9.3.x Resolved Issues¶
9.3.1-31¶
Release Date: 01/11/2022
| Issue Key | Component | Description |
|---|---|---|
| VER-80161 | UI - Management Console | The CVE-2021-45105 security vulnerability was found in earlier versions of the Apache log4j library used by the MC. The library has been updated to resolve this issue. 在较早版本的 MC 使用的 Apache log4j 库中发现了安全漏洞 CVE-2021-45105。该库已更新以解决此问题。 |
| VER-80172 | Kafka Integration | Security vulnerabilities CVE-2021-45105 and CVE-2021-44832 were found in earlier versions of the log4j library used by the Vertica/Apache Kafka integration. The library has been updated to resolve this issue. 在较早版本的 Vertica/Apache Kafka 集成使用的 log4j 库中发现了安全漏洞 CVE-2021-45105 和 CVE-2021-44832。该库已更新以解决此问题。 |
| VER-80245 | UI - Management Console | The CVE-2021-44832 security vulnerability was found in earlier versions of the Apache log4j library used by the MC. The library has been updated to resolve this issue. 在较早版本的 MC 使用的 Apache log4j 库中发现了安全漏洞 CVE-2021-44832。该库已更新以解决此问题。 |
9.3.1-30¶
Release Date: 12/23/2021
| Issue Key | Component | Description |
|---|---|---|
| VER-80114 | Kafka Integration | Security vulnerabilities CVE-2021-44228 and CVE-2021-45046 were found in earlier versions of the log4j library used by the Vertica/Apache Kafka integration. The library has been updated to resolve this issue. 在较早版本的 Vertica/Apache Kafka 集成使用的 log4j 库中发现了安全漏洞 CVE-2021-44228 和 CVE-2021-45046。该库已更新以解决此问题。 |
| VER-80113 | UI - Management Console | Security vulnerabilities CVE-2021-44228 and CVE-2021-45046 were found in earlier versions of the log4j library used by the MC. The library has been updated to resolve this issue. 在较早版本的 MC 使用的 log4j 库中发现了安全漏洞 CVE-2021-44228 和 CVE-2021-45046。该库已更新以解决此问题。 |
9.3.1-28¶
Release Date: 10/20/2021
| Issue Key | Component | Description |
|---|---|---|
| VER-78126 | Catalog Engine | Vertica now restarts properly for nodes with very large checkpoint files. Vertica 现在可以正确重启具有非常大检查点文件的节点。 |
| VER-79048 | Optimizer | Queries with multiple distinct aggregates sometimes produced wrong results when inputs appeared segmented on same columns as distinct aggregate arguments — now fixed. 当输入在与 DISTINCT 聚合参数相同的列上分段时,包含多个 DISTINCT 聚合的查询有时会产生错误结果 — 现已修复。 |
9.3.1-27¶
Release Date: 9/27/2021
| Issue Key | Component | Description |
|---|---|---|
| VER-78948 | UI - Management Console | MC returned errors when configuring email gateway aliases with hyphen characters — resolved. 当使用连字符配置电子邮件网关别名时,MC 返回错误 — 已解决。 |
9.3.1-26¶
Release Date: 8/11/2021
| Issue Key | Component | Description |
|---|---|---|
| VER-77636 | Data Export | Added support for exporting UUID types via s3export; previously would sometimes crash the initiator node. 增加了通过 s3export 导出 UUID 类型的支持;此前有时会导致发起节点崩溃。 |
| VER-77846 | Kafka Integration | Kafka Scheduler now allows an initial offset of -3 (begin reading from consumer group offset). Kafka Scheduler 现在允许初始偏移量为 -3(从消费者组偏移量开始读取)。 |
| VER-78134 | Kafka Integration | The Kafka scheduler no longer hangs in rare cases. Kafka 调度器在极少数情况下不再挂起。 |
| VER-78349 | Data Networking | In rare cases, the socket for internal connections could erroneously close and send many error messages — resolved. 在极少数情况下,用于内部连接的套接字可能错误关闭并发送大量错误消息 — 已解决。 |
| VER-78368 | ILM, Tuple Mover | Force-split operations (copy/drop/move/swap partitions) could fail when Tuple Mover attempted to merge split ROS containers before step 2 completed — now the two stages are part of a single transaction. 当 Tuple Mover 在第 2 步完成之前尝试合并已拆分的 ROS 容器时,强制拆分操作(复制/删除/移动/交换分区)可能会失败 — 现在这两个阶段是单个事务的一部分。 |
9.3.1-25¶
Release Date: 6/10/2021
| Issue Key | Component | Description |
|---|---|---|
| VER-76628 | Recovery | Restarting a node with --force on a node with missing/corrupted DFS file now recovers from an available UP node or drops with a warning. 在缺少/损坏 DFS 文件的节点上使用 --force 重启节点,现在可以从可用的 UP 节点恢复或发出警告后删除。 |
| VER-77284 | Security | VoltageSecureProtect() now guarantees a properly sized buffer for ciphertexts with Unicode FPE formats. VoltageSecureProtect() 现在保证为 Unicode FPE 格式的密文提供适当大小的缓冲区。 |
| VER-77297 | Backup/DR | Vbr can now delete restore points that are already partially deleted. Vbr 现在可以删除已经部分删除的还原点。 |
| VER-77411 | Hadoop | Vertica now properly parses uncompressed parquet files. Vertica 现在可以正确解析未压缩的 parquet 文件。 |
| VER-77525 | Security | Vertica now auto-creates needed default key projections for a user with DML access when inserting into a table with a primary key and no projections. 当具有 DML 访问权限的用户向具有主键但无投影的表中插入数据时,Vertica 现在会自动创建所需的默认键投影。 |
9.3.1-24¶
Release Date: 5/21/2021
| Issue Key | Component | Description |
|---|---|---|
| VER-77193 | Execution Engine | Queries on v_internal tables with equality predicates on relname/nspname columns were not optimized when expressions contained embedded characters (underscore/percent) or multiple predicates queried the same column — now optimized. 对 v_internal 表的查询在 relname/nspname 列上使用等值谓词时,如果表达式包含嵌入字符(下划线/百分号)或多个谓词查询同一列,则未进行优化 — 现已优化。 |
| VER-77183 | Spread | Vertica now properly detects duplicate tokens with the same arq ID. Vertica 现在可以正确检测具有相同 arq ID 的重复令牌。 |
| VER-77156 | Tuple Mover | Previously, Tuple Mover merged all eligible ROS containers without considering resource pool capacity, causing mergeout failure — now it accounts for capacity and adjusts container count. 以前,Tuple Mover 合并所有符合条件的 ROS 容器而不考虑资源池容量,导致 mergeout 失败 — 现在它会考虑容量并调整容器数量。 |
| VER-76974 | Backup/DR | Object restore now functions properly for objects >1 TB on Eon Mode clusters running on S3. 在 S3 上运行的 Eon 模式集群中,对象恢复现在可以正确处理大于 1 TB 的对象。 |
| VER-76657 | UI - Management Console | MC now properly allows completing database design for a non-K-Safe database. MC 现在允许为非 K-Safe 数据库完成数据库设计。 |
| VER-76623 | Installation Program, UI - Management Console | The MC installer has been optimized to reduce its storage footprint. MC 安装程序已优化以减小其存储占用。 |
9.3.1-23¶
Release Date: 4/20/2021
| Issue Key | Component | Description |
|---|---|---|
| VER-76860 | Optimizer | Incorrect usage of EXPORT_TABLES in EXPORT TO VERTICA caused database crash — now returns an error instead. 在 EXPORT TO VERTICA 中错误使用 EXPORT_TABLES 导致数据库崩溃 — 现在改为返回错误。 |
9.3.1-22¶
Release Date: 4/5/2021
| Issue Key | Component | Description |
|---|---|---|
| VER-75789 | Kafka Integration | Data ingested through Kafka during a timezone change could result in duplicate records — resolved. 在时区变更期间通过 Kafka 摄入的数据可能导致重复记录 — 已解决。 |
| VER-75903 | Tuple Mover | With Mergeout Cache enabled, dc_mergeout_requests now contains valid transaction IDs instead of zero. 启用 Mergeout Cache 后,dc_mergeout_requests 现在包含有效的事务 ID,而不是零。 |
| VER-75941 | Tuple Mover | Tuple Mover logged excessive PURGE requests on a projection while another MERGEOUT ran on the same projection — resolved. 当另一个 MERGEOUT 在同一投影上运行时,Tuple Mover 记录了过多的 PURGE 请求 — 已解决。 |
| VER-75975 | Hadoop | Fixed a bug in Parquet Predicate Pushdown where correct rowgroups were not pruned based on predicates on HDFS files. 修复了 Parquet 谓词下推中的一个 bug,即未根据 HDFS 文件上的谓词正确剪裁 rowgroup。 |
| VER-76102 | Tuple Mover | Tuple Mover dequeued DVMERGEOUT and MERGEOUT simultaneously and only executed DVMERGEOUT, leaving MERGEOUT pending — now it checks for outstanding MERGEOUT after DVMERGEOUT completes. Tuple Mover 同时出队了 DVMERGEOUT 和 MERGEOUT,但只执行了 DVMERGEOUT,导致 MERGEOUT 一直挂起 — 现在它会在 DVMERGEOUT 完成后检查是否有未完成的 MERGEOUT。 |
| VER-76126 | Monitoring | Output of get_expected_recovery_epoch is no longer truncated in clusters with >20 nodes. 在超过 20 个节点的集群中,get_expected_recovery_epoch 的输出不再被截断。 |
| VER-76254 | Data Export | Parquet files exported by Vertica now properly display date values when imported to Impala. Vertica 导出的 Parquet 文件在导入到 Impala 时现在可以正确显示日期值。 |
| VER-76258 | Hadoop | Vertica now properly handles queries of external tables of Parquet files with ZSTD compression. Vertica 现在可以正确处理对使用 ZSTD 压缩的 Parquet 文件外部表的查询。 |
| VER-76402 | Kafka Integration | Kafka UDXs now properly support sasl.mechanism types SCRAM-SHA-512 and SCRAM-SHA-256. Kafka UDX 现在正确支持 sasl.mechanism 类型 SCRAM-SHA-512 和 SCRAM-SHA-256。 |
| VER-76478 | Execution Engine | Complex memory-intensive queries with an analytic function feeding into a merge could crash on high core counts with high EXECUTIONPARALLELISM — resolved. 在高核心数和高 EXECUTIONPARALLELISM 下,包含递送到 merge 的分析函数的复杂内存密集型查询可能崩溃 — 已解决。 |
9.3.1-21¶
Release Date: 2/9/2021
| Issue Key | Component | Description |
|---|---|---|
| VER-75640 | Optimizer | Queries with subqueries on tables with TopK projections sometimes returned an error — resolved. 对具有 TopK 投影的表使用子查询有时会返回错误 — 已解决。 |
| VER-75785 | Optimizer | Setting a projection column comment to empty string with COMMENT ON COLUMN caused database crash — resolved. 使用 COMMENT ON COLUMN 将投影列注释设置为空字符串导致数据库崩溃 — 已解决。 |
| VER-75787 | Machine Learning | RF_CLASSIFIER no longer triggers out-of-memory errors in response to some parameter combinations. RF_CLASSIFIER 在某些参数组合下不再触发内存不足错误。 |
9.3.1-20¶
Release Date: 1/22/2021
| Issue Key | Component | Description |
|---|---|---|
| VER-75394 | Execution Engine | Subqueries used in expressions with correlated joins and aggregated output sometimes returned multiple rows per join key — resolved. 在带有关联连接和聚合输出的表达式中使用的子查询有时会为每个连接键返回多行 — 已解决。 |
| VER-75505 | Admin Tools, Spread | A new database no longer fails to start after quickly stopping and starting a different database due to stale Spread connections. 由于过期的 Spread 连接,快速停止和启动不同数据库后,新数据库不再无法启动。 |
| VER-75512 | Security | LDAP Link now supports ranges of LDAP attribute values. LDAP Link 现在支持 LDAP 属性值的范围。 |
| VER-75533 | Security, UI - Agent | Client initiated SSL/TLS renegotiation is now disabled on port 5444. 客户端发起的 SSL/TLS 重新协商现已在端口 5444 上禁用。 |
| VER-75534 | Optimizer | Analytic functions triggered an Optimizer error when used with a LIMIT OVER clause — resolved. 分析函数与 LIMIT OVER 子句一起使用时触发了优化器错误 — 已解决。 |
9.3.1-19¶
Release Date: 12/14/2020
| Issue Key | Component | Description |
|---|---|---|
| VER-74582 | Execution Engine | The size of the METADATA resource pool was calculated incorrectly under some circumstances — resolved. 在某些情况下,METADATA 资源池的大小计算不正确 — 已解决。 |
| VER-74963 | Storage and Access Layer | Under certain conditions, Vertica would crash while cancelling an insert/copy transaction on EON — corrected. 在特定条件下,Vertica 在取消 EON 上的插入/复制事务时会崩溃 — 已纠正。 |
| VER-75012 | S3 | Clearing AWSSessionToken session-level parameter no longer displays a VIAssert error. 清除 AWSSessionToken 会话级参数不再显示 VIAssert 错误。 |
| VER-75013 | Tuple Mover | Enabling both mergeout caching and reflexive mergeout adversely affected performance on new/restarted databases — now reflexive mergeout auto-disables mergeout caching. 同时启用 mergeout 缓存和 reflexive mergeout 对新启动/重启的数据库性能产生了不利影响 — 现在 reflexive mergeout 会自动禁用 mergeout 缓存。 |
| VER-75016 | Hadoop | Vertica node no longer fails when querying an external table copying from partitioned parquet with incorrect filter data type. 查询从具有错误过滤器数据类型的已分区 parquet 复制的外部表时,Vertica 节点不再失败。 |
| VER-75094 | Execution Engine | Inter-node encryption adversely affected query performance — addressed with improved performance in many cases. 节点间加密对查询性能产生了不利影响 — 已在许多情况下改善了性能。 |
| VER-75248 | Security | CREATE OR REPLACE VIEW now requires only DROP privilege (plus CREATE on schema) rather than ownership to replace a view. CREATE OR REPLACE VIEW 现在只需要 DROP 权限(加上 schema 上的 CREATE 权限)而不是所有权来替换视图。 |
9.3.1-18¶
Release Date: 11/19/2020
| Issue Key | Component | Description |
|---|---|---|
| VER-74649 | UI - Management Console | With multiple IPs on a cluster, Vertica now prioritizes reachable IP host addresses when trying JDBC connections. 当集群有多个 IP 时,Vertica 现在在尝试 JDBC 连接时优先选择可达的 IP 主机地址。 |
| VER-74714 | UI - Management Console | MC now confirms the Agent IP is reachable from the MC host when multiple IPs are available. 当有多个 IP 可用时,MC 现在会确认 MC 主机可以到达 Agent IP。 |
9.3.1-17¶
Release Date: 10/5/2020
| Issue Key | Component | Description |
|---|---|---|
| VER-73872 | Execution Engine | Querying UUID data types with an IN operator ran significantly slower than using OR — resolved. 使用 IN 运算符查询 UUID 数据类型的速度明显慢于使用 OR — 已解决。 |
| VER-74017 | Backup/DR | Users can now restore full database backups into an Eon mode cluster where a node has a changed IP address. 用户现在可以将完整数据库备份还原到节点 IP 地址已更改的 Eon 模式集群中。 |
| VER-74169 | Data Load / COPY | Added support guidance parameter 'UDLMaxDataBufferSize' (default 256 MB) to avoid insufficient memory errors during User Defined Load. 添加了支持指导参数 'UDLMaxDataBufferSize'(默认 256 MB),以避免在用户自定义加载期间出现内存不足错误。 |
9.3.1-16¶
Release Date: 9/15/2020
| Issue Key | Component | Description |
|---|---|---|
| VER-74003 | Execution Engine | The query_profiles table now reports correct rows processed when querying external tables stored in parquet on HDFS. 在查询存储在 HDFS 上 parquet 格式的外部表时,query_profiles 表现在报告正确的处理行数。 |
| VER-74134 | Execution Engine | Queries sometimes threw a Sort Order Violation error when join output sorted like the target projection was consumed by a single Analytic function — resolved. 当按目标投影排序的连接输出被单个分析函数消费时,查询有时会抛出排序顺序违规错误 — 已解决。 |
9.3.1-15¶
Release Date: 8/27/2020
| Issue Key | Component | Description |
|---|---|---|
| VER-73871 | Tuple Mover | If mergeout on a purge request failed due to table lock, subsequent purge mergeout requests for the same table were blocked — resolved. 如果由于表锁导致清理请求的 mergeout 失败,同一表的后续清理 mergeout 请求会被阻塞 — 已解决。 |
| VER-73875 | Backup/DR | Added a meaningful error message when a user terminated a database snapshot session and objects were deleted. 当用户终止数据库快照会话且对象被删除时,添加了有意义的错误消息。 |
| VER-73870 | Storage and Access Layer | After reading from Google Cloud Storage, Vertica now closes TCP connections when users close their sessions. 从 Google Cloud Storage 读取后,Vertica 现在会在用户关闭会话时关闭 TCP 连接。 |
9.3.1-14¶
Release Date: 8/12/2020
| Issue Key | Component | Description |
|---|---|---|
| VER-73366 | Optimizer | Common sub-expressions in INSERT...SELECT were not reused, causing performance degradation; EXPLAIN plans also rendered them incorrectly — both resolved. INSERT...SELECT 中的公共子表达式未被重用,导致性能下降;EXPLAIN 计划也错误地呈现了它们 — 两者均已解决。 |
| VER-73465 | Optimizer | Queries on certain projections failed to find appropriate partition-level statistics — now queries use partition-level statistics as needed. 对某些投影的查询未能找到合适的分区级统计信息 — 现在查询会根据需要使用分区级统计信息。 |
| VER-73469 | DDL | Renaming a schema was not propagated to table DEFAULT expressions using sequences of that schema — now all sequences are updated with the new schema name. 重命名 schema 未传播到使用该 schema 序列的表 DEFAULT 表达式 — 现在所有序列都会更新为新的 schema 名称。 |
| VER-73471 | Data Load / COPY | FJsonParser rejecting rows could cause the database to panic under certain circumstances — resolved. FJsonParser 拒绝行在某些情况下可能导致数据库 panic — 已解决。 |
| VER-73481 | Complex Types, Optimizer | Directed queries sometimes caused "Unknown Expr" warnings with queries involving arrays — resolved. 定向查询有时会在涉及数组的查询中导致 "Unknown Expr" 警告 — 已解决。 |
| VER-73484 | Tuple Mover | A large ROS of inactive partitions was excluded from mergeout — now merged with smaller ROS when total containers are near pushback threshold and total size is within limit. 非活动分区的大型 ROS 被排除在 mergeout 之外 — 现在当容器总数接近反压阈值且总大小在限制内时,会与较小的 ROS 合并。 |
| VER-73723 | Data Load / COPY | Copy of poorly written parquet files (generated by GO) with wide varchar columns was very slow — now fixed. 复制包含宽 varchar 列的编写不良的 parquet 文件(由 GO 生成)非常慢 — 现已修复。 |
9.3.1-13¶
Release Date: 7/17/2020
| Issue Key | Component | Description |
|---|---|---|
| VER-73137 | Hadoop | Vertica would sometimes connect to the wrong HDFS cluster when several namenodes from different clusters had the same hostname — fixed. 当来自不同集群的多个 namenode 具有相同主机名时,Vertica 有时会连接到错误的 HDFS 集群 — 已修复。 |
| VER-73260 | DDL - Projection | Users without required privileges could create projections using createType hints L, D, or P — resolved. 没有所需权限的用户可以使用 createType 提示 L、D 或 P 创建投影 — 已解决。 |
| VER-73263 | Execution Engine | Cancelling a query caused subsequent queries in the same transaction to return with a cancellation error — resolved. 取消查询导致同一事务中的后续查询返回取消错误 — 已解决。 |
| VER-73266 | Control Networking | Adding a column to a wide table with many storage containers could cause server failure — fixed. 向具有多个存储容器的宽表添加列可能导致服务器故障 — 已修复。 |
9.3.1-12¶
Release Date: 7/1/2020
| Issue Key | Component | Description |
|---|---|---|
| VER-72905 | Data Load / COPY, SDK | The fenced mode protocol for user-defined parsers had a bug causing the parser to process the same data multiple times — resolved. 用户自定义解析器的隔离模式协议存在一个 bug,导致解析器多次处理相同数据 — 已解决。 |
| VER-72909 | Security | The LDAPLinkSearchTimeout configuration parameter has been restored. LDAPLinkSearchTimeout 配置参数已恢复。 |
| VER-72948 | Backup/DR | Backups were failing to delete old restore points when backup objects exceeded 10,000 — resolved. 当备份对象超过 10,000 个时,备份无法删除旧的还原点 — 已解决。 |
| VER-72968 | Backup/DR | Vertica no longer fails if an object is being restored at the same time an object is being created with the same name in different character cases. 如果正在还原的对象同时有另一个不同大小写的同名对象正在创建,Vertica 不再失败。 |
| VER-72974 | Kafka Integration, Security | When omitting --ssl-ca-alias, the scheduler loads all certificates into the trust store; when given an alias, only that alias is loaded. 省略 --ssl-ca-alias 时,调度器将所有证书加载到信任库;指定别名时,只加载该别名。 |
| VER-72976 | Kafka Integration | The scheduler now supports CA bundles at the UDX and vkconfig level. 调度器现在支持在 UDX 和 vkconfig 级别使用 CA 捆绑包。 |
| VER-72997 | Cloud - Amazon, Security | Vertica instances can only access the S3 bucket specified for communal storage. Vertica 实例只能访问为公共存储指定的 S3 存储桶。 |
| VER-73011 | Backup/DR | The error message "Trying to delete untracked object" has been replaced with a more useful message providing guidance about backup metadata inconsistency. 错误消息 "Trying to delete untracked object" 已被更改为提供有关备份元数据不一致的指导信息。 |
9.3.1-11¶
Release Date: 6/9/2020
| Issue Key | Component | Description |
|---|---|---|
| VER-72839 | Catalog Engine, Execution Engine | Dropping a column no longer corrupts partition-level statistics, preventing subsequent ANALYZE_STATISTICS_PARTITION runs from failing. 删除列不再损坏分区级统计信息,从而防止后续 ANALYZE_STATISTICS_PARTITION 运行失败。 |
9.3.1-10¶
Release Date: 5/21/2020
| Issue Key | Component | Description |
|---|---|---|
| VER-72707 | DDL | Setting NULL on a column's DEFAULT expression is equivalent to no default; if DEFAULT/SET USING was already NULL, changing the column's data type removes the DEFAULT/SET USING expression. 在列的 DEFAULT 表达式上设置 NULL 等同于没有默认值;如果 DEFAULT/SET USING 已经为 NULL,更改列的数据类型会移除 DEFAULT/SET USING 表达式。 |
9.3.1-9¶
Release Date: 5/18/2020
| Issue Key | Component | Description |
|---|---|---|
| VER-72569 | DDL - Table | ALTER TABLE...RENAME could not rename multiple tables if projection names conflicted — resolved by maintaining local snapshots of renamed objects. 如果投影名称冲突,ALTER TABLE...RENAME 无法重命名多个表 — 通过维护重命名对象的本地快照解决。 |
| VER-72607 | Data Load / Copy | With EnableFastConstLoad ON, INSERT SELECT without specifying columns with default values preceding specified columns could cause Sort Order Violation — resolved. 在启用 EnableFastConstLoad 的情况下,INSERT SELECT 未指定在指定列之前具有默认值的列可能导致排序顺序违规 — 已解决。 |
| VER-72630 | DDL - Table | Concurrent CREATE TABLE...AS SELECT statements sometimes loaded duplicate data into the same table — resolved. 并发的 CREATE TABLE...AS SELECT 语句有时会向同一表加载重复数据 — 已解决。 |
| VER-72637 | Tuple Mover | Previously, TM resource pool always allocated two mergeout threads to inactive partitions regardless of MAXCONCURRENCY — now can allocate up to half of MAXCONCURRENCY threads to inactive partitions. 以前,TM 资源池总是为非活动分区分配两个 mergeout 线程,不考虑 MAXCONCURRENCY — 现在可以分配最多 MAXCONCURRENCY 一半的线程给非活动分区。 |
| VER-72710 | DDL | Creating/renaming a table in a user-defined schema with same name as an existing non-super unsegmented projection no longer displays an error. 在用户定义的 schema 中创建/重命名与现有非超级未分段投影同名的表不再显示错误。 |
9.3.1-8¶
Release Date: 5/7/2020
| Issue Key | Component | Description |
|---|---|---|
| VER-72106 | Optimizer | Optimizer required excessive time to generate plans for tables with complex live aggregate projections — resolved. 优化器为具有复杂实时聚合投影的表生成计划需要过多时间 — 已解决。 |
| VER-72255 | Execution Engine | Malformed queries on specific columns (constraint definitions with unrecognized values or improper transformations) caused database failure — resolved. 针对特定列的错误查询(包含无法识别的值或不正确转换的约束定义)导致数据库故障 — 已解决。 |
| VER-72346 | Optimizer | Queries mixing single-table predicates with EXISTS subqueries sometimes returned incorrect results — resolved. 混合单表谓词和 EXISTS 子查询的查询有时返回不正确的结果 — 已解决。 |
| VER-72352 | UI - Management Console | An exception occurred if a database with no password was used as production database for Extended Monitoring — resolved. 如果使用没有密码的数据库作为扩展监控的生产数据库,会发生异常 — 已解决。 |
| VER-72417 | Hadoop | Better error messages and cancelability for HDFS datanode failures during queries; previously hung during SASL handshaking. 改进了查询期间 HDFS 数据节点失败时的错误消息和可取消性;以前会在 SASL 握手期间挂起。 |
9.3.1-7¶
Release Date: 4/16/2020
| Issue Key | Component | Description |
|---|---|---|
| VER-72066 | Execution Engine | Loading Kafka messages larger than 5MB no longer causes insufficient memory error. 加载大于 5MB 的 Kafka 消息不再导致内存不足错误。 |
| VER-72107 | Hadoop | INFER_EXTERNAL_TABLE_DDL is now compatible with Parquet files using 2-level encoding for lists. INFER_EXTERNAL_TABLE_DDL 现在与使用二级列表编码的 Parquet 文件兼容。 |
| VER-72109 | DDL - Table | Columns with constraints containing ASCII and non-ASCII characters in names could cause database failure on insert — resolved. 名称中含有 ASCII 和非 ASCII 字符约束的列可能导致插入时数据库故障 — 已解决。 |
| VER-72112 | Kafka Integration, Supported Platforms | Vertica now supports rdkafka in RHEL 8.0, Ubuntu 19.10, and Debian 10. Vertica 现在支持 RHEL 8.0、Ubuntu 19.10 和 Debian 10 上的 rdkafka。 |
| VER-72113 | Data Removal - Delete, Purge, Partitioning | Clearer messages provided when drop_partitions fails due to insufficient resource pool. 当 drop_partitions 因资源池不足而失败时,提供更清晰的消息。 |
| VER-72115 | Hadoop | Vertica now retries HDFS operations that return 403 (Server too busy) instead of failing immediately. 对于返回 403(服务器太忙)的 HDFS 操作,Vertica 现在会重试而不是立即失败。 |
| VER-72117 | Optimizer, Recovery | Projection data now remains consistent following a MERGE during node recovery in Enterprise mode. 在 Enterprise 模式下节点恢复期间执行 MERGE 后,投影数据现在保持一致。 |
9.3.1-6¶
Release Date: 4/10/2020
Internal hotfix — no customer-facing issues.
9.3.1-5¶
Release Date: 4/2/2020
| Issue Key | Component | Description |
|---|---|---|
| VER-71833 | EON | The clean_communal_storage meta-function is now up to 200X faster. clean_communal_storage 元函数现在最多快 200 倍。 |
| VER-71866 | Subscriptions | In Eon Mode, primary and ETL session nodes now realign when subscription or node state changes. 在 Eon 模式下,当订阅或节点状态更改时,主节点和 ETL 会话节点现在会重新对齐。 |
9.3.1-4¶
Release Date: 3/13/2020
| Issue Key | Component | Description |
|---|---|---|
| VER-71622 | Backup/DR | Vbr no longer includes the dbadmin password in logging files when debug level is 3 and password is stored in a password file. 当调试级别为 3 且密码存储在密码文件中时,Vbr 不再在日志文件中包含 dbadmin 密码。 |
| VER-71688 | License | Vertica no longer fails if you audit a flextable created using SET USING. 如果审计使用 SET USING 创建的 flextable,Vertica 不再失败。 |
9.3.1-3¶
Release Date: 3/5/2020
| Issue Key | Component | Description |
|---|---|---|
| VER-71128 | Depot | DEPOT_FILES, DEPOT_EVICTIONS, and DEPOT_FETCHES now contain STORAGE_OID column with the storage container oid. DEPOT_FILES、DEPOT_EVICTIONS 和 DEPOT_FETCHES 现在包含 STORAGE_OID 列,提供存储容器 oid。 |
| VER-71139 | Optimizer | Moving CONSTRAINTS from ALTER TABLE to CREATE TABLE now only moves constraints of TEMP tables. 将 CONSTRAINTS 从 ALTER TABLE 移动到 CREATE TABLE 现在仅移动 TEMP 表的约束。 |
| VER-71362 | Admin Tools | Log rotate now functions properly in databases upgraded from 9.2.x to 9.3.x. 从 9.2.x 升级到 9.3.x 的数据库中,日志轮转现在正常工作。 |
| VER-71363 | Kafka Integration | rdkafka could create an infinite loop in rare cases — resolved. rdkafka 在极少数情况下可能创建无限循环 — 已解决。 |
| VER-71488 | Supported Platforms | vertica_agent.service can now stop gracefully on RHEL 7.7. vertica_agent.service 现在可以在 RHEL 7.7 上优雅停止。 |
9.3.1-2¶
Release Date: 2/20/2020
| Issue Key | Component | Description |
|---|---|---|
| VER-71253 | Admin Tools | Database Designer no longer produces an error when targeting a schema other than "public". 当目标 schema 不是 "public" 时,Database Designer 不再产生错误。 |
| VER-71067 | Third Party Tools Integration | NULL input to VoltageSecureProtect and VoltageSecureAccess now returns NULL value. 对 VoltageSecureProtect 和 VoltageSecureAccess 的 NULL 输入现在返回 NULL 值。 |
9.3.1-1¶
Release Date: 2/6/2020
| Issue Key | Component | Description |
|---|---|---|
| VER-70802 | Build and Release | The TZ environment variable now uses the latest time zone data from IANA. TZ 环境变量现在使用来自 IANA 的最新时区数据。 |
| VER-70962 | Third Party Tools Integration | VoltageSecureProtect now correctly sizes output when specifying custom Voltage number formats. VoltageSecureProtect 在指定自定义 Voltage 数字格式时现在正确计算输出大小。 |
| VER-71003 | Catalog Engine | In Eon Mode, projection checkpoint epochs on down nodes become consistent with the current checkpoint epoch when nodes resume. 在 Eon 模式下,当节点恢复时,故障节点上的投影检查点时期与当前检查点时期保持一致。 |
9.3.1-0¶
Release Date: 1/14/2020
| Issue Key | Component | Description |
|---|---|---|
| VER-42720 | Execution Engine | Vertica occasionally generated misleading error messages in certain distributed query executions — resolved. Vertica 在某些分布式查询执行中偶尔生成误导性错误消息 — 已解决。 |
| VER-63714 | Cloud - Google | Exporting parquet files to Google Cloud Storage caused an error — fixed. 将 parquet 文件导出到 Google Cloud Storage 时导致错误 — 已修复。 |
| VER-66216 | Admin Tools | Passwords are now validated when creating databases. 创建数据库时现在会验证密码。 |
| VER-66903 | Kafka Integration | Users can now use kafka_conf to override the Kafka SSL setting to allow SASL+SSL authentication. 用户现在可以使用 kafka_conf 覆盖 Kafka SSL 设置以允许 SASL+SSL 认证。 |
| VER-67029 | Client Drivers - JDBC | The maxconnection limitation check no longer fails after running a distributed query. 运行分布式查询后,最大连接限制检查不再失败。 |
| VER-67201 | Client Drivers - ODBC | Non-ASCII characters in filenames for copy local operations (and rejection/exception files) now work correctly in Windows ODBC/OLEDB drivers. 复制本地操作文件名中的非 ASCII 字符(以及拒绝/异常文件)现在在 Windows ODBC/OLEDB 驱动程序中正常工作。 |
| VER-67497 | Client Drivers - VSQL | VSQL options -B (backup server/port) and -k/-K (Kerberos) are now compatible and can be set together. VSQL 选项 -B(备份服务器/端口)和 -k/-K(Kerberos)现在兼容,可以同时设置。 |
| VER-67511 | Tuple Mover | Tuple Mover threads spawned by DML operations are now subject to MergeOutInterval, preventing excessive checks for storage location move requests. DML 操作生成的 Tuple Mover 线程现在受 MergeOutInterval 约束,防止对存储位置移动请求进行过多检查。 |
| VER-67701 | LocalPlanner | Vertica occasionally failed running COUNT DISTINCT queries with predicates on lead columns in projection sort order — resolved. Vertica 偶尔在投影排序顺序的前导列上带有谓词的 COUNT DISTINCT 查询中失败 — 已解决。 |
| VER-67964 | Execution Engine | In rare circumstances with short-lived partitioning due to network problems, the whole cluster could go down — resolved. 在因网络问题导致短暂分区的极少数情况下,整个集群可能宕机 — 已解决。 |
| VER-67966 | Backup/DR | Concurrent copycluster and swap partition tasks could cause data participating in swap partition to go missing on target — resolved. 并发的 copycluster 和 swap partition 任务可能导致参与交换分区的数据在目标上丢失 — 已解决。 |
| VER-67980 | Client Drivers - ODBC | Error messages containing Japanese column names now display correctly. 包含日文列名的错误消息现在正确显示。 |
| VER-68414 | SAL | Queries could fail when reading ROS file format from Vertica 6.0 or earlier — now properly handled. 从 Vertica 6.0 或更早版本读取 ROS 文件格式时,查询可能失败 — 现已正确处理。 |
| VER-68549 | UI - Management Console | LDAP anonymous binding setting in MC did not take effect despite clicking Apply and Done — fixed. MC 中的 LDAP 匿名绑定设置尽管点击了 Apply 和 Done 却不生效 — 已修复。 |
| VER-68633 | Security | SESSIONS system tables are now populated with the username value from the catalog for case consistency. SESSIONS 系统表现在使用目录中的用户名值填充,以保证大小写一致性。 |
| VER-68644 | Tuple Mover | Vertica now ignores unsafe projections when determining the AHM. 在确定 AHM 时,Vertica 现在忽略不安全的投影。 |
| VER-68720 | UDX | GRANT and REVOKE now support ALTER and DROP privileges for user-defined functions. GRANT 和 REVOKE 现在支持用户自定义函数的 ALTER 和 DROP 权限。 |
| VER-68726 | Kafka Integration | Vertica no longer fails when encountering improper Kafka SSL key/certificate setup error information on RedHat Linux. 在 RedHat Linux 上遇到不正确的 Kafka SSL 密钥/证书设置错误信息时,Vertica 不再失败。 |
| VER-68790 | Backup/DR | Deleting an object during a backup on Eon mode could cause backup failure — resolved. 在 Eon 模式下备份期间删除对象可能导致备份失败 — 已解决。 |
| VER-68853 | Subscriptions | A failure in one Eon Mode subcluster no longer affects another subcluster. 一个 Eon 模式子集群的故障不再影响另一个子集群。 |
| VER-68898 | Data Removal - Delete, Purge, Partitioning | DROP_PARTITIONS error message improved when force-split fails due to concurrent mergeout — now provides hint to try again. 当强制拆分因并发 mergeout 失败时,DROP_PARTITIONS 的错误消息已改进 — 现在提示重试。 |
| VER-69184 | Execution Engine | COUNT(DISTINCT) with aggregates like AVG on numeric datatype caused memory conflict when input size > output size — resolved. COUNT(DISTINCT) 与 AVG 等聚合函数在数值数据类型上使用时,当输入大小 > 输出大小时导致内存冲突 — 已解决。 |
| VER-69217 | Data Load / COPY | Excessively long inputs to a NUMERIC column in COPY could crash Vertica — resolved. 对 COPY 中 NUMERIC 列的过长输入可能导致 Vertica 崩溃 — 已解决。 |
| VER-69229 | Execution Engine | Queries failed to sort UUIDs correctly if ORDER BY did not specify UUID column first — resolved. 如果 ORDER BY 未首先指定 UUID 列,查询无法正确排序 UUID — 已解决。 |
| VER-69292 | Execution Engine, UDX, Vertica Log Text Search | v_txtindex.StringTokenizerDelim no longer makes copies of input for each token, reducing memory issues. v_txtindex.StringTokenizerDelim 不再为每个标记复制输入,减少了内存问题。 |
| VER-69383 | Optimizer - Statistics and Histogram | ANALYZE_ROW_COUNT can no longer change STATISTICS_TYPE from FULL to ROWCOUNT. ANALYZE_ROW_COUNT 不能再将 STATISTICS_TYPE 从 FULL 更改为 ROWCOUNT。 |
| VER-69431 | Data Export | S3EXPORT now supports a null_as parameter to specify how null values are exported. S3EXPORT 现在支持 null_as 参数来指定如何导出空值。 |
| VER-69486 | Security | LDAP Link now only requires a global catalog lock when mapping LDAP entries to Vertica users/roles, not the entire operation. LDAP Link 仅在将 LDAP 条目映射到 Vertica 用户/角色时需要全局目录锁,而不是整个操作。 |
| VER-69489 | Data Load / COPY | COPY statements that auto-generated rejected data filenames too long for the filesystem now truncate them. 自动生成的文件名对于文件系统过长的拒绝数据文件名,COPY 语句现在会截断它们。 |
| VER-69535 | Data Load / COPY | A bug in tracking disk usage of rejected data tables could rarely crash Vertica — resolved. 跟踪拒绝数据表磁盘使用量的一个 bug 在极少数情况下可能导致 Vertica 崩溃 — 已解决。 |
| VER-69606 | Catalog Engine | ETL queries no longer fail with "Active subscriptions changed during query planning" when REBALANCE_SHARDS runs on another subcluster. 当 REBALANCE_SHARDS 在另一个子集群上运行时,ETL 查询不再因 "Active subscriptions changed during query planning" 而失败。 |
| VER-69702 | Communications/messaging | Queries against virtual tables no longer display duplicate error messages when one or more nodes is down. 当一个或多个节点宕机时,对虚拟表的查询不再显示重复的错误消息。 |
| VER-70487 | Admin Tools | Starting a subcluster in Eon Mode no longer writes duplicate entries to admintools.log. 在 Eon 模式下启动子集群不再向 admintools.log 写入重复条目。 |
9.3.0-2¶
Release Date: 12/18/2019
| Issue Key | Component | Description |
|---|---|---|
| VER-69250 | UDX | GRANT and REVOKE now support ALTER and DROP privileges for user-defined functions. GRANT 和 REVOKE 现在支持用户自定义函数的 ALTER 和 DROP 权限。 |
| VER-69666 | Backup/DR | Concurrent copycluster and swap partition tasks could cause data going missing on target cluster — resolved. 并发的 copycluster 和 swap partition 任务可能导致目标集群上的数据丢失 — 已解决。 |
| VER-69783 | Execution Engine | COUNT(DISTINCT) with AVG on numeric datatype caused memory conflict when input size > output size — resolved. COUNT(DISTINCT) 与 AVG 在数值数据类型上使用时,输入大小 > 输出大小时导致内存冲突 — 已解决。 |
| VER-69787 | Cloud - Google | Exporting parquet files to Google Cloud Storage caused an error — fixed. 将 parquet 文件导出到 Google Cloud Storage 时导致错误 — 已修复。 |
| VER-70085 | Backup/DR | Deleting an object during an Eon mode backup could cause backup failure — resolved. 在 Eon 模式备份期间删除对象可能导致备份失败 — 已解决。 |
| VER-70144 | Communications/messaging | Queries against virtual tables no longer show duplicate error messages when nodes are down. 当节点宕机时,对虚拟表的查询不再显示重复的错误消息。 |
| VER-70148 | Client Drivers - VSQL | VSQL options -B and -k/-K are now compatible and can be set together. VSQL 选项 -B 和 -k/-K 现在兼容,可以同时设置。 |
| VER-70150 | Catalog Engine | ETL queries no longer fail with "Active subscriptions changed during query planning" when running REBALANCE_SHARDS on another subcluster. 当在另一个子集群上运行 REBALANCE_SHARDS 时,ETL 查询不再因 "Active subscriptions changed during query planning" 而失败。 |
| VER-70151 | Subscriptions | A failure in one Eon Mode subcluster no longer affects another subcluster. 一个 Eon 模式子集群的故障不再影响另一个子集群。 |
| VER-70235 | Data Removal - Delete, Purge, Partitioning | DROP_PARTITIONS error message improved when concurrent mergeout prevents force-split. 当并发 mergeout 阻止强制拆分时,DROP_PARTITIONS 的错误消息已改进。 |
| VER-70460 | AP-Advanced | Certain User-Defined Aggregate functions could cause "Data Area overflow" error in rare situations — resolved. 某些用户自定义聚合函数在极少数情况下可能导致 "Data Area overflow" 错误 — 已解决。 |
| VER-70466 | Admin Tools | Starting a subcluster in Eon Mode no longer writes duplicate entries to admintools.log. 在 Eon 模式下启动子集群不再向 admintools.log 写入重复条目。 |
9.3.0-1¶
Release Date: 10/31/2019
| Issue Key | Component | Description |
|---|---|---|
| VER-69388 | SAL | Queries could fail when Vertica 9.2 tried to read ROS file format from Vertica 6.0 or earlier — now properly handled. 当 Vertica 9.2 尝试从 Vertica 6.0 或更早版本读取 ROS 文件格式时,查询可能失败 — 现已正确处理。 |
9.3.0-0¶
Release Date: 10/14/2019
| Issue Key | Component | Description |
|---|---|---|
| VER-67087 | Admin Tools | During database revive, admintools treated s3 and hdfs user storage locations as local filesystem paths — resolved. 在数据库恢复期间,admintools 将 s3 和 hdfs 用户存储位置视为本地文件系统路径 — 已解决。 |
| VER-68531 | Admin Tools | Environment variables for admintools SSH operations were set incorrectly on remote hosts — resolved. admintools SSH 操作的环境变量在远程主机上设置不正确 — 已解决。 |
| VER-64171 | Backup/DR | If hardlink failed during backup, vbr switched to copying data instead of failing — resolved. 如果备份期间硬链接失败,vbr 会转换为复制数据而不是失败 — 已解决。 |
| VER-66956 | Backup/DR | Dropping a user that owns an object during replicate/restore could cause nodes to fail. 在复制/还原期间删除拥有对象的用户可能导致节点故障。 |
| VER-62334 | Catalog Engine | DROP statement rollback now generates at most 10 dependency messages, then shows total count instead of flooding logs. DROP 语句回滚现在最多生成 10 条依赖消息,然后显示总计数,而不是淹没日志。 |
| VER-67734 | Catalog Engine | In Eon mode, queries sometimes failed if submitted at the same time a new node was added — resolved. 在 Eon 模式下,如果在添加新节点的同时提交查询,有时会失败 — 已解决。 |
| VER-68188 | Catalog Engine | Exporting a catalog on a view that references itself could cause Vertica to fail — resolved. 在引用自身的视图上导出目录可能导致 Vertica 失败 — 已解决。 |
| VER-68603 | Catalog Sync and Revive | Database startup occasionally failed when catalog truncation failed due to full disk — resolved. 当因磁盘已满导致目录截断失败时,数据库启动偶尔会失败 — 已解决。 |
| VER-68494 | Client Drivers - Misc | Node failures now update the available node list for connection load balancing. 节点故障现在会更新连接负载均衡的可用节点列表。 |
| VER-67342 | Data Removal - Delete, Purge, Partitioning | SWAP_PARTITIONS_BETWEEN_TABLES made atomic to fix empty results when swap committed between query launch and local planning. SWAP_PARTITIONS_BETWEEN_TABLES 已改为原子操作,修复了在查询启动和本地计划之间提交交换时返回空结果的问题。 |
| VER-66546 | Cloud - Amazon | verticad script on Amazon Linux sometimes could not detect the operating system — resolved. Amazon Linux 上的 verticad 脚本有时无法检测操作系统 — 已解决。 |
| VER-53370 | DDL - Projection | Renaming a table now also renames projections prefixed by the anchor table name. 重命名数据库表也会重命名以锚表名为前缀的投影。 |
| VER-66882 | DDL - Table | SELECT ANALYZE_STATISTICS no longer acquires a GCL-X lock when running against local temp tables. 对本地临时表运行时,SELECT ANALYZE_STATISTICS 不再获取 GCL-X 锁。 |
| VER-67105 | S3 Data Export | s3export now supports CSV format via new enclosed_by and escape_as parameters. s3export 现在通过新的 enclosed_by 和 escape_as 参数支持 CSV 格式。 |
| VER-68619 | Hadoop Data Export | Date columns now include Parquet Logical type so other tools recognize them as Date. 日期列现在包含 Parquet 逻辑类型,以便其他工具将其识别为日期。 |
| VER-66272 | Data Removal - Delete, Purge, Partitioning | SWAP_PARTITIONS_BETWEEN_TABLES made atomic to fix empty results (same as VER-67342). SWAP_PARTITIONS_BETWEEN_TABLES 已改为原子操作(与 VER-67342 相同)。 |
| VER-65427 | Data Load / COPY | Vertica could crash if client disconnected during COPY LOCAL with REJECTED DATA — resolved. 如果在使用 REJECTED DATA 的 COPY LOCAL 期间客户端断开连接,Vertica 可能崩溃 — 已解决。 |
| VER-65659 | Data Load / COPY | Copy or external table query could crash a node — resolved. 复制或外部表查询可能导致节点崩溃 — 已解决。 |
| VER-68033 | Data Load / COPY | A bug in Apache Parquet C++ library could cause Vertica to fail when reading Parquet files with large Varchar statistics — resolved. Apache Parquet C++ 库中的一个 bug 可能导致 Vertica 在读取包含大 Varchar 统计信息的 Parquet 文件时失败 — 已解决。 |
| VER-53981 | Database Designer Core | DESIGNER_DESIGN_PROJECTION_ENCODINGS mistakenly removed comments from target projections — resolved. DESIGNER_DESIGN_PROJECTION_ENCODINGS 错误地从目标投影中删除了注释 — 已解决。 |
| VER-62628 | Execution Engine | Subquery error message now provides Join label and localplan_id to help identify which subquery returned multiple rows. 子查询错误消息现在提供连接标签和 localplan_id,以帮助识别哪个子查询返回了多行。 |
| VER-66224 | Execution Engine | Queries exceeding runtime cap auto-restarted instead of reporting timeout error — resolved. 超过运行时间上限的查询自动重启而不是报告超时错误 — 已解决。 |
| VER-67069 | Execution Engine | Queries with complex predicates ignored cancel attempts and runtime caps — now cancelable. 具有复杂谓词的查询忽略了取消尝试和运行时间上限 — 现在可以取消。 |
| VER-67102 | Execution Engine | Users were unable to cancel meta-function RUN_INDEX_TOOL — resolved. 用户无法取消元函数 RUN_INDEX_TOOL — 已解决。 |
| VER-69066 | Execution Engine | Queries failed to sort UUIDs correctly if ORDER BY didn't specify UUID column first — resolved. 如果 ORDER BY 未首先指定 UUID 列,查询无法正确排序 UUID — 已解决。 |
| VER-64680 | Kafka Integration | kafkacat utility bundled with Vertica can now make TLS/SSL encrypted connections. Vertica 捆绑的 kafkacat 工具现在可以进行 TLS/SSL 加密连接。 |
| VER-68192 | License | Upgrading 8.1.1-x databases with Autopass license to 9.0+ could cause license tampering startup issues — resolved. 将具有 Autopass 许可证的 8.1.1-x 数据库升级到 9.0+ 可能导致许可证篡改启动问题 — 已解决。 |
| VER-67573 | Nimbus | DDL transactions now complete while file deletions continue in the background, preventing delays in Eon mode. DDL 事务现在在后台继续文件删除时完成,防止 Eon 模式下的延迟。 |
| VER-26260 | Optimizer | Vertica can now optimize queries on system tables with guaranteed unique values by pruning unnecessary joins. Vertica 现在可以通过修剪不必要的连接来优化对具有保证唯一值的系统表的查询。 |
| VER-66423 | Optimizer | The Optimizer is better able to derive transitive selection predicates from subqueries. 优化器现在能够更好地从子查询推导传递选择谓词。 |
| VER-66902 | Optimizer | EXPORT TO VERTICA returned error if exporting a flattened table that existed in both source and target — resolved. 如果导出在源和目标中都存在的扁平化表,EXPORT TO VERTICA 返回错误 — 已解决。 |
| VER-66933 | Optimizer | Export operations now always export constraints (except foreign keys) as part of the CREATE statement, fixing issues with temporary tables. 导出操作现在始终将约束(外键除外)作为 CREATE 语句的一部分导出,修复了临时表的问题。 |
| VER-66968 | Optimizer | Queries performing inner join with grouping now return consistent results. 执行带有分组的内连接的查询现在返回一致的结果。 |
| VER-67138 | Optimizer | Vertica could move subqueries to the ON clause when flattening, which is not supported — resolved. Vertica 在展平时可能将子查询移动到 ON 子句,这是不支持的 — 已解决。 |
| VER-67443 | Optimizer | ALTER TABLE...ALTER CONSTRAINT can now be used when nodes are down. ALTER TABLE...ALTER CONSTRAINT 现在可以在节点宕机时使用。 |
| VER-67740 | Optimizer | Vertica sometimes crashed when certain analytic function combinations applied to merge join output — resolved. 当某些分析函数组合应用于 merge join 输出时,Vertica 有时会崩溃 — 已解决。 |
| VER-67908 | Optimizer | Prepared statements now throw a warning instead of an error for WITH clause materialization and process without materializing. 对于 WITH 子句物化,预处理语句现在抛出警告而不是错误,并在不物化的情况下处理。 |
| VER-68306 | Optimizer | Queries with multiple analytic functions over complex expressions and different PARTITION BY/ORDER BY sets produced inconsistent results — resolved. 具有多个分析函数、复杂表达式和不同 PARTITION BY/ORDER BY 集的查询产生不一致的结果 — 已解决。 |
| VER-68379 | Optimizer | REFRESH_COLUMNS calls can now be monitored via the new "dc_refresh_columns" table. REFRESH_COLUMNS 调用现在可以通过新的 "dc_refresh_columns" 表进行监控。 |
| VER-68594 | Optimizer | Queries sorting on subquery outputs sometimes produced inconsistent results — resolved. 对子查询输出进行排序的查询有时产生不一致的结果 — 已解决。 |
| VER-68828 | Optimizer | MERGE statements in serializable transaction isolation mode no longer return 'Can't run historical queries at epochs prior to the Ancient History Mark' error. 可序列化事务隔离模式下的 MERGE 语句不再返回 'Can't run historical queries at epochs prior to the Ancient History Mark' 错误。 |
| VER-49136 | Optimizer | TRUNCATE TABLE caused all table- and partition-level statistics to be dropped — resolved. TRUNCATE TABLE 导致所有表和分区级统计信息被删除 — 已解决。 |
| VER-68826 | Optimizer - Statistics and Histogram | EXPORT_STATISTICS_PARTITION wrote empty content to target file for date-partitioned tables — resolved. 对于日期分区表,EXPORT_STATISTICS_PARTITION 向目标文件写入空内容 — 已解决。 |
| VER-67647 | Build and Release Performance tests | ST_IsValid() with geography data type saw ~10% performance degradation. 带地理数据类型的 ST_IsValid() 性能下降了约 10%。 |
| VER-66648 | QA | "ERROR 8617: Request size too big" message incorrectly suggested excessive memory consumption — resolved. "ERROR 8617: Request size too big" 消息错误地暗示了过度的内存消耗 — 已解决。 |
| VER-64783 | Recovery | Vertica now considers recovery complete when all regular projections are recovered, disregarding live aggregate projection recovery status. 当所有常规投影都恢复时,Vertica 现在认为恢复完成,忽略实时聚合投影的恢复状态。 |
| VER-68504 | Execution Engine S3 | Refreshing a large projection in Eon mode using temp space on S3 could crash the node — resolved. 在 Eon 模式下使用 S3 上的临时空间刷新大型投影可能导致节点崩溃 — 已解决。 |
| VER-60036 | SDK | All C++ UDXs now require c++11 standard (via -std=c++11 flag at compilation). 所有 C++ UDX 现在需要 c++11 标准(通过在编译时使用 -std=c++11 标志)。 |
| VER-61780 | Scrutinize | Scrutinize previously generated UnicodeEncodeError if system locale used non-ASCII characters — resolved. 如果系统区域设置使用非 ASCII 字符,Scrutinize 以前会生成 UnicodeEncodeError — 已解决。 |
| VER-66988 | Scrutinize | The db_password entry has been removed from scrutinize_collection.log. db_password 条目已从 scrutinize_collection.log 中移除。 |
| VER-62276 | Security | Changes to SSLCertificate, SSLPrivateKey, and SSLCA parameters take effect for all new connections without requiring a restart. 对 SSLCertificate、SSLPrivateKey 和 SSLCA 参数的更改无需重启即可对所有新连接生效。 |
| VER-65168 | Security | Default permissions on Database Designer scripts reverted to 666 (rw-rw-rw-). Database Designer 脚本的默认权限恢复为 666(rw-rw-rw-)。 |
| VER-65890 | Security | Two new system tables added: inheriting_objects and inherited_privileges for monitoring inherited privileges. 新增两个系统表:inheriting_objects 和 inherited_privileges,用于监控继承的权限。 |
| VER-68616 | Security | Queries by non-owners with inherited privileges on views now correctly ensure the view owner has SELECT WITH GRANT OPTION on anchor relations. 具有继承权限的非所有者对视图的查询现在正确确保视图所有者对锚关系具有 SELECT WITH GRANT OPTION。 |
| VER-67658 | Spread | In unstable networks, lost UDP-based control messages could cause hanging uncancellable sessions — resolved. 在不稳定的网络中,丢失的基于 UDP 的控制消息可能导致挂起且不可取消的会话 — 已解决。 |
| VER-47639 | Supported Platforms | Vertica now supports the XFS file storage system. Vertica 现在支持 XFS 文件存储系统。 |
| VER-67234 | Tuple Mover | Mergeout prioritization algorithm changed to prevent slow-loading jobs from being overlooked, avoiding ROS pushback. 更改了 Mergeout 优先级算法,防止慢速加载作业被忽略,避免 ROS 反压。 |
| VER-67275 | Tuple Mover | The active partition mergeout thread now scans the entire queue for eligible jobs instead of stopping at a non-eligible job. 活动分区 mergeout 线程现在扫描整个队列以查找符合条件的作业,而不是在遇到不符合条件的作业时停止。 |
| VER-68383 | Tuple Mover | Mergeout no longer generates and queues purge requests with invalid partition keys, preventing rapid log growth. Mergeout 不再生成和排队具有无效分区键的清理请求,防止日志快速增长。 |
| VER-65744 | UDX | Flex table views now properly show UTF-8 encoded multi-byte characters. Flex 表视图现在正确显示 UTF-8 编码的多字节字符。 |
| VER-62048 | UI - Management Console | Design steps in MC Events history tab appeared in wrong order with incorrect timestamps — resolved. MC 事件历史标签中的设计步骤以错误的顺序出现,时间戳不正确 — 已解决。 |
| VER-65561 | UI - Management Console | Password could appear in the scrutinize collection log under some MC conditions — resolved. 在某些 MC 条件下,密码可能出现在 scrutinize 收集日志中 — 已解决。 |
| VER-67903 | UI - Management Console | MC was not correctly handling design tables containing Japanese characters — resolved. MC 未正确处理包含日文字符的设计表 — 已解决。 |
| VER-68357 | UI - Management Console | MC now correctly updates Clockskew alerts with the timestamp when first detected, and displays resolved/unresolved status. MC 现在用首次检测到的时间戳正确更新时钟偏差警报,并显示已解决/未解决状态。 |
| VER-68448 | UI - Management Console | MC now generates alerts for the CRC Mismatch SNMP trap. MC 现在为 CRC 不匹配 SNMP 陷阱生成警报。 |