.


Per APAR IT13174

Topic: Backing up your data -> Backing up VMware virtual machines

The following Note is added before the table that lists the backup and restore operations for VMware virtual machines that the backup-archive client can implement on Windows platforms:
  • Note: You can complete VMware backup and restore operations with the backup-archive client only on 64-bit Windows operating systems. Backup-archive clients on Windows 32-bit operating systems cannot complete VMware backup and restore operations.


Per APAR IC92811_1

Solaris global zone and non-global zones backup

The Solaris global zone and non-global zone backup topic has been rewritten to improve its clarity. The new text reads as follows:

Oracle Solaris operating systems
Solaris global zone and non-global zones backups

For Solaris zones, perform incremental and selective backups of file systems within the zone where these file systems were created.

Treat each non-global zone as a separate system that has its own Tivoli® Storage Manager node name and run backups from within each of the zones.

If you run incremental or selective backups of non-global zones from the global zone, the global-zone administrator must decide which files in the non-global zone are included or excluded in backups. For example, device, system and kernel files of the non-global zones are not automatically excluded from backups, but they must not be backed up. Restoring such files can make a non-global zone unusable.




Per APAR IT14125

The following topic is updated with additional information:

Back up and restore data with backup-archive clients > Backing up your data > Display backup processing status

Two additional messages are documented:

Total number of objects grew:
The total number of files that grew larger as a result of compression.

Total number of retries:
The total number of retries during a backup operation. Depending on the settings for the serialization attribute and the changingretries option, a file that is opened by another process might not be backed up on the first backup try. The backup-archive client might try to back up a file several times during a backup operation. This message indicates the total retries for all files that are included in the backup operation.


Per APAR IT01977

The following topics have been revised to clarify the requirements for backing up and restoring Oracle Solaris Zettabyte file systems:

>Backing up Solaris Zettabyte file systems (changed title)
>Restoring Solaris Zettabyte (ZFS) file systems

Backing up Solaris Zettabyte file systems

On Solaris SPARC and Solaris x86 systems, you can backup Zettabyte file systems (ZFS), by using ZFS snapshots. The advantage of this approach, over an ordinary incremental or selective backup, is that the files and folders in a snapshot are always in a read-only state, so they cannot be changed during a backup.

About this task

You create a ZFS snapshot by using Oracle Solaris ZFS commands. For example:


zfs snapshot tank/myZFS@mySnapshot

In this example, the ZFS pool name is called tank and the ZFS file system name is myZFS. Files that belong to this ZFS snapshot are in the subdirectory named tank/myZFS/.zfs/snapshot/mySnapshot/.

Procedure

Use either of these two methods to backup a ZFS snapshot.

· Backup each file of the snapshot by using the Tivoli® Storage Manager snapshotroot option. For example:

dsmc inc -snapshotroot=/tank/myZFS/.zfs/snapshot/mySnapshot /tank/myZFS



This option allows the administrator to replace the current snapshot path with the ZFS file system path, so that the files and folders are backed up under the original file system.

· Backup the complete snapshot by using Oracle Solaris ZFS commands. For example:

zfs send tank/myZFS@mySnapshot > /tmpdir/mySnapshotFile



The advantage of backing up the complete snapshot is that the full file system can be restored, in a disaster recovery scenario.



Restoring Solaris Zettabyte (ZFS) file systems

Zettabyte File Systems (ZFS) use storage pools to manage physical storage.

How you restore a ZFS file system depends on how it was backed up.

  • If you backed up all files and folders as separate objects, you can restore them by performing a file-level restore. For example:

dsmc restore /tank/myZFS/ -subdir=yes -replace=all

Do not perform a file-level restore operation in a disaster recovery scenario. Even though you successfully restore all system files and folders from a Tivoli® Storage Manager client-created backup, the restored system might be unstable or fail.

  • If you backed up an entire ZFS snapshot as a single file, you need to restore the snapshot file from the server into a temporary location. For example:

dsmc restore /tmpdir/mySnapshotfile

You can then restore the file system from the snapshot file by using the Oracle Solaris ZFS commands. For example:

zfs receive tank/myZFS@mySnapshot < /tmpdir/mySnapshotFile

The advantage of restoring ZFS from a snapshot file is that the full file system can be restored, in a disaster recovery scenario.

Refer to the Oracle Solaris ZFS Administration Guide for detailed information about restoring data on ZFS file systems. If you are restoring a ZFS root pool, see the topics that describe how to re-create your root pool and recover root pool snapshots.




Backup and restore data with backup-archive clients

Per APAR IC02964

Topic: Back up NAS file systems using Network Data Management Protocol

This topic applies only to AIX, Solaris, and Windows clients. The topic, and subtopics, have been revised to remove the icons for HP-UX, Linux, and Mac, because those clients cannot be used for backing up or restoring data from NAS file servers, using NDMP. Additionally, the introductory text has been clarified to indicate that the tape drives and libraries can be on either the file server or the Tivoli Storage Manager server. The new text reads as follows:
  • Tivoli® Storage Manager Windows, AIX®, and Solaris backup-archive clients can use Network Data Management Protocol (NDMP) to efficiently back up and restore network attached storage (NAS) file system images. The file system images can be backed up to, or be restored from, automated tape drives or libraries that are locally attached to Network Appliance or EMC Celerra NAS file servers, or to or from tape drives or libraries that are locally attached to a Tivoli Storage Manager server.NDMP backups.


Backup and restore data with backup-archive clients

Per APAR IC04910

Topic: Restoring your data

The following paragraph is intended for Windows backup-archive clients only. This note does not apply to UNIX/Linux backup-archive clients:

Note: When restoring a directory, its modification date and time is set to the date and time of the restore, not to the date and time the directory had when it was backed up. This is because Tivoli Storage Manager restores the directories first, then adds the files to the directories.




Storage management policies

Per Defect IT03448

Display information about management classes and copy groups > Copy mode attribute
The publication does not mention directory objects in the explanation of the absolute and modified copy-group modes. The absolute and modified copy-group modes apply to directory objects and to file objects.




Client options

Per defect IT03448

Topic: Client options reference> ABSOLUTE

The publication does not mention directory objects in the explanation of the absolute option. The absolute option applies to directory objects and file objects.


Per APAR IT11049

Topic: Client options reference> ASNODENAME

The explanation is changed as follows:
  • Asnodename

    Use the asnodename option to allow agent nodes to back up or restore data on behalf of another node (the target node). This enables concurrent operations from multiple nodes to store data to the same target node and file space in parallel.

    Your client node must be granted access to the target node by the Tivoli® Storage Manager server administrative client grant proxynode command, and you must be a root user to use the asnodename option.

    When the Tivoli Storage Manager administrator grants a node proxy authority, and you use the asnodename option to become that node, you can query and restore all files as if you had root authority.

    An agent node is a client node that has been granted authority to perform client operations on behalf of a target node.

    A target node is a client node that grants authority to one or more agent nodes to perform client operations on its behalf.

    A proxy operation uses the target node's settings (such as maxnummp and deduplication) and schedules that are defined on the Tivoli Storage Manager server. The Tivoli Storage Manager server node settings and schedules for the agent node are ignored.

    ...



Per APAR IT11049

Topic: Client options reference> ASNODENAME > Session settings and schedules for a proxy operation

The complete, updated topic is printed:
  • Session settings and schedules for a proxy operation

    A proxy operation occurs when an agent node uses the asnodename target_node_name option to complete operations on behalf of the specified target node.

    A proxy operation uses the target node's settings (such as maxnummpcloptset, and deduplication) and schedules that are defined on the Tivoli® Storage Manager server. The Tivoli Storage Managerserver node settings and schedules for the agent node are ignored.

    The following considerations apply to proxy operations.

    • All operations use the target node's policy domain settings and constructs, even if the agent node belongs to a different domain. The agent node's policy domain settings and constructs are ignored.
    • The agent node authenticates to the Tivoli Storage Manager server by using the agent node's password.
    • In order to run proxy operations, the agent node and target node must not be locked on the Tivoli Storage Manager server.
    • Proxy node relationships are not transitive. If a target node is itself defined as a proxy node for some other node, the agent node cannot be used to run operations on that other node unless the agent is also defined as a proxy node for that other node.

    For example, assume the following proxy definitions among nodes TAURUS, SCORPIO, and GEMINI:
    • TAURUS is a proxy node for SCORPIO.
    • TAURUS is not a proxy node for GEMINI.
    • SCORPIO is a proxy node for GEMINI.
    The proxy definitions yield the following results:
    • TAURUS can run operations on behalf of SCORPIO.
    • SCORPIO can run operations on behalf of GEMINI.
    • TAURUS cannot run operations on behalf of GEMINI.


Per APAR IT12617

Topic: Backup-archive client options and commands > Processing options > Client options reference > COMPRESSALWAYS

The COMPRESSALWAYS topic adds the following caveat:

The compressalways option is ignored when client-side deduplication is enabled.


Per APAR IT04057

Topic: Client options reference>DOMAIN

The following clarification has been added to the description of the DOMAIN option:

If a domain statement excludes one or more objects and no domain statement includes any objects, the result is an empty domain (nothing is backed up). You must specify the objects to include in the domain if any domain statements exclude objects.

Example 1: This example uses one domain statement to back up all local file systems except for /fs1:

domain all-local -/fs1

Example 2: This example uses multiple domain statements to back up all local file systems except for /fs1:

domain all-local


domain -/fs1

Example 3: This example excludes /fs1 from backup. If no other domain statement is used, the result is an empty domain. Nothing is backed up.

domain -/fs1

Example 1: This example uses one domain statement to back up all local file systems except for the system state:

domain all-local -systemstate

Example 2: This example uses multiple domain statements to back up all local file systems except for the system state:

domain all-local


domain -systemstate

Example 3: This example excludes the system state from backup. If no other domain statement is used, the result is an empty domain. Nothing is backed up.

domain -systemstate

If you start the incremental command with a file specification, Tivoli Storage Manager ignores any domain definitions and backs up only the file specification.



Per APAR IC94432

Topic: Client options reference> DOMAIN.VMFILE
Topic: Client options reference> DOMAIN.VMFULL

The value that you specify for the vmhost parameters on these options must match the host name, as it is displayed in the vCenter server, in the Hosts and Clusters view.

The new descriptions read as follows:

Description for DOMAIN.VMFULL

vmhost=hostname
Process all virtual machines that are defined to the Virtual Center or to the ESX server that is specified on the vmchost option. The host name that you specify must match the fully qualified host name or IP address, as it is displayed in the vCenter server in the "Host and Clusters" view.

All virtual machines that are added to this host are automatically included in backup and restore processing.

This parameter can include multiple ESX servers that are separated by commas. When the Virtual Center contains multiple ESX servers, this option does not determine the ESX server from which a snapshot is taken. The ESX server from which a snapshot is taken is determined by the VMware VirtualCenter web service.

When you connect directly to an ESXi or ESX host, the vmchost option applies only if the vmhost is the server that you connect to. If it is not, a warning level message is issued to the console and is recorded in the client dsmerror.log; it is also recorded as a server event message.

If the vmenabletemplatebackups option is set to yes, and VM templates are part of the domain, they are included in the backup.

Restriction: VMware templates for virtual machines cannot be backed up when they are in an ESX or ESXi host because ESX and ESXi hosts do not support templates.


Description for DOMAIN.VMFILE

vmhost=hostname
Process all virtual machines that are defined to the Virtual Center or to the ESX server that is specified on the vmchost option. The host name that you specify must match the fully qualified host name or IP address, as it is displayed in the vCenter server in the "Host and Clusters" view.

All newly added to this host are automatically included in backup and restore processing. To be included, the virtual machines must also be running on the ESX server that is specified by the host name.

This parameter can include multiple ESX servers separated by commas. When the Virtual Center contains multiple ESX servers, this option does not determine the ESX server from which a snapshot is taken. The ESX server from which a snapshot is taken is determined by the VMware VirtualCenter web service.

When you connect directly to an ESXi or ESX host, the vmhost keyword applies only if the vmhost is the ESXi you connect to. If it is not, a warning level message is issued to the console and is recorded in the client dsmerror.log file; it is also recorded as a server event message.


Per APAR IT02572

Topic: Client options reference> Include options

In the following examples, the include.image statement has been revised to remove the wildcard characters in the file system specification. Wildcard characters are not allowed on include image statements.

Examples
Options file:
AIX operating systemsHP-UX operating systemsLinux operating systemsOracle Solaris operating systems
include /home/proj/text/devel.*
include /home/proj/text/* textfiles
include * managall
include /WAS_ND_NDNODE mgmtclass
include /WAS_APPNODE mgmtclass
include.image /home
/*/*
include.archive /home/proj/text/
 * myarchiveclass
include.backup /home/proj/text/
 * mybackupclass
include.compression /home/proj/text/
 devel.*
include.encrypt /home/proj/gordon/*
include.fs.nas netappsj/vol/vol0
 homemgmtclass



Per APAR IC97752

Topic: Client options reference> Passwordaccess

New text has been added to the bulleted list of text at the top of this topic that describes the why users might be prompted to enter their password when establishing a session with the server, even if the server is not configured to require authentication. The new text is in the last bullet item in the following text:

If a password is required, you can choose one of the following methods:
  • Set the password for your client node yourself and have Tivoli® Storage Manager prompt for it each time you request services.
  • Let Tivoli Storage Manager automatically generate a new password for your client node each time it expires, encrypt and store the password in a file, and retrieve the password from that file when you request services. You are not prompted for the password.
  • If the server is not configured to require a password to log on to it, you can still be prompted to enter your node password when the backup-archive client establishes a connection with the server. This behavior occurs if this option, passwordaccess, is allowed to default or if you set it to passwordaccess prompt. The password that you supply in response to the prompt is used only to encrypt your login information; it is not used to log onto the server. In this configuration, you can avoid entering a password by setting this option to passwordaccess generate. Setting passwordaccess generate causes the client to create, store, and submit the password for you. When passwordaccess generate is set, the password option is ignored.


Per APAR IT02889

Topic: Client options reference> Resourceutilization

The following caveat is added to the list of potentially undesirable aspects of running multiple sessions:
  • It is possible that files are restored instead of hard links.
Restoring files instead of hard links can occur when the following criteria are all true:
  • You restore an entire file system.
  • During the restore operation, the value of the resourceutilization option is greater than 1.
  • The file system contained hard links when the file system was backed up.
The chance of restoring linked files instead of hard links increases as the number of sessions increases. When you restore a file system that contained hard links when the file system was backed up, set resourceutilization=1 to ensure that hard links are restored.


Client options reference > Vmmaxparallel

Per internal defect 115223

The vmmaxparallel option is not valid in the dsm.opt options file.

 




Commands

Per APAR IT14149



Using commands>Restore VM

The description of the HOST parameter for VMware virtual machines has been revised to read as follows:

Specifies the domain name of the ESX host server to restore to as it is defined
in the vSphere vCenter.

This parameter is case-sensitive and must be the same value as the host name
that is shown in the VMware vSphere Web Client. To confirm the host name
in the vSphere Web client, select a host and click Manage > Networking >
TCP/IP configuration > DNS.

This parameter is not valid for restoring VMware virtual machines that were
backed up using VCB.
 

Per APAR IT03448



Backup-archive client options and commands > Using commands > Incremental
The publication does not mention directory objects in the explanation of the absolute and modified copy-group modes. The absolute and modified copy-group modes apply to directory objects and to file objects.

Per APAR IC98148

IBM Tivoli Storage Manager > Clients > Backup-archive client options and commands > Using commands> Restore VM



The descriptions of the :vmdk and :-vmdk parameters have been updated to include notes that describe that hard disk names must be entered in ASCII, as they are displayed in the output of a Backup VM command that uses the -preview option. The new text reads as follows:

:vmdk=disk label
This parameter is not valid for restoring VMware virtual machines that were backed up using VCB.

This option is used to specify the disk label of the virtual disks to include in the restore operation. You specify this option only if you want to selectively restore data from specific disks.

Note: On the Restore VM command, the label names of the vmdk files that you want to include (:vmdk= parameter) in a Restore VM operation must be specified as the English-language label name, as it is displayed in the output of the Backup VM vmname -preview command. Examples of the English vmdk labels are "Hard Disk 1", "Hard Disk 2", and so on.

:-vmdk=disk label
This parameter is not valid for restoring VMware virtual machines that were backed up using VCB.

This option is used to specify the disk label of one or more virtual disks to exclude from the restore operation.

Note: On the Restore VM command, the label names of the vmdk files that you want to exclude (:-vmdk= parameter) from a Restore VM operation must be specified as the English-language label name, as it is displayed in the output of the Backup VM vmname -preview command. Examples of the English vmdk labels are "Hard Disk 1", "Hard Disk 2", and so on.

[{"Product":{"code":"SSGSG7","label":"Tivoli Storage Manager"},"Business Unit":{"code":"BU054","label":"Systems w\/TPS"},"Component":"Client","Platform":[{"code":"PF002","label":"AIX"},{"code":"PF010","label":"HP-UX"},{"code":"PF016","label":"Linux"},{"code":"PF022","label":"OS X"},{"code":"PF027","label":"Solaris"}],"Version":"7.1;7.1.1;7.1.2;7.1.3;7.1.4;7.1.6;7.1.8","Edition":"All Editions","Line of Business":{"code":"LOB26","label":"Storage"}}]

Document Information

Modified date:
10 July 2020

UID

swg27039351