Release Notes. EMC NetWorker Module for Databases and Applications. Release Notes. Version 8.2 and Service Packs REV 12.

(1)

EMC

®

NetWorker

®

Module for Databases and

Applications

Version 8.2 and Service Packs

Release Notes

302-000-386

REV 12

April 25, 2016

These release notes contain supplemental information about EMC NetWorker Module for Databases and Applications (NMDA) release 8.2 and Service Packs.

NOTICE

Before you install NMDA and periodically after the installation, refer to the latest revision of this document on EMC Online Support at https://support.emc.com. On the Support by Product page, search for NMDA by typing the full product name in the Find a Product field. The product page provides the product-specific downloads, documentation, and additional content and tools.

Topics include:

l Revision history... 2

l Product description...5

l New features and changes... 8

l Fixed problems... 14

l Environment and system requirements... 17

l Known problems and limitations...18

l Technical notes...31

l Documentation... 39

l Software media, organization, and files... 40

l Installation... 41

l Troubleshooting and getting help...42

(2)

Revision history

The following table presents the revision history of this document.

Table 1 Revision history

Revision Date Description

12 April 25, 2016 Updated the following information:

l _{MySQL and MEB support on page 17}_{- Changed}

the MySQL 5.7.9 reference in the table to MySQL 5.7.

l _{Compatibility matrix on page 43}_{- Updated}

details in the compatibility matrix. 11 March 25, 2016 Updated the following information:

l _{Fixed problems in NMDA 8.2 SP3 on page 14}

-Added this topic about problem resolutions in NMDA 8.2 SP3.

l MySQL and MEB support on page 17 - Added

information about the new support of MEB 4.0 and MySQL 5.7.9 in NMDA 8.2 SP3.

Added to Known problems and limitations:

l _{ESC 25733 on page 23}_{, "NMDA backup might}

fail with many error messages in the NetWorker server log"

Added to Technical notes:

l Sybase backup cancellation in the NMC GUI on

page 39

10 December 21, 2015 Updated the following information:

l _{Sybase scheduled backup might fail if Sybase}

server and NetWorker client run in different locales on page 31 - Fixed two references to #!/bin/sh in the workaround paragraph.

l _{DB2 HPU restore might fail with an incremental}

backup on page 23 - Updated the final paragraph about the IBM HPU version that contains the DB2 bug resolution.

details in the compatibility matrix. 09 November 9, 2015 Updated the following information:

l _{Sybase ASE 16.0 SP02 or later backup server on}

AIX and Solaris SPARC on page 13 - Added this topic about NMDA support of Sybase ASE 16.0 SP02 or later on AIX and Solaris SPARC. 08 September 30, 2015 Updated the following information:

(3)

Table 1 Revision history (continued)

l _{Backup configuration changes with the NetWorker}

9.0 server on page 8 - Added this topic about the NMDA configuration requirements with the NetWorker 9.0 server.

l _{Support of DB2 10.5 fp5 or later and DB2 10.1 fp5}

or later on page 9 - Added this topic about new DB2 support in NMDA 8.2 SP2.

-Added this topic about problem resolutions in NMDA 8.2 SP2.

-Added 201099 to the fixed problems table. Added to Known problems and limitations:

l _{201099 on page 29}_{, "Oracle backup}

configuration with the wizard might time out with a large database"

l 229925 (ESC 23290) on page 30, "Oracle

backup fails with a core dump when a channel is allocated, released, and allocated again"

l _{236152 (ESC 24005) on page 30}_{, "The nsrdasv}

program might report an incorrect exit code to an Oracle backup script"

l 225756 (ESC 23001) on page 31, "Sybase

backup might fail when the backup server is busy and NSR_PARALLELISM is greater than 1" 07 September 4, 2015 Updated the following information:

l Fixed problems in NMDA 8.2 SP1 on page 15

-Added 177446 to the fixed problems table.

l _{MySQL and MEB support on page 17}_{- Added}

information about the support of MEB 3.12.

l Compatibility matrix on page 43 - Updated

details in the compatibility matrix. Added to Known problems and limitations:

l _{206229 on page 21}_{, "Push install of 32-bit}

NMDA fails on a 64-bit OS" 06 July 13, 2015 Updated the following information:

l _{Added the escalation number in the following}

topic headings:

n _{227747 (ESC 23153) on page 23} n _{228292 (ESC 23260) on page 23}

n _{234458, 235572 (ESC 23787) on page 23}

(4)

n _{233072, 233552 (ESC 23496) on page 25} n _{234431 (ESC 23780) on page 25}

l _{Updated information in the following topics:} n _{234458, 235572 (ESC 23787) on page 23} n _{233072, 233552 (ESC 23496) on page 25} l _{Compatibility matrix on page 43}_{- Updated}

details in the compatibility matrix. 05 June 30, 2015 Added to Known problems and limitations:

l _{227747 on page 23}_{, "DB2 backups might}

continue to fail after a network disconnection causes a backup failure"

l 228292 on page 23, "DB2 rollforward might

appear to become suspended"

l _{234458, 235572 on page 23}_{, "DB2 backups and}

restores might fail due to DB2 10.5 and 10.1 issues"

l 233072, 233552 on page 25, "Notes database

restore might fail after scanner re-creates the backup index"

l _{234431 on page 25}_{, "Lotus transaction log}

backup fails if NSR_CATALOGFILE is set" Updated the following information:

l Compatibility matrix on page 43 - Updated

details in the compatibility matrix. 04 May 6, 2015 Updated the following information:

l MySQL and MEB support on page 17 - Added

information about the support of MEB 3.11. Added to Known problems and limitations:

l _{227400, "Oracle backup with postcommand script}

might fail"

03 January 28, 2015 Updated for the release of NMDA 8.2 SP1. 02 August 29, 2014 Updated the following information:

l Fixed problems on page 14 - Added the

following bugs to the table of fixed problems: NW159657 (186596), NW160571 (186399), NW161178 (191984), NW161337 (193052). Updated the NW158246 problem summary.

l _{Table 8 on page 17}_{- Updated the list of}

(5)

details in the compatibility matrix. 01 June 18, 2014 Initial release of NMDA 8.2.

Product description

EMC®_NetWorker®_{Module for Databases and Applications (NMDA) is an add-on module}

for the NetWorker server and client that provides backup and recovery services for DB2, Informix, Lotus Domino/Notes, MySQL, Oracle, and Sybase ASE data.

NMDA replaces all the following legacy NetWorker modules on the supported UNIX, Linux, or Windows operating systems:

l NetWorker Module for DB2 (NMDB2) l NetWorker Module for Informix (NMI) l NetWorker Module for Lotus (NML) l NetWorker Module for Oracle (NMO) l NetWorker Module for Sybase (NMS)

NOTICE

You can use NMDA to restore backups that were performed with the legacy NetWorker modules.

You can use the existing legacy NetWorker module enabler to perform NMDA backups after the upgrade.

Major configuration changes from legacy NetWorker modules to NMDA on page 6

describes major configuration differences between the legacy modules and NMDA.

New features and changes on page 8 describes the major new features and enhancements in NMDA 8.2 and Service Packs. The appropriate release of the EMC NetWorker Module for Databases and Applications Administration Guide describes all the features that NMDA supports in the corresponding software release.

Environment and system requirements on page 17 provides information about the supported versions of the NetWorker software.

Known problems and limitations on page 18 describes the known problems and limitations in NMDA 8.2 and Service Packs.

Technical notes on page 31 describes existing NMDA features that require special considerations or require a later version of the NetWorker software.

Documentation on page 39 describes the NMDA 8.2 and Service Packs documentation and related documents. Installation on page 41 provides details about installing NMDA.

Before you install NMDA, review the compatibility matrix at the end of this document and the NMDA installation guide, administration guide, and other documentation at EMC Online Support.

(6)

Major configuration changes from legacy NetWorker modules to NMDA

The following table compares the major configurations in the legacy NetWorker modules and NMDA, and outlines any changes in those configurations.

You must set most NMDA parameters in the NMDA configuration file. The NMDA configuration file templates are in the following directories:

l On UNIX and Linux: /nsr/apps/config

l On Windows: NetWorker_install_path\apps\config

Table 2 Configuration changes from legacy NetWorker modules to NMDA

Feature or parameter in legacy NetWorker module Corresponding feature or parameter in NMDA

NSR_DEBUG_FILE = debug_file_full_path. Replaced by NSR_DIAGNOSTIC_DEST = debug_directory_full_path.

NSR_ENCRYPTION parameter. Parameter is dropped. Specify backup encryption through the NSR_AES_ENCRYPTION parameter.

NetWorker Module for DB2 (NMDB2):

Scheduled backup (configured without the wizard) performed through the nsrdb2sv -f configuration_file command.

Replaced by scheduled backup (configured without the wizard) performed through the nsrdasv -z configuration_file command.

Note

The NMDA configuration_file supports the NMDB2 configuration_file format.

Manual backup through the db2 commands. No change. Restore through the db2 commands. No change.

DB2_PSWD parameter. Replaced by USER_PSWD parameter. Set the password value in the parameter through the nsrdaadmin -P command.

DB2_USR parameter. Replaced by DB2_USER parameter.

NSR_DB2_BACKUP_INFO parameter. Replaced by NSR_DR_BACKUP_INFO parameter. If NSR_DB2_BACKUP_INFO was not set, NSR_DR_BACKUP_INFO is set to TRUE (default). NSR_DB2_CONFIG_FILE parameter. Replaced by NSR_DR_FILE_LIST parameter. NSR_*_DEBUG* parameters. Replaced by NSR_DEBUG_LEVEL parameter. NetWorker Module for Informix (NMI):

Scheduled backup performed through the nsrdbmi shell/batch script.

Replaced by scheduled backup performed through the nsrdasv -z configuration_file command.

Manual backup performed through the onbar commands. No change. Restore performed through the onbar commands. No change.

(7)

Table 2 Configuration changes from legacy NetWorker modules to NMDA (continued)

DO_BOOTFILE_BACKUPS parameter. Replaced by NSR_DR_BACKUP_INFO parameter. NetWorker Module for Lotus (NML):

Scheduled backup (configured without the wizard) or manual backup performed through the nsrnotesv -z

configuration_file command.

Replaced by scheduled backup (configured without the wizard) or manual backup performed through the nsrdasv -z configuration_file command.

Note

The NMDA configuration_file supports the NML configuration_file format.

Restore performed through the nsrnotesrc -z configuration_file command.

No change. Manual backup or restore performed through the NetWorker

User for Lotus GUI (nwbml.exe) on Windows only.

No change. Document-level recovery performed through the nsrdocrc -z

configuration_file command.

No change. Document-level recovery performed through the Notes client GUI

on Windows only.

No change.

LOTUSUSER parameter. Replaced by LOTUS_USER parameter.

NetWorker Module for Oracle (NMO):

Scheduled backup (configured without the wizard) performed through the nsrnmo shell/batch script.

Replaced by scheduled backup (configured without wizard) performed through the nsrdasv -z configuration_file command.

Manual backup performed through the Oracle RMAN interface. No change. Restore performed through the Oracle RMAN interface. No change.

NSR_SB_DEBUG_FILE parameter. Replaced by NSR_DEBUG_LEVEL=1 parameter. NetWorker Module for Sybase (NMS):

Scheduled backup performed through the nsrsyb shell/batch script.

Replaced by scheduled backup performed through the nsrdasv -z configuration_file command.

Manual backup performed through the nsrsybsv command_line_options command.

Replaced by manual backup performed through the nsrdasv -z configuration_file command.

Database consistency check performed through the nsrsybcc command_line_options command.

No change. Restore performed through the nsrsybrc

command_line_options command.

No change. Manual backup or restore performed through the NetWorker

User for Sybase GUI (nwbms.exe) on Windows only.

No change.

BACKUP_OPT parameter. Parameter is dropped.

-G, -R, and -T options of nsrsybsv command. Replaced by NSR_DUMP_LOG_OPT parameter.

(8)

Table 2 Configuration changes from legacy NetWorker modules to NMDA (continued)

-l option of nsrsybsv command. Replaced by NSR_BACKUP_LEVEL parameter. -p option of nsrsybsv command. Replaced by NSR_PROMOTE_FULL parameter. -P option of nsrsybsv command, or password setting in the

Client resource.

Replaced by USER_PSWD parameter. Set the password value in the parameter through the nsrdaadmin -P command. -r option of nsrsybsv command. Replaced by NSR_ASE_PASSWORD parameter.

-U option of nsrsybsv command, or remote user setting in the Client resource.

Replaced by SYBASE_USER parameter. -V option of nsrsybsv command. Replaced by NSR_ASE_VERIFY parameter.

New features and changes

The following topics describe the new features and changes in NMDA 8.2 and the related Service Packs.

NMDA 8.2 and Service Packs replace the previous NMDA releases on the supported Linux, UNIX, and Windows operating systems. You can use NMDA 8.2 and Service Packs to restore any backups performed with previous NMDA releases.

Note

Deprecated features are features supported in this NMDA release that will be removed and unsupported in a future NMDA release.

Backup configuration changes with the NetWorker 9.0 server

If you update the NetWorker server to release 9.0, you must change certain settings in the NMDA backup configurations to enable the NMDA backup operations with the updated NetWorker server. This requirement applies to all the NMDA 8.2 and Service Packs configurations.

With the NetWorker 9.0 server, you cannot create any new configurations of Avamar deduplication backups. However, you can continue to use existing backup configurations of Avamar backups, and you can continue to restore Avamar backups by using NMDA with the NetWorker 9.0 server.

The NetWorker 9.0 server does not support browse policy settings, and sets the browse policy to the same value as the retention policy. The browse policy field and resource are removed from the NMC 9.0 GUI. The NetWorker 9.0 server ignores any setting of the NSR_SAVESET_BROWSE parameter.

With the NetWorker 9.0 server, the NSR_DATA_VOLUME_POOL and

NSR_SAVESET_RETENTION parameters are no longer supported for scheduled backups. However, these parameters are still supported for manual backups. You can continue to use a scheduled backup configuration that was created with a pre-9.0 server. However, ensure that the destination pool and retention period values are set in the backup action that is configured for the scheduled backup. You can configure a backup action by using the Policy Action Wizard, described in the EMC NetWorker Version 9.0 Administration Guide.

(9)

The NetWorker 9.0 server introduces the new concepts of a data protection policy, backup workflow, and backup action for a scheduled backup configuration. The EMC NetWorker Version 9.0 Administration Guide and the EMC NetWorker Module for Databases and Applications Version 9.0 Administration Guide provide details about configuring a backup action, workflow, and data protection policy for a scheduled backup.

You can also specify the scheduled backup pool or retention policy in the Client resource. With a NetWorker 9.0 server, the attribute settings in the Client resource override the settings in the backup action. In the Client resource in the NMC 9.0 GUI, the Group attribute is renamed as Protection group list.

The NetWorker 9.0 server no longer supports a scheduled backup pool that is specified with the NSR_DATA_VOLUME_POOL parameter or associated with a backup group. For manual backups with the NetWorker 9.0 server, the parameter settings in the NMDA configuration file have a higher precedence than the corresponding settings in the Client resource.

NOTICE

The NetWorker 9.0 server does not support NSM backups and restores with pre-9.0 NetWorker client and NMDA software. To enable NSM operations with the NetWorker 9.0 server, update the NetWorker client and NMDA to release 9.0. Ensure that you install the NetWorker 9.0 extended client software. NSM backups that are configured with a pre-9.0 server continue to operate after the software update. However, you cannot perform a NetWorker 9.0 clone operation on any snapshot that was created with pre-9.0 NMDA, even after you have updated NMDA and the NetWorker client.

NMDA 8.2 SP2

The following topics describe the major new or enhanced features of NMDA 8.2 SP2. The EMC NetWorker Module for Databases and Applications Version 8.2 and Service Packs Administration Guide provides complete configuration details for the new features. The compatibility matrix at the end of this document provides details about the supported environments.

Support of DB2 10.5 fp5 or later and DB2 10.1 fp5 or later

NMDA 8.2 SP2 provides support of DB2 10.5 fp5 or later and DB2 10.1 fp5 or later. These DB2 fix packs provide an enhancement of DB2 native encryption. There is a known DB2 limitation due to the enhancement that might cause the NMDA wizard configuration or scheduled backup on AIX to fail. Refer to the DB2 troubleshooting information in the

EMC NetWorker Module for Databases and Applications Version 8.2 and Service Packs Administration Guide for details about a workaround that is recommended by IBM DB2.

NMDA 8.2 SP1

The following topics describe the major new or enhanced features of NMDA 8.2 SP1. The EMC NetWorker Module for Databases and Applications Version 8.2 and Service Packs Administration Guide provides complete configuration details for the new features. The compatibility matrix at the end of this document provides details about the supported environments.

Data Domain IPv6 support

NMDA 8.2 SP1 provides support for both the EMC®_{Data Domain}®_{IPv6 and mixed IPv4}

and IPv6 networks.

(10)

The EMC NetWorker Release 8.2 and Services Packs Release Notes provides information about the new support of Data Domain IPv6 networks with NetWorker software.

DB2 HADR support

NMDA 8.2 SP1 supports DB2 backups and restores in a DB2 High Availability Disaster Recovery (HADR) environment. The NetWorker scheduled backups always run on the active node in the HADR environment.

IBM DB2 has the following restrictions with HADR support that impact NMDA functionality:

l Multiple standby mode is available only with DB2 10.1 and later. l HADR is not available in partitioned database environments.

l DB2 does not support backups on the standby host, even if the Reads on Standby

(ROS) feature is enabled.

l Tablespace-level restore or rollforward is not supported on either a primary or

standby database.

NMDA supports DB2 HADR configuration with and without Virtual IP (VIP). NMDA has the following limitations with DB2 HADR support:

l You cannot use the NMDA wizard to configure an HADR backup in a VIP environment.

You must use the client-side manual configuration procedure through NMC.

l NMDA does not support DB2 HADR in a DB2 pureScale environment.

The EMC NetWorker Module for Databases and Applications Version 8.2 and Service Packs Administration Guide provides details about how to configure and perform DB2 HADR backups and restores.

DB2 load with copy yes option support

NMDA supports the use of the DB2 load command with the copy yes option to save a copy of DB2 database changes during the load operation to NetWorker for later use in the database recovery. Rollforward recovery must be enabled for the database before you can use the copy yes option with the load command.

The EMC NetWorker Module for Databases and Applications Version 8.2 and Service Packs Administration Guide provides details about how to use a load operation with the copy yes option to save a load-copy file and use a rollforward operation to recover the file into a database.

Sybase listonly option support

NMDA 8.2 SP1 supports the use of the Sybase restore command, nsrsybrc, with the listonly option for obtaining structure information from the latest backup of a Sybase instance or databases, without restoring the backup.

Sybase ASE 15.7 ESD#2 and later provides the listonly option support.

The EMC NetWorker Module for Databases and Applications Version 8.2 and Service Packs Administration Guide provides details about how to obtain the structure information from an NMDA Sybase backup by using the Sybase listonly option.

Sybase restore parameter support

In previous NMDA releases, the Sybase restore parameters were set as nsrsybrc command line options, described in the EMC NetWorker Module for Databases and Applications Command Reference Guide. These command line options are deprecated in NMDA 8.2 SP1, and support of the command line options will be removed in the future.

(11)

NMDA 8.2 SP1 provides new support for Sybase restore parameters in the NMDA configuration file. After setting the required restore parameters in the configuration file, you can perform a Sybase restore/recovery with the nsrsybrc -z

configuration_file_path command.

This NMDA release provides separate configuration file templates for the Sybase backup and restore parameters:

l nmda_sybase_backup.cfg—Template for Sybase backup parameters l nmda_sybase_restore.cfg—Template for Sybase restore parameters

The following table lists the parameters that you can set in the NMDA 8.2 SP1

configuration file for a Sybase restore/recovery. The table also lists the corresponding nsrsybrc command line options. The EMC NetWorker Module for Databases and

Applications Version 8.2 and Service Packs Administration Guide provides details about how to set the Sybase parameters.

Table 3 NMDA Sybase restore parameters and nsrsybrc command line options

Sybase restore parameters in NMDA 8.2

SP1 Corresponding command line options

DATABASE_ONLINE (new parameter) nsrsybrc -p

NSR_ASE_PASSWORD nsrsybrc -r backup_password

NSR_ASE_VERIFY nsrsybrc -V {header | full |

verifyonly}

NSR_BACKUP_PATHS nsrsybrc { SYBASE:/ASE_server_name | SYBASE:/ASE_server_name/database_name [ SYBASE:/ASE_server_name/

database_name...] }

NSR_CLIENT nsrsybrc -c client_name

NSR_DEBUG_LEVEL nsrsybrc -D debug_level

NSR_DIAGNOSTIC_DEST (new parameter) Not applicable NSR_DPRINTF (new debug parameter) Not applicable

NSR_ENCRYPTION_PHRASES nsrsybrc -e AES_pass_phrase

NSR_NWPATH Not applicable

NSR_RECOVER_POOL nsrsybrc -w recover_pool

NSR_RELOCATION_DEST (new parameter) nsrsybrc -d SYBASE:/ASE_server_name[/ database_name]

NSR_SERVER nsrsybrc -s server_name

NSR_XBSA_DEBUG (new debug parameter) Not applicable

RECOVER_LISTONLY (new parameter) nsrsybrc -l {load_sql | create_sql} RECOVER_UNTIL (new parameter) nsrsybrc -m {up_to_min |

up_to_min_noskip} -t backup_time STRIPE_DEVICE_DEST (new parameter) nsrsybrc -L temporary_file_pathname

(12)

Table 3 NMDA Sybase restore parameters and nsrsybrc command line options (continued)

Sybase restore parameters in NMDA 8.2

SP1 Corresponding command line options

SYBASE_USER nsrsybrc -U username

USE_CONSISTENCY_CHECK nsrsybrc -k

USER_PSWD nsrsybrc -P password

Updated NMDA 8.2 SP1 parameters

NMDA 8.2 SP1 supports the new parameter DB2_ACS_METADATA_DELETION_FORCE. You can use this parameter to enforce the deletion of the DB2 snapshot metadata save sets of a snapshot backup. You can enforce this deletion after a deletion of the

corresponding NSM snapshot backup succeeded in removing the snapshot save sets but failed to remove the associated metadata save sets.

NMDA 8.2 SP1 also provides new support for the Sybase restore parameters described in the preceding topic.

The EMC NetWorker Module for Databases and Applications Version 8.2 and Service Packs Administration Guide provides complete details about all the supported NMDA

parameters.

NMDA 8.2

The following topics describe the major new or enhanced features of NMDA 8.2.

The EMC NetWorker Module for Databases and Applications Release 8.2 Administration Guide

provides complete configuration details for the new features. The compatibility matrix at the end of this document provides details about the supported environments.

DB2 data recovery with the HPU utility

NMDA 8.2 introduces support for the IBM Optim High Performance Unload (HPU) utility, which is a high-speed stand-alone utility that unloads, extracts, and repartitions backup data into DB2 databases.

The HPU utility unloads data directly from an NMDA DB2 full, incremental, or delta backup to staging files and then writes the data to output files for use with the DB2 load utility. By reading the data directly from a backup image instead of through the DB2 software layer or database engine, the utility provides quick and efficient recovery of discrete volumes of data.

The IBM DB2 documentation provides details about HPU and how to create the control file for HPU operations.

NMDA support of the HPU utility requires HPU 4.2 Fixpack 1 or later.

Lotus transaction logs backup with multiple concurrent streams

NMDA 8.2 introduces support for the backup of Lotus archived transaction logs through multiple concurrent (parallel) streams during a full, incremental, or logs only backup. You can improve the backup performance by specifying that multiple concurrent data streams are sent to the NetWorker server for the backup.

For a logs only backup, you must set the NSR_PARALLELISM parameter or

(13)

send to the NetWorker server. During a full or incremental backup, the parameter or option that you set for the data backup is automatically used for the logs backup.

Oracle recovery wizard support of TNS_ADMIN environment variable

NMDA 8.2 enables you to set the Oracle TNS_ADMIN environment variable in the Oracle recovery configuration wizard to specify a nondefault location for the tnsnames.ora file. You need to set the parameter only if the file is not in the default location. The tnsnames.ora file is required to enable the proper database connections during an NMDA Oracle recovery.

Sybase ASE 16.0 SP02 or later backup server on AIX and Solaris SPARC

Sybase ASE 16.0 SP02 or later installs both 32-bit and 64-bit versions of the backup server on AIX and Solaris SPARC. NMDA supports only the 32-bit backup server with Sybase ASE 16.0 SP02 or later.

To enable the NMDA operations with the 32-bit backup server, follow the linking instructions in the latest revision of the EMC NetWorker Module for Databases and Applications Version 8.2 and Service Packs Installation Guide.

Sybase network password encryption

NMDA 8.2 supports the Sybase network password encryption feature. This feature encrypts login passwords for secure password transmission between the Sybase server and client.

The Sybase network password encryption was introduced in Sybase ASE 15.0.2 ESD #2, with support of a value from 0 to 2. A value of 1 restricts clients to use either the RSA public key encryption algorithms or the Sybase proprietary algorithms. A value of 2 restricts clients to use only the RSA encryption algorithms.

Starting with Sybase ASE 15.7 ESD #2, the network password encryption also supports a value of 3, which indicates that the server allows only incoming clients that use the EPEP login protocol. The Sybase documentation provides more details about network

password encryption.

NMDA automatically encrypts the Sybase login passwords by using the algorithms configured on the Sybase ASE server. No additional configuration is needed in NMDA.

Updated NMDA 8.2 parameters

NMDA 8.2 provides support for the new DB2 parameter SOURCE_DBNAME, which specifies the name of the DB2 database used for a backup of the archived logs. The parameter is mandatory for a rollforward recovery of the destination database after a restore to a different database.

NMDA 8.2 no longer supports the NSR_NO_BUSY_ERRORS parameter. Use the

NSR_MAX_START_RETRIES parameter instead. You do not need to use this parameter for regular operations, but only in the case of connectivity issues with the NetWorker server.

The NMDA Oracle parameter NSR_RETENTION_DISABLED is unsupported. The following results occurred if you used the parameter in a prior NMDA release:

l Backups remained on the backup device forever unless the DBA ran the delete

obsolete command on a regular basis with permissions to delete the backup entries.

l Clone operations failed.

If you set the NSR_RETENTION_DISABLED parameter in an RMAN backup script, the NMDA Oracle backup fails with an error message. Instead of using this parameter, either

(14)

you must use the NSR_ORACLE_RETENTION parameter if you use Oracle window-based policies or you must set the NetWorker retention policies correctly going forward. The EMC NetWorker Module for Databases and Applications Release 8.2 Administration Guide

provides details.

Fixed problems

The following topics provide details about the major fixes implemented in NMDA 8.2 and the related Service Packs.

Fixed problems in NMDA 8.2 SP3

The following table lists the major fixes implemented in NMDA 8.2 SP3.

Table 4 Problems fixed in NMDA 8.2 SP3

Issue number Problem summary Application

246782 (ESC 25797)

NMDA binaries might core dump on a Solaris SPARC 64-bit M-series system. All 222878 A DB2 NSM restore to a non-existent database might not restore the ACL of the database

correctly.

DB2 234237 (ESC

24795)

A DB2, Informix, Lotus, or Sybase backup might fail to start after many instances of the following error are recorded in the NetWorker server daemon log and the server becomes unresponsive:

nsrexecd GSS warning Session information (number #:#) for 'account/ client@domain' has expired after 120 minutes because a daemon has not requested the information

DB2, Informix, Lotus, Sybase

242534 (ESC 24583)

A Lotus restore is slow and appears to become suspended if there are many Lotus backups that include many DAOS files.

Lotus 247615 (ESC

25237)

After you perform an RMAN retention query, you cannot open the output file from the query. Oracle 251315 (ESC

25556)

After you update the NMDA software to release 8.2 SP2, an Oracle RMAN backup might fail with the following type of error message and produce a core dump:

RMAN-03009: failure of allocate command on c1 channel at 12/26/2015 04:43:33

ORA-03113: end-of-file on communication channel

Oracle

248187 (ESC 24921)

A Sybase scheduled backup does not stop immediately when you stop the backup group in the NMC GUI.

(15)

Fixed problems in NMDA 8.2 SP2

227747 (ESC 23153)

After a DB2 backup fails due to a network disconnection with the NetWorker server, subsequent DB2 backups might continue to fail in certain cases.

DB2 228292 (ESC

23260)

A DB2 rollforward operation might appear to become suspended if the NetWorker client file index includes many entries for non-DB2 backups.

DB2 234458,

235572 (ESC 23787)

Due to issues in DB2 10.5 and 10.1, such as APAR IT05079, a DB2 backup configuration in the wizard or a scheduled backup might fail. A restore of tablespaces in a DPF environment might fail.

DB2

233072, 233552 (ESC 23496)

After the NetWorker backup index is re-created by using the scanner program, a restore of a Lotus Domino database backup might fail.

Lotus Domino

234431 (ESC 23780)

If the NSR_CATALOGFILE parameter value is set in the configuration file or in the backup configuration wizard, then a Lotus transaction log backup fails.

Lotus Domino 227400 An Oracle scheduled backup with a postcommand script might fail if the script name starts with

nsrnmodr.

Oracle 229925 (ESC

23290)

An Oracle backup fails with a core dump when you allocate a channel, release the channel, and then allocate another channel in the RMAN backup script.

Oracle 236152 (ESC

24005)

When you use a backup script that runs the nsrdasv program to perform an Oracle RMAN backup, the nsrdasv command always reports the exit code 0.

Oracle 225756 (ESC

23001)

A Sybase backup might fail when the backup server is busy and NSR_PARALLELISM is greater than zero.

Sybase

Fixed problems in NMDA 8.2 SP1

177446 Client Direct backups fail with 32-bit NMDA on 64-bit Linux because the 64-bit NetWorker client release 8.2 does not include the 32-bit libDDBoost.so library. The issue is fixed in the 64-bit NetWorker client release 8.2 SP1.

Note

To enable Client Direct backups of a 32-bit database on 64-bit Linux, install the 64-bit NetWorker client release 8.2 SP1.

All

203697 The 32-bit NMDA software files are missing in the 64-bit NMDA zipped download file on Windows.

All

(16)

Table 6 Problems fixed in NMDA 8.2 SP1 (continued)

NW161970 For DB2 NSM backups, the DB2 metadata save set is not removed when a point-in-time save set is removed or released.

DB2 NW161515 A misleading message "Unknown error category 211" might cause confusion with a Domino

backup failure.

Lotus NW161419 An Oracle backup might fail with the following error message when the nsrmmd process is

busy:

Volume ID 'volume_ID' is not in media database. (2:3:73)

Oracle

NW161762 The nsrdasv program fails to correctly process the NLS_DATE_FORMAT parameter value that contains spaces.

Oracle NW161809 NSR_ORACLE_RETENTION is not honored if the parameter is set by the wizard. Oracle NW161962 Oracle backups complete successfully but do not release the FIPS140 resource correctly at the

end, which might cause Oracle to dump a core file.

Oracle 201099 An Oracle backup configuration operation with the wizard might time out when the

configuration includes a large Oracle database.

Oracle NW160422 The nsrsybrc command does not permit spaces between the -P option and the password. Sybase ASE NW161343 An NMDA Sybase restore cannot find a backup performed during Standard Time when the

restore is performed during Daylight Saving Time.

Sybase ASE

Fixed problems in NMDA 8.2

The following table lists the major fixes implemented in NMDA 8.2.

Table 7 Problems fixed in NMDA 8.2

NW154527, 186742

DB2 backup on Windows fails due to a db2sysc.exe process crash. DB2 NW158246 The restore of a DB2 NSM snapshot backup might fail with a crash of the db2vend process. DB2 NW161337,

193052

DB2 redirected recovery with the db2 "rollforward db" command fails to recover the transaction logs from the backup media.

DB2 NW155418 With the NSR_PREFETCH_LOGS parameter set to a nonzero value, a Lotus restore with the

nsrdocrc command does not prefetch any transaction logs.

Lotus Domino NW157119 Lotus recovery of a Domino .nsf file with many transaction logs does not stop when End

restore is selected, and the recovery continues to completion.

Lotus Domino NW160571,

186399

On Linux, a Lotus scheduled backup fails with a crash of the nsrdasv program when the checkpoint restart option (CPR) is enabled.

Lotus Domino NW161178,

191984

When you run the NetWorker User for Lotus GUI to recover a Lotus backup, the GUI might not display all the backed-up files included in the backup.

(17)

Table 7 Problems fixed in NMDA 8.2 (continued)

NW154716 Oracle recovery configuration wizard fails if the Oracle TNS_ADMIN environment variable specifies a nondefault location for the Oracle Net service files.

Oracle NW157752 NSR_RMAN_OUTPUT=SHELL setting on Windows does not generate the expected names for the

RMAN output log files.

Oracle NW159657,

186596

The nsrdasv program always returns a zero exit code for an Oracle backup failure. Oracle

Environment and system requirements

As client-side products, NMDA and the NetWorker client must reside on each database or application server that participates in the backup or restore. The NetWorker server and storage node can reside on either the same system or separate systems.

Unless otherwise specified, NMDA requires the NetWorker server, storage node, and client release 8.0 or later.

The compatibility matrix at the end of this document provides the current details on the software that NMDA 8.2 and related Service Packs require, such as supported operating systems and versions, database servers or application servers, NetWorker, and other supporting software.

The EMC NetWorker Module for Databases and Applications Administration Guide for the corresponding NMDA release provides details about the environment and system configurations required to operate NMDA.

MySQL and MEB support

NMDA interacts with the MySQL Enterprise Backup (MEB) software to perform MySQL backups and restores. You must install a supported version of MEB with the proper version of MySQL as shown in the following table.

Table 8 MySQL and MEB version dependencies

MySQL version Supported MEB versions

5.1 3.6, 3.7, 3.8, 3.9, 3.10

5.5 3.6, 3.7, 3.8, 3.9, 3.10, 3.11, 3.12

5.6 3.6, 3.7, 3.8, 3.9, 3.10, 3.11, 3.12

5.7 4.0

The following considerations apply to NMDA MySQL operations and MEB dependencies:

l 32-bit MEB must be installed for 32-bit MySQL support, and 64-bit MEB must be

installed for 64-bit MySQL support.

l NMDA does not support the MEB 3.9 “copy-back-and-apply-log” operation.

l With MEB 3.11, 3.12, and 4.0, NMDA does not support wizard configurations of the

partial MySQL backups of non-InnoDB databases. However, you can use the NMDA wizard to configure the partial MySQL backups of InnoDB databases.

(18)

l With MEB 3.11, 3.12, and 4.0, NMDA supports MySQL binary log backups only if you

disable the MEB log file backups by setting MYSQL_MEB_OPTIONS=--skip-binlog.

l With MEB 3.12 and 4.0, NMDA does not support cloud backup and restore with

OpenStack Object Storage (Swift) 1.0.

l With MEB 4.0, NMDA does not support the mysql.backup_sbt_history table on

the backed-up MySQL instance for accessing information about online backups to tape. You must use the mminfo or nsrinfo command to obtain information about NMDA MySQL backups to a NetWorker server.

The MySQL documentation provides details about MEB 4.0 limitations, which also affect the NMDA MySQL operations.

MySQL 5.6 support

NMDA supports MySQL release 5.6, including the new features except for the use of an encrypted login file, mylogin.cnf, for setting the backup credentials.

NMDA support of new features in MySQL 5.6 requires MEB version 3.8.1 or later to be installed on the MySQL server host. For a list of the new features supported only with MEB 3.8.1 or later, refer to the topic “Compatibility Note for MySQL Version” in the MySQL Enterprise Backup User’s Guide for version 3.8.1 or later.

Known problems and limitations

The following tables list the known problems and limitations that apply to NMDA 8.2 and the related Service Packs:

l Limitations in NMDA 8.2 and Service Packs on page 18

l Limitations in other products that affect NMDA 8.2 and Service Packs on page 20

The issues are grouped according to the specific application to which they apply. The first group includes the common issues that apply to multiple applications.

When you contact Customer Service about an issue, use the listed issue number. Unless the entry for a known limitation indicates that it is resolved for a specific release, the limitation applies to NMDA 8.2 and all the Service Packs.

A resolved limitation is listed in the fixed problems table for the release in which it is resolved.

Table 9 Limitations in NMDA 8.2 and Service Packs

Issue

number Description Operatingsystem Product feature Application

206229 on page 21

A push install operation fails to update the 32-bit NMDA software on a 64-bit OS.

All Push install All

224226 on page 21

On Linux, NMDA installations and operations fail if the NetWorker client package is installed in a nondefault location.

Linux Installation, backup, restore

(19)

Table 9 Limitations in NMDA 8.2 and Service Packs (continued)

Issue

ESC 25733 on page 23

An NMDA backup might fail and produce many instances of the following error message in the NetWorker server log file:

An existing connection was forcibly closed by the remote host

AIX, Linux Backup All

227747 (ESC 23153) on page 23

After a DB2 backup fails due to a network disconnection with the NetWorker server, subsequent DB2 backups might continue to fail in certain cases.

All Backup, archive DB2 228292 (ESC 23260) on page 23

A DB2 rollforward operation might appear to become suspended if the NetWorker client file index includes many entries for non-DB2 backups.

All Restore DB2

NW151464 on page 24

The savegroup report for a DB2 snapshot and rollover backup does not include any save set information from the rollover part of the backup.

Linux Backup,

NSM

DB2

NW158881 on page 25

With 32-bit NMDA on AIX, an Informix backup fails with the following error message:

Dependent module /usr/lib/liblocktbl.so could not be loaded.

AIX Backup Informix

233072, 233552 (ESC 23496) on page 25

After the NetWorker backup index is re-created by using the scanner program, a restore of a Lotus Domino database backup might fail.

All Restore Lotus Domino

234431 (ESC 23780) on page 25

All Backup Lotus Domino

NW101758, NW113768 on page 26

A Lotus backup of 64-bit Domino on AIX might cause the Domino server to crash or become suspended or the backup might fail with a specific error message.

AIX Backup Lotus Domino

NW121290 on page 27

A Lotus full database backup might expire before the dependent transaction log backups because NMDA Lotus does not create dependencies between the backups.

All Backup Lotus Domino

NW143964 on page 27

Conversion of a MySQL client-side to server-side configuration with the nsrdaadmin -W command fails. Linux Configuration conversion MySQL NW148580 on page 27

When you run the nsrinfo command to obtain information about MySQL backups, the command displays an error message.

Linux Backup MySQL

LGTsc06110 o n page 28

Japanese Oracle backup information might be displayed incorrectly in the NMC program.

All Backup,

internationalization

Oracle

(20)

Table 9 Limitations in NMDA 8.2 and Service Packs (continued)

Issue

NW150321 on page 28

NMDA Oracle save set bundling or policy uniformity fails for a multinode Oracle RAC backup, but the backup is successful and fully recoverable.

All Backup Oracle

201099 on page 29

An Oracle backup configuration operation with the wizard might time out when the configuration includes a large Oracle database.

All Configuration wizard Oracle

227400 An Oracle scheduled backup with a postcommand script might fail if the script name starts with nsrnmodr.

All Backup Oracle

229925 (ESC 23290) on page 30

All Backup Oracle

236152 (ESC 24005) on page 30

When you use a backup script that runs the nsrdasv program to perform an Oracle RMAN backup, the nsrdasv command always reports the exit code 0.

All Backup Oracle

NW105216 on page 30

If a Sybase manual or scheduled backup attempts to back up more than 200 databases, the backup might fail.

All Backup Sybase ASE

NW146800 on page 31

A Sybase scheduled backup might fail if the Sybase server and NetWorker client run in different locales.

225756 (ESC 23001) on page 31

A Sybase backup might fail when the backup server is busy and NSR_PARALLELISM is greater than 1.

Table 10 Limitations in other products that affect NMDA 8.2 and Service Packs

Issue

NW007000 on page 22

Due to a NetWorker limitation in non-English locales, backup configuration with the wizard fails.

All Configuration wizard, internationalization

All 181210 on

page 23

Due to a DB2 bug, if you use the HPU utility to unload and restore an NMDA DB2 incremental backup, then the operation might fail with an error.

All Restore DB2

234458, 235572 (ESC 23787) on page 23

Due to issues in DB2 10.5 and 10.1, such as APAR IT05079, a DB2 backup configuration in the wizard or a scheduled backup might fail. A restore of tablespaces in a DPF environment might fail.

All Backup,

restore

DB2

NW152810 on page 24

Due to a DB2 bug, if you run a multisession DB2 backup and use the db2 prune command to delete backup entries from the DB2 history file, the command might leave some entries in the file.

All Backup DB2

NW154550 on page 24

Due to a DB2 limitation, NMDA does not delete the logs from the NetWorker server during a db2 prune

(21)

Table 10 Limitations in other products that affect NMDA 8.2 and Service Packs (continued)

Issue

operation if the NMDA configuration file pathname in the LOGARCHOPTn parameter exceeds 30 ASCII characters.

NW019057 on page 28

Due to an Oracle limitation, Oracle backup problems might occur when the RMAN script contains non-ASCII data. All Backup, internationalization Oracle NW152664 on page 29

Due to Oracle bug 17080042, a point-in-time recovery of an Oracle 12c table, table partition, or pluggable database (PDB) might fail if you use manual channel allocation.

All Restore Oracle

NW154492 on page 29

Due to Oracle bug 17770652, an Oracle 12cR1 container or pluggable database snapshot backup with NSM fails on Windows if some of the datafiles are not located on a snapshotable device.

Windows Backup, NSM

Oracle

Common NMDA problems and limitations

This section contains the NMDA problems and limitations that apply to all the

applications (DB2, Informix, Lotus Domino/Notes, MySQL, Oracle, Sybase) that NMDA supports.

Push install of 32-bit NMDA fails on a 64-bit OS

206229

If you try to update 32-bit NMDA on a 64-bit OS by using the software distribution (push install) feature from the NetWorker server, the push install operation fails with the following type of error message:

nsrcpd NSR warning No upgrade clients available for the selected products

As a workaround, update the 32-bit NMDA software manually on a 64-bit OS:

1. Uninstall the existing 32-bit NMDA software according to the uninstall instructions in the NMDA installation guide for the module version.

2. Install the updated 32-bit NMDA software according to the install instructions in the latest NMDA installation guide.

Installations and operations fail on Linux with NetWorker client installed in nondefault location

224226

On Linux, if the NetWorker client package is installed in a nondefault location, then NMDA operations cannot load the FIPS140 shared libraries. As a result, the following issues occur:

l The NMDA installation fails with the following error message:

Unable to load FIPS-140 module: The FIPS140 shared library could not be found

(22)

The NMDA binaries are installed correctly, but the automated wizard registration fails with the error.

l All other NMDA operations fail with the same error message, including backups,

restores, and backup deletions.

As a workaround, perform the following steps:

1. Run the following command as the root user to create the required symbolic link: # ln -s /relocation_path/lib/nsr /usr/lib/nsr

2. If you want to use the NMDA wizard, then run the appropriate command from the following table as the root user to register the wizard.

Note

To manually unregister the wizard, use the command from the table but replace the option -i with -u.

Table 11 Command to register the wizard

Database or application Wizard registration command

DB2 nsrdb2ra -i Informix nsrifmxra -i Lotus nsrlotusra -i MySQL nsrmysqlra -i Oracle nsrorara -i Sybase nsrsybra -i

3. To enable NMDA MySQL operations, run the following commands as the root user to create the required symbolic links:

# ln -s /relocation_path/lib64/libmysqlapi.so.2 /usr/lib64/ libmysqlapi.so.2

# ln -s /relocation_path/lib64/libmysqlapiwrap.so /usr/lib64/ libmysqlapiwrap.so

# ln -s /relocation_path/lib/libnsrmysql.so /usr/lib/libnsrmysql.so

Backup configuration with the wizard fails in non-English locale

NW007000

Due to a NetWorker limitation in non-English locales, if the NetWorker server is installed in a nondefault path that includes any non-ASCII characters, the configuration of a backup with the wizard fails with the following error message:

Failed to open lockbox non-ASCII_dir_name\nsr\lockbox\client_name \clb.lb

As a workaround, perform either of the following actions:

l Configure the backup manually with the NMC method, without the wizard.

l Ensure that the installation path of the NetWorker server does not include any

(23)

NMDA backup might fail with many error messages in the NetWorker server log

ESC 25733

An NMDA backup might fail and produce many instances of the following error message in the NetWorker server log file:

An existing connection was forcibly closed by the remote host

The NetWorker server might also become suspended.

As a workaround, remove any suspended NMDA process, and restart the NetWorker server if required.

DB2 problems and limitations

This section contains NMDA problems and limitations specific to DB2 operations.

DB2 HPU restore might fail with an incremental backup

181210

Due to a DB2 bug, if you use the IBM Optim High Performance Unload (HPU) utility to unload and restore an NMDA DB2 incremental backup, then the unload and restore operation might fail with the INZU366I HPU return code 8 (reason code 0x283a001). The DB2 bug is resolved in IBM HPU v5.1.0.1.

DB2 backups might continue to fail after a network disconnection causes a backup failure

227747 (ESC 23153)

After a DB2 backup fails due to a network disconnection with the NetWorker server, subsequent DB2 backups might continue to fail in certain cases because the libnsrdb2 shared library blocks the subsequent backups even when the network connection is resumed.

The issue is resolved in NMDA 8.2 SP2.

DB2 rollforward might appear to become suspended

228292 (ESC 23260)

A DB2 rollforward operation might appear to become suspended if the NetWorker client file index includes many entries for non-DB2 backups. The rollforward operation incorrectly searches index entries in all namespaces instead of searching only the backup entries in the DB2 namespace, which might take an extended period of time. The issue is resolved in NMDA 8.2 SP2.

DB2 backups and restores might fail due to DB2 10.5 and 10.1 issues

234458, 235572 (ESC 23787)

Due to issues in DB2 10.5 and DB2 10.1, such as APAR IT05079, the following NMDA operations might fail:

l The configuration of a DB2 scheduled backup in the wizard might fail. l A DB2 scheduled backup might fail.

(24)

l A restore of tablespaces in a DPF environment might fail due to a change in the

vendor context input from DB2 with respect to the type of restore. The restore failure produces the following error message:

No match for a database image file was found based on the source database alias "database_name" and timestamp "backup_timestamp" provided

DB2 snapshot and rollover backup does not report the rollover save set details

NW151464

When you use NMDA with NSM to perform a DB2 snapshot and rollover backup, the savegroup report does not include any save set information from the rollover part of the backup. The savegroup report includes information from the snapshot backup only.

DB2 prune operation might fail to delete all entries from DB2 history file

NW152810

Due to a DB2 bug, if you run a multisession DB2 backup and then use the db2 prune command to delete the backup entries from the DB2 history file, the command deletes the index entries on the NetWorker server but might leave some of the entries in the DB2 history file.

The DB2 bug is fixed in DB2 9.7 fp4.

As a workaround, run the db2 prune command again without the and delete option to delete the remaining entries in the DB2 history file. For example:

db2 prune history 20130605180753 with force option and delete db2 prune history 20130605180753 with force option

DB2 prune operation does not delete the logs from the NetWorker server in certain cases

NW154550

NMDA does not delete the DB2 transaction logs from the NetWorker server during a DB2 prune operation if the NMDA configuration file pathname specified in the LOGARCHOPTn

parameter exceeds 30 ASCII characters. This character length limitation is a DB2

limitation. The LOGARCHOPTn parameter is set when configuring an automatic backup of DB2 transaction logs.

The issue is a result of DB2 not setting the NMDA configuration file pathname correctly in the Comment field of the corresponding DB2 history file record when the pathname exceeds 30 ASCII characters.

As a workaround, you can perform the following steps.

1. Run the db2 update history command to update the Comment field in the DB2 history file to point to an NMDA configuration file pathname that does not exceed 30 ASCII characters.

For example, the following db2 list history command output shows that the Comment field in the history file does not contain a valid NMDA configuration file pathname:

db2 list history archive log all for db hat List History File for hat

Op Obj Timestamp+Sequence Type Dev Earliest Log Current Log Backup ID -- --- ---- --- --- --

(25)

X D 20130920133146 1 O S0000017.LOG C0000000 Comment: Start Time: 20130920133146 End Time: 20130920133155 Status: P EID: 25 Location: /usr/lib/libnsrdb2.so

You can run the following db2 update history command to update the Comment field with a valid configuration file pathname:

db2 update history eid 25 with comment \"@/space/db2inst1/nmda.cfg \"

2. Rerun the db2 prune command.

To prevent this issue in future archive log backups, use an NMDA configuration file pathname that does not exceed 30 ASCII characters in the LOGARCHOPTn parameter.

Informix problems and limitations

This section contains NMDA problems and limitations specific to Informix operations.

Informix backup fails with 32-bit NMDA on AIX

NW158881

With 32-bit NMDA on AIX, an Informix backup fails with the following error message:

Dependent module /usr/lib/liblocktbl.so could not be loaded.

As a workaround, run the following command as root, to remove the following symbolic link before you perform any Informix backups:

rm /usr/lib/liblocktbl.so

This link removal ensures that the bit NMDA Informix library can locate the correct 32-bit libraries.

Lotus Domino/Notes problems and limitations

This section contains NMDA problems and limitations specific to Lotus Domino/Notes operations.

Notes database restore might fail after scanner re-creates the backup index

233072, 233552 (ESC 23496)

After the NetWorker backup index is re-created by using the scanner program, a restore of a Lotus Domino database backup might fail. The failure is caused when the backup index is not rebuilt correctly by the scanner program. The scanner program stops rebuilding the backup index if a backup save set includes a file that was skipped during the backup, for example, due to errors in opening the file.

Lotus transaction log backup fails if NSR_CATALOGFILE is set

234431 (ESC 23780)

(26)

Backup of 64-bit Domino on AIX might fail

NW101758, NW113768

An NMDA Lotus backup of 64-bit Domino on AIX might cause the Domino server to crash or become suspended or the backup might fail with the following error message:

notes_sess_open: Exiting with error: The call to NotesInitExtended failed: 433

To prevent the issues NW101758 and NW113768, perform the following steps on the AIX Domino host after the NMDA software is installed:

1. Log in as the Notes user, run the following commands, and record the command output:

> ulimit -a

> echo $LDR_CNTRL

2. Log in as the root user, and run the following commands: # ulimit -a

# echo $LDR_CNTRL

# dump -o /usr/bin/nsrexecd

If any of the following conditions is true, perform step 3 on page 26 to step 5 on page 26:

a. The data limit reported by the ulimit command for the root user is lower than for the Notes user.

b. The LDR_CNTRL value for the root user is different than for the Notes user. c. The output from the dump nsrexecd command contains the following

information:

maxDATA 0x80000000

3. Shut down the NetWorker client that is running on the Domino host by using the nsr_shutdown command.

4. Perform the appropriate action, depending on the issues found:

a. Set the data limit for the root user to the same value as for the Notes user. b. Set the LDR_CNTRL parameter for the root user to the same value as for the Notes

user.

c. Run the following ldedit command on the nsrexecd executable to decrease the amount of memory that it uses:

# /usr/ccs/bin/ldedit -bmaxdata:0x00000000 /usr/bin/nsrexecd

5. Restart the NetWorker client.

If the issue occurs on the AIX Domino host before you have performed the preceding steps:

1. Shut down the Domino server. 2. Reboot the whole machine.

(27)

Lotus full backups might expire before dependent transaction logs

NW121290

NMDA Lotus does not create dependencies between any transaction log backups (performed during an incremental backup or transaction log only backup) and full

database backups. As a result, a database backup might expire before the corresponding transaction log backups. You cannot use the transaction log backups to restore and recover the data without the full backup.

If you have encountered this issue with your backups, the backups expired from the NetWorker online indexes might not be deleted immediately from the backup media. As a workaround, run the scanner –i command or the nsrck –L7 command to rebuild the online indexes so that you can still use full backups for restore operations.

To reduce the possibility of encountering this issue, set the browse policy and retention policy of the Domino and Notes backups to a much longer time than the recovery point objective (RPO). For example, if the RPO is 1 week and you perform full backups on Sundays and incremental backups for the rest of the week, set the browse policy and retention policy to 2 weeks for both full backups and incremental backups. This setting ensures that the software retains the backups on the media to achieve the RPO of 1 week.

MySQL problems and limitations

This section contains NMDA problems and limitations specific to MySQL operations.

Conversion of MySQL client-side to server-side configuration fails

NW143964

When you use the nsrdaadmin -W command to convert a MySQL client-side

configuration to a server-side configuration, the conversion fails with the following error message:

75525:nsrdaadmin: NMDA wizard conversion is not supported for application type specified: 7.

As a workaround, use the wizard and select your existing client-side configuration file on the Specify the NMDA configuration file template page to create a new configuration. You can delete the old client-side configuration when the wizard configuration has completed successfully.

The nsrinfo command does not display MySQL backup information

NW148580

When you run the nsrinfo command to obtain information about MySQL backups from the NetWorker client file index, the command does not provide the backup information. You need to use the nsrinfo command to obtain specific types of information about MySQL backups, such as the binary log start number and log position information. The nsrinfo command used to query MySQL backup information displays the following error message:

90955:nsrinfo: The MySQL client attributes could not be decoded.

As a workaround, update the NetWorker client software to release 8.0.1.3 or later.

(28)

Oracle problems and limitations

This section contains NMDA problems and limitations specific to Oracle operations.

Japanese Oracle backup information might be displayed incorrectly in NMC program

LGTsc06110

When a scheduled NMDA Oracle backup is run in a Japanese environment and the output from RMAN or from a preprocessing or postprocessing script (used for the backup) is not

UTF8 text, the Japanese backup information displayed in the NMC program might appear incorrectly as box or question mark characters.

As a workaround, perform the following actions:

l Use the NSR_RMAN_ARGUMENTS parameter to redirect the RMAN output to a

specified log file.

l Specify "echo off" in any preprocessing and postprocessing scripts used for the

scheduled NMDA Oracle backup.

Backup problems might occur due to non-ASCII data in RMAN script

NW019057

Due to Oracle bug 3627132, an NMDA Oracle backup might fail or some parameters might be ignored in a non-English locale when the following conditions exist:

l NMDA internationalization is set up.

l The character sets of the RMAN script and Oracle database are different. l A non-ASCII character is included in the parms option or send command in the

RMAN script.

For example, if the database uses the UTF8 character set and the RMAN script uses the EUC-JP character set, an NMDA Oracle backup fails when the parms option or send command in the RMAN script contains any non-ASCII data.

As a result, the NMDA software does not support non-ASCII data in the RMAN script when the character set of the RMAN script (set through the NLS_LANG environment variable) differs from the character set of the Oracle database in a non-English locale.

NMDA Oracle save set bundling or policy uniformity fails for multinode RAC backups

NW150321

If you configure the NMDA Oracle save set bundling or policy uniformity feature, then a multinode Oracle RAC backup generates the following warning:

No information could be obtained from the backup piece file.

The backup is successful and fully recoverable, but the save set bundling or policy uniformity feature fails.

As a workaround:

l On UNIX or Linux, create a symbolic link that points the /nsr/apps directory on

every RAC node to a file system pathname on a shared disk, which is accessible from all the nodes.

l On Windows, set NSR_TMPDIR in the nwora.res file on every RAC node to a file

(29)

Oracle 12c point-in-time recovery might fail with manual channel allocation

NW152664

Due to Oracle bug 17080042, a point-in-time recovery of an Oracle 12c table, table partition, or pluggable database (PDB) might fail if you use manual channel allocation with the allocate channel command.

As a workaround, use automatic channel allocation with the configure channel command for a point-in-time recovery of an Oracle 12c table, table partition, or PDB.

Oracle 12cR1 container or pluggable database snapshot backup fails on Windows if some datafiles

are not on a snapshotable device

NW154492

Due to Oracle bug 17770652, an Oracle 12cR1 snapshot backup of a container or pluggable database on Windows fails when some of the datafiles are not located on a snapshotable device.

For example, you run the following RMAN command to back up a database that consists of datafiles 1 to 5, but datafiles 3 and 4 are not on a snapshotable device:

backup proxy full

format '%d_%U' database;

The backup fails with the following error message:

ORA-00600: internal error code, arguments: [krbbbdr_inconsistent_pdbid]

As a workaround, exclude from the proxy backup any datafiles that are not on a snapshotable device, and back up those datafiles in a separate backup command without the proxy option.

For example, in the previous example where datafiles 3 and 4 are not on a snapshotable device, run two separate backups by using the following separate RMAN commands:

backup proxy full format '%d_%U' datafile 1,2,5; backup full format '%d_%U' datafile 3,4;

Oracle backup configuration with the wizard might time out with a large database

201099

When you use the configuration wizard to create or modify an Oracle backup

configuration, the wizard operation might time out when the configuration includes a large Oracle database.

(30)

Oracle backup with postcommand script might fail

227400

You can use a postcommand script during an Oracle scheduled backup to back up files to prepare an Oracle server for disaster recovery. The backup with the postcommand script might fail if the script name starts with nsrnmodr.

Oracle backup fails with a core dump when a channel is allocated, released, and allocated again

229925 (ESC 23290)

As a workaround, do not allocate an additional channel after you release a channel in the RMAN backup script.

The nsrdasv program might report an incorrect exit code to an Oracle backup script

236152 (ESC 24005)

When you use a backup script that runs the nsrdasv program to perform an Oracle RMAN backup, the nsrdasv command always reports the exit code 0, which indicates a successful backup. If the backup fails, this exit code is incorrect, which prevents the proper diagnosis of the backup issue.

Sybase ASE problems and limitations

This section contains NMDA problems and limitations specific to Sybase ASE operations.

Backup of more than 200 databases might fail

NW105216

If a Sybase manual or scheduled backup attempts to back up more than 200 databases, the backup might fail.

Note

The limit of 200 databases is based on a maximum size of 30 bytes for a Sybase database name. The database limit might be higher for a backup of databases with names of smaller sizes.

As a workaround, perform two or more smaller Sybase backups instead of one large Sybase backup, where each smaller backup backs up fewer than 200 databases. For example, perform one of the following operations:

l Perform two or more Sybase manual backups with the nsrdasv command, with

each nsrdasv command backing up fewer than 200 databases.

l Perform two or more Sybase scheduled backups that each back up fewer than 200

databases. For each backup, create a separate Client resource and separate

configuration file with each configuration file specifying a difference set of databases for the backup.

(31)

Sybase scheduled backup might fail if Sybase server and NetWorker client run in different locales

NW146800

A Sybase scheduled backup might fail if the Sybase server and NetWorker client run in different locales.

As described in the EMC NetWorker Module for Databases and Applications Administration Guide, you must set the NSR_LOCALE parameter to the required locale in the NMDA configuration file for a Sybase manual backup configuration or a client-side scheduled backup configuration. However, NSR_LOCALE is not supported for a server-side scheduled backup or a probe-based backup.

As a workaround for a server-side scheduled backup or a probe-based backup, you can set the LANG environment variable in the NetWorker start script that is installed with the NetWorker client software on the Sybase server host. For example on Linux, insert the LANG setting lines at the beginning of the start script after the "#!/bin/sh" line:

#!/bin/sh

LANG=en_US.UTF-8 export LANG

Sybase backup might fail when the backup server is busy and NSR_PARALLELISM is greater than 1

225756 (ESC 23001)

An NMDA Sybase backup does not honor the NSR_MAX_START_RETRIES parameter setting when the backup server is busy and NSR_PARALLELISM is set to a value greater than 1. As a result, the backup fails with the following error message:

_nwbsa_is_retryable_error: received a network error (Severity 0 Number -13): busy

As a workaround, set the NSR_PARALLELISM parameter to the value 1. The issue is resolved in NMDA 8.2 SP2.

Technical notes

This section provides supplementary technical notes for implementing NMDA 8.2 and the related Service Packs. Some of these notes describe issues that are dependent on the functionality of other products used with NMDA.

Client Direct backups and restores

NMDA supports Client Direct backups to an AFTD or DD Boost device. NMDA also supports the Client Direct feature for NSM snapshot backups of DB2 or Oracle data to these types of devices.

Client Direct operations with an AFTD

NMDA supports Client Direct backups to and restores from an AFTD. This support requires specific NetWorker software, depending on the operating system:

l For an AFTD on UNIX or Linux, NetWorker server and storage node release 8.1 or later l For an AFTD on Windows, NetWorker server and storage node release 8.0 or later

Client Direct operations with a DD Boost device

NMDA supports Client Direct backups to and restores from a DD Boost device over an IP or Fibre Channel (FC) connection. NMDA supports the FC connection on all the platforms where NetWorker 8.1 or later provides the same support.