Resolved Issues

This section includes fixes that do not change the user interface and/or API behavior of the product. For a summary of changes, see Changes in this release.

If a resolution is associated with a reported issue, a Zendesk reference number is displayed, for example (10283).

Resolved in NuoDB 5.0

This section provides a summary of the issues resolved in the release of NuoDB 5.0 compared with the original NuoDB 4.3 release.

  • It was easy for users to misuse the 'Y' and 'y' date symbols in a date format string which resulted in inconsistent query results. Error checking and messaging to enforce correct use of these symbols has been added. For more information, see SQL Data Type Conversion Functions. (10102)
    The issue has been resolved.

  • Referencing the system.procedures table in statements was not case insensitive like other system meta data tables. (7032, 11255)
    The issue has been resolved.

  • A Transaction Engine (TE) could exit due to an Out-of-Memory (OOM) condition if a statement included many subqueries. (13176, 13253)
    The issue has been resolved.

  • nuoarchive check could fail with "ASSERT: Platform/PlatformRecords.cpp:4740: FAILED: transaction" while attempting to repair an index. (12254)
    The issue has been resolved.

  • NuoDB could fail with "corrupted atom (fingerprint mismatch, got <number>, expected 0)" when using database archives created with NuoDB versions prior to 3.0.2. (13231)
    The issue has been resolved.

  • An index could report "unable to verify uniqueness in index <index name> after 10 attempts" for each attempt to insert a particular key. For example, this error could be reported for a system table index when trying to create or truncate a table. (12986, 11883, 11901, 12258)
    The issue has been resolved.

  • After unique index creation, the exception "Uniqueness violation encountered while populating index <index name>, key = '<key>'" could be reported, resulting in unexpected database process shut down with message "FAILED: !objectIds.isFree(objId)". (13019)
    The issue has been resolved.

  • When using the Connection.setSchema function in the NuoDB JDBC Driver, the function would fail if the schema name included supported special characters. (13005)
    The issue has been resolved.

  • A Storage Manager (SM) could shutdown abruptly due to running out of memory while syncing a database containing extremely large records.
    The issue has been resolved.

  • Executing thousands of UNION ALL statements could result in an abnormal shutdown of the Transaction Engine (TE) due to an out-of-memory condition. (12966, 13087, 13094, 13105, 13129)
    The issue has been resolved.

  • A SELECT statement that included a GROUP BY, an ORDER BY, and window functions could return results in the incorrect order. In addition, available performance optimizations may not have been chosen.
    The issue has been resolved.

  • A Storage Manager (SM) container could not enter the ready state due to missing details on the generated command line that are required by the Kubernetes readiness probe. (13053)
    The issue has been resolved.

  • A Transaction Engine (TE) could shut down abruptly due to running out of memory when executing multiple queries with large number of query parameters. (13038)
    The issue has been resolved.

  • If it was detected at compilation time that a query could not return any results, the result set metadata would not populate.
    The issue has been resolved.

  • The exception "can’t find ResultSet" could occur due to the loss of ResultSet metadata after the commit.
    The issue has been resolved.

  • The nuocmd subcommands hotcopy-status and hotcopy-info with the --show-json option could report incorrect destinationDirectory information. (13011)
    The issue has been resolved.

  • Referencing a sequence with NEXT VALUE could result in connection shutdown with message, "Error 58000: Unexpected internal error requires closing client connection".
    The issue has been resolved.

  • Docker related subcommands could appear twice in the help text output when invoking, nuocmd <verb> -h and nuodocker <verb> -h.
    The issue has been resolved.

  • A Transaction Engine (TE) could not start when there is a Storage Manager (SM) in the syncing state. (12884)
    The issue has been resolved.

  • Creating a large number of tables in the same transaction could result in a performance degradation. (13016)
    The issue has been resolved.

  • If no data is sent from a client connection to the NuoDB Admin agentPort (48004), then the connection may close and produce error messages, "Server Failed to process message" and "Unable to parse Tag/XML from message". This error could occur when deploying NuoDB in Kubernetes using the NuoDB Helm Charts with admin.externalAccess.enabled=true.
    The issue has been resolved.

  • The Admin Process (AP) would shut down a Storage Manager (SM) during startup if journal recovery processed more than 8192 unreaped journal files. When this occurred, a java.lang.NumberFormatException error would appear in the nuoadmin.log file.
    The issue has been resolved.

  • When running NuoDB in containerized deployments, if a non-root user runs the ping operation in a Docker container, the "ping: socket: Operation not permitted." error is returned.
    The issue has been resolved.

    This issue exists in NuoDB 4.3.2 only.
  • The nuoarchive check command could not be run on an archive if a Storage Manager (SM) is shut down in the middle of transitioning the state of the archive from encrypted to decrypted.
    The issue has been resolved.

  • Creating multiple FOREIGN KEY constraints referencing the same table, including the same column, would results in an error indicating a duplicate value in a unique index. (9588)
    The issue has been resolved.

  • The nuoarchive check command could stop responding when referencing missing table atoms.
    The issue has been resolved.

  • If a statement included ten or more subqueries, a Transaction Engine (TE) could shut down abnormally with "Unable to free memory for 300 seconds". (13107, 13190, 13197)
    The issue has been resolved.

  • When the KILL CONNECTION SQL command was executed, in addition to the database connection, a Transaction Engine (TE) could shut down unexpectedly with "Engine connection to admin process closed". (13206)
    The issue has been resolved.

  • Renaming a column used to partition a table would return "Error: cannot rename column: table partitioned by this column". (12997)
    The issue has been resolved.

  • In a bare metal (physical/VM) deployment environment, if the Admin Process (AP) stopped while NuoDB engines (TE or SM) were still running on the same host, restarting the AP could fail with an error “Address already in use”. (12162)
    The issue has been resolved.

  • Using complex expressions with literal values in SQL Aggregate Functions or SQL Window Functions could be cast incorrectly and return wrong results.
    The issue has been resolved.

  • A SQL statement using "PARTITIONS" to name an identifier would return an error. (13249)
    The issue has been resolved.

  • From within a Data Definition Language (DDL) trigger it was possible to trigger another DDL trigger that could lead to a recursion error which returned, "Error 58000: maximum call stack depth of 15 exceeded".
    The issue has been resolved.

  • When using nuosql, every third line of a multi-line statement would mistakenly prompt the user, "missing semicolon?".(8566)
    The issue has been resolved.

  • If running ALTER TABLE ALTER COLUMN and a foreign key reference was added, a Transaction Engine (TE) could shut down unexpectedly with "Connection lost. Use RECONNECT command to continue."
    The issue has been resolved.

  • A SQL statement that failed to execute due to insufficient authorization was not recorded in the NuoDB Admin log file, nuoadmin.log.
    The issue has been resolved.

  • Executing GRANT …​ ON ALL TABLES IN SCHEMA …​ on a schema that contains a views object would return an error. The issue has been resolved.