Netezza Performance Server 11.2.1.14-IF1 release notes

Netezza Performance Server for Cloud Pak for Data System

Version 11.2.1.14-IF1 improves security, backup and restore reliability, and platform support, while fixing critical issues across connectivity, query processing, and system stability to deliver better performance, resilience, and diagnostics.

Note: Netezza Performance Server 11.2.1.14-IF1 is compatible with Cloud Pak for Data System 1.0.8.0 and later.

New features and enhancements

Authentication and security
  • The SET AUTHENTICATION command is enhanced to allow explicit selection of the Kerberos authentication mechanism (krb5 or SSPI) for JDBC and ODBC drivers.
  • Support tools are enhanced to function correctly when executed by Kerberos-authenticated users.
  • SSL certificates are renewed to maintain security compliance.
Backup and restore
  • Restore operations now support restoring views only.
  • Backup and Restore (BnR) queries now record actual user and application information in query history using CLIENT_USER_ID and CLIENT_APPLICATION_NAME.
  • Logging is enhanced to provide additional diagnostics for Backup and Restore chain inconsistency failures.
  • nzrestore logging is improved to clearly reflect enableSplitDelete behavior. In addition, the numRowsPerDelete threshold is reduced from 13.4 million to 13 million rows.
SPU, Monitoring, and Alerts
  • Enhanced the SendMail utility to support OAuth 2.0 and STARTTLS, enabling secure and modern SMTP authentication mechanisms.
Platform and installation
  • Installation issues in INZA caused by incorrectly generated dependency backup commands are resolved.
  • During nzupgrade, an SQL file is now generated that provides steps to drop user-created objects found in the definition or information_schema schemas.

Resolved issues

Client, driver and connectivity
  • Fixed an nzsql library issue for Red Hat Enterprise Linux 9 platform.
  • Fixed the issue where the legacy Netezza ODBC driver silently failed in minimal container environments due to the absence of the ps utility. Error handling and diagnostic visibility is improved.
  • Excessive "FATAL 1: Connection couldn't be established" messages in pg.log is reduced.
Backup, restore, and scheduler
  • Fixed data inconsistencies in backups caused by system restarts immediately after insert, update, or delete operations.
  • Scheduler rules are now correctly removed when the associated group is deleted.
  • Incremental restore update queries now correctly handle column names containing single quotes.
Query processing and engine stability
  • Fixed a crash during query abort involving subqueries.
  • Query hangs at the PostgreSQL layer caused by orphaned libraries is resolved.
  • Query crashes caused by a large number of constant expressions is resolved.
  • Accessing NULL UDX objects no longer results in segmentation faults.
  • Defensive fixes and additional logging are added to prevent NPS restarts caused by suspected double-free memory conditions.
  • Fixed an issue with administrative sessions getting stuck
  • Overall system stability is improved by ensuring proper termination of child processes (including SIGKILL fallback) and by preventing timer handlers from executing after resources are released.
SQL, metadata, and catalog
  • Identifiers containing colon characters (:) now return a validation error.
  • ALTER DATABASE OWNER no longer fails with an invalid schema error.
  • Duplicate entries in the _t_object catalog table are prevented.
  • Fixed a timing issue where creating tables with the same name under the same database and schema caused inconsistencies in catalog indexes.
  • The NZ_QUERY_HISTORY.SPOOLED_MB column now correctly reports spooled data usage.
Logging and configuration
  • Fixed the mention of Host Err logged at pg/dbos logs when the error is actually originating from SPU and should be mentioned as SPU err in these logs.
  • Fixed PostgreSQL log rotation issues
  • Diagnostic logging is enhanced to identify sessions involved when a deadlock is detected.
  • A System configuration parameter dbosWindowAggrWorkBlocks's value increased to 40960. This controls the amount of memory (in blocks/pages) allocated for SQL window aggregation operations on the host.
  • The wag_size parameter has been reverted to its default value of 200.
    Note: The default value was increased to 300 in 11.2.1.11 but restored to 200 in 11.2.1.14-IF1 after it caused compatibility issues with customer scripts.
    The following system views and built-in functions are affected by the wag_size:

    System views

    • _v_sys_columns
    • _v_sys_columns_xdb
    • _v_relation_column
    • _v_relation_column_xdb
    • _v_table_organize_column
    • _v_aggregate
    • _v_function
    • _v_procedure
    • _v_rules
    • _v_view
    • _v_view_xdb
    • _v_odbc_columns1
    • _v_odbc_columns2
    • _v_odbc_columns3
    • _v_jdbc_columns1
    • _v_jdbc_columns2
    • _v_jdbc_columns3
    • _v_oledb_views1
    • _v_dotnet_views1
    • _v_dotnet_views3
    • _v_sched_rule

    Built-in functions

    • format_type
    • format_sql_type
    • format_sql_type_table
    • oidvectortypes
    • obj_description
    • version
    • timeofday
    • array_dims
    • timezone
    • timestamp_tz
    • to_char
    • to_ascii
    • get_viewdef
    • get_ruledef
    • get_schedruledef
    • get_indexdef
    • json_typeof
    • jsonb_typeof
  • The disable_mem_check option is enabled by default to improve performance for complex queries.
  • The enable_sproc_hist_logging option is disabled by default to avoid unnecessary system restarts.
Security updates
  • The IBM Java package is upgraded to version 8.0.8.55, addressing multiple security vulnerabilities across CVEs from 2021 through 2025.