7.9 Maintenance Release Changelog (Cloud Only)

Fixed some error handling issues with unlimited storage download and upload processes.

The SHOW TABLE STATUS command now displays the memory usage by GLOBAL TEMPORARY tables.

Fixed an issue that may cause the old master aggregator to lose users/grants metdata when it rejoins the cluster after a new master aggregator was promoted using the AGGREGATOR SET AS MASTER command.

Fixed a crash when parsing certain Parquet data into a pipeline.

Added support for using a connection link for inserting data with the FORMAT PARQUET option.

Fixed an issue with aggregate functions using incorrect enum/set types that may result in inaccurate output in the operator tree.

Fixed an issue with a transaction leak on the master aggregator when running CREATE TABLE AS SELECT on a child aggregator using autocommit=0 on the connection.

Fixed a bug that may cause a query to hang when comparing an utf8 column with an utf8mb4 constant. This issue occurs when collation_server is set to one of the utf8mb4 collations.

Improved the accuracy of network time reporting in query profiles regarding the time spent sending the results back to the user connection.

Fixed an edge case issue causing a potential memory leak when running an UPSERT statement against a columnstore table.

Fixed an issue that could cause the engine to crash from info_schema query submissions.

Improved the performance of bushy join rewrites.

Fixed an edge case issue where the engine could crash when performing multi-inserts.

The avro schema registry URL portion of the CREATE PIPELINE syntax is now redacted in processlist.

Fixed an issue where the engine could crash during recursive set operations.

The information_schema.statistics "collation" column now correctly indicates whether an index is ascending ("A") or descending ("D").

Improved performance of comparing utf8mb4 strings.

Fixed an edge case issue which could cause the engine to hang during shutdown.

Added the skip_segelim_with_inlist_threshold engine variable, which will skip segment elimination with the IN list if its size is larger than threshold (default is 1000 elements).

Fixed a performance issue when creating materialized aggregator result tables.

Fixed an issue when auto-creation is turned off (auto.create.topics.enable=false); SELECT INTO KAFKA returns a success message even if the insert fails.

Fixed an issue with tables created on child aggregators not using the correct create_user in the information_schema.tables.

Improved the ALTER TABLE error message received when run on a view on a child aggregator.

Added the session variable histogram_column_correlation, which controls the assumed correlation between columns when using histograms for selectivity estimation.

Added ability to use ALTER and TRUNCATE table commands when BACKUP is running.

Fixed an issue where wrong results may occur when a shard key column is set to the value of a non-deterministic function (RAND(), UUID(), etc.).

SingleStoreDB Cloud no longer schedules pipelines to run internally if there is no writable database attached.

Improved join estimations when using a unique key.

Fixed an incorrect error message when GRANTing table-level permission if the default_user_require_ssl is set.

Improved incremental backup performance.

Fixed a potential issue that could cause a NFS backup error to trigger a crash.

Fixed an issue where having too many rows on the same columnstore hash key could return "Read failed with error 14 (Bad address)".

Added the ability to handle 429 (too many requests) errors the same way as 503 (slow down) errors.

Enabled histograms to estimate predicates that use Stored Procedure parameters or implicit/explicit typecasted literals. This change is controlled via the enable_histogram_to_unwrap_literals sync variable. This is Off by default. It can be enabled by setting the variable to ‘On’ or 1.

Added the ability to limit the application of join predicate transitivity to predicates that may be estimated. To use this feature, the session variable verify_fields_in_transitivity needs to be set to ON.

Fixed an issue that may cause a restarted master aggregator to crash when the STOP PIPELINE command runs on a detached pipeline.

Fixed a potential issue that could cause a lost second prepare statement when using a binary protocol over SSL.

Fixed an issue where an incorrect result could be returned when a query contains multiple joins that all match a multi-column hash index.

Fixed a performance issue with count distinct aggregation and streaming subqueries.

Reduced the memory usage of querying information_schema.pipeline_errors when there are many past pipeline errors.

Decreased the cost of working with large project lists during optimization.

ATTACH DATABASE...READ ONLY no longer blocks the DDL or code-generation against the primary writable workspace when the read-only workspace is being ATTACHED. This also applies to DETACH of a read-only site.

Introduced new syntax for DROP FROM PLANCACHE to drop a plan based on query text instead of plan_id. The syntax also supports dropping plans across all aggregators instead of on a particular node_id.

Introduced a new engine variable: clamp_histogram_date_estimates. When enabled, SingleStoreDB will use a heuristic to more accurately estimate date predicates that are larger than the histograms range. This variable is set to AUTO (OFF) by default.

Added the ability to allow code-generation of queries against the information schema when the cluster is very low on disk/memory.

Fixed a potential issue when a database is ATTACH ... READ ONLY is re-attached to the cluster for the first time. Auto-increment columns metadata was not resyncing properly, causing false, duplicate key errors.

Fixed a potential issue that may cause an infinite loop when some information_schema tables are filtered using the utf8m4 string constant.

Fixed a potential issue using multi-statement batches inside of prepared statements without parameters that may result in a crash.

Fixed a potential issue that may cause unlimited databases to rebuild unnecessary auto-stats.

Fixed an issue which caused unlimited database S3 uploads to fail due to an ExpiredToken error. This issue impacted the refresh of EC2 instance meta credentials caused by the public open-source AWS C++ SDK after encountering a transient network issue.

Allow REVOKE from ROLE permissions for tables that have been dropped.

Fixed a specific query shape which failed CodeGen for columnstore unique keys.

Added engine variable: exporter_ssl_key_passphrase. It is used for the decryption of the exporter_ssl_key file by the monitoring process if the file is encrypted (password protected).

Fixed an issue where memsql_exporter could start with an incorrect user.

Fixed a potential issue regarding diagnostic messages appearing to start successfully but were silently failing.

Improved the hash function used for strings.

Updated behavior of the DROP FROM PLANCACHE command; it now supports ON NODE syntax to drop a query plan from a specific node.

Fixed an inconsistent error message on CREATE USER when the target user already exists.

Added ability to assign a custom name to a Wasm UDF in CREATE FUNCTION by specifying the USING EXPORT clause.

Fixed an incorrect error message for EXPLAIN REBALANCE ALL DATABASES when a read-only database was attached to a workspace.

Added ability for information_schema.mv_plancache to be workspace aware. Only displays plans executed on the connected user's workspaces.

Added support for SELECT INTO ... FORMAT PARQUET.

Added ability to show CTE (common table expressions) in SHOW CREATE VIEW.

Fixed a memory leak issue due to alloc_warnings when using reference table pipelines.

Fixed an issue where commands forwarded to the master aggregator during an online upgrade could cause a crash.

Fixed an invalid op tree error when CCM matching is enabled on a query with CTEs.

Fixed a performance issue for sampling queries with large IN lists executed by the query optimizer.

Fixed an issue with the aggregate result tables not being parameterized correctly, which resulted in unnecessary plans being generated.

Increased the default value of the following engine variables: http_api_pool_max_idle to 1024 and http_api_pool_capacity to 2048.

Added http_api_max_lifetime_seconds engine variable which controls how long a pooled connection can remain open.

Added http_api_max_idle_seconds engine variable which controls how long a pooled connection can remain idle (or unused) before it will be closed.

Improved snapshot trigger for databases with a slow rate of writes. A snapshot occurs when at least 128MB of logs have been written in the last hour.

The CREATE DATABASE command now performs a disk space availability check to ensure enough space exists to store some initial logs and blobs for the database. If the disk space is too low, the create command will be rejected.

Added ability for automatically triggered rebalance operations to restart any long-running write queries which would prevent the rebalance from making progress. The user executed rebalance commands now have a KILL option to trigger this behavior from a manually executed REBALANCE.

Added support for running INSERT SELECTS using an aggregator result table on a child aggregator.

Fixed an issue where canceling a RESTORE would not completely drop the partially restored database.

Fixed a lockdown feature error that may occur for queries with multiple in-subselects.

Fixed an issue where SHOW GRANTS output was incorrect for users with REQUIRE SSL WITH GRANT OPTIONS.

Improved reliability of full and incremental cloud backups that experience network connectivity issues.

Improved the accuracy of the FIRST_LOADED and LAST_EXECUTED timestamp columns in information_schema.plancache.

Fixed an issue where S3 can silently fail to upload a file for an unlimited database.

Fixed an issue that may cause error 200 to occur when running CREATE DATABASE or RESTORE DATABASE commands on an unlimited database.

Fixed an edge case issue where the blob cache eviction algorithm could cause an infinite loop.

Added the ability to query MV_BACKUP_STATUS for users with SHOW METADATA permissions.

Fixed a crash when creating a WASM user-defined function from an S3 bucket with an invalid prefix in the path name.

Fixed an incorrect error message when a computed column refers to a two part table name (i.e., table.column).

Fixed an issue that may cause insert into mv_queries to become invalid when the cluster is in ONLY_FULL_GROUP_BY sql_mode.

Fixed an issue that returned 'zero rows' when sampling fails or there is no sampling data. The corrected behavior will use heuristic or histogram estimate depending on what is available. To enable this behavior, set estimate_zero_rows_when_sampling_data_is_missing to ‘false’.

Fixed parsing of microseconds in STR_TO_DATE to give fewer than six digits a higher priority.

Fixed an issue where queries that contain CTE (common table expressions) with the same name may not traverse correctly.

Fixed an issue where S3 can silently fail to upload a file for an unlimited database.

Added the ability to configure SingleStoreDB to estimate the equality predicate for uncorrelated scalar subselects as 1/cardinality; and use heuristics for all other scalar subselect predicates. To enable the new estimation logic, set exclude_scalar_subselects_from_filters to 'false'.

Improved error messages generated when a PITR cannot find a valid time range for restoring.

Added support for Negative Lookahead, Positive Lookahead, Non-greedy quantifiers, and Non-capturing group regular expressions.

Fixed an issue where large amounts of consecutive aborted messages were not being handled correctly in a Kafka pipeline.

Improved the Subselect To Join rewrite to recognize more cases with subselects in filter conditions for left join.

Fixed an issue with window functions which refer to the window definition at the end of a SELECT statement.

Removed ability to use utf8_general_mysql500_ci collation. This collation is not supported by SingleStoreDB Cloud.

Added support for a new AES encryption mode called CBC (cipher block chaining) for 128-bit and 256-bit keys. See AES_ENCRYPT for more information.

Fixed an issue with unnecessary data sorting when single partition queries contain an ORDER BY clause. The sort is redundant because the data is sorted on the leaves for these types of queries.

Fixed a crash caused by UPDATE statements with a joined table where the optimizer pushes the JOIN into a derived table whose name is automatically generated and different from the original table being updated.

Added support for s3_force_path_style to S3 backup. s3_force_path_style (optional boolean, defaults to true) specifies whether to use path style (region.amazonaws.com/bucket) or virtual address style (bucket.region.amazonaws.com) syntax when specifying the location of the bucket. For S3 compatible services that only support virtual address paths, specify "s3_force_path_style":"false" in the configuration options. Refer to the BACKUP DATABASE section for more information.

Fixed an error regarding failover in progress that occurs when running EXPLAIN REBALANCE PARTITIONS FULL.

Fixed an issue where auto-attach of leaf nodes was not being retried on repeatable errors due to excessive errors when checking on the state of the leaf partitions.

Fixed an issue where DETACH DATABASE of a read-only replica was not correctly removing the partitions from offline leaves.

Fixed an issue where CLEAR ORPHAN commands were not correctly removing the multiple replica partitions for databases attached as read replicas.

Fixed an issue where ATTACH DATABASE would attach a replica partition for an existing online read replica partition.

Fixed an issue with REBALANCE PARTITION operations that do not move master partitions. Adding or removing a read replica will not impact temporary tables on the primary writable copy of the database.

Added ability to SHOW METADATA permissions so users can query information_schema.mv_events.

Added requirement that ALTER USER statements in privilege_transfer_mode now need GRANT permissions.

Added ability to enable/disable AWS EKS IRSA auth for S3 pipelines. The default state for the variable enable_eks_irsa is OFF.

Improved GLOBAL temporary table errors for queries from a read replica.

Fixed an error that was possible while querying information_schema.TABLE_STATISTICS.

Fixed an issue causing some queries with table value functions to report the error, "Partition's table metadata has an unexpected table type," when flexible parallelism is enabled.

Improved CREATE RESULT TABLE error when run from a read replica. This feature is not supported for read replicas.

Fixed an issue where replication between primaries and read replicas may not pause properly during an online upgrade.

Fixed a typo in the system variable aes_default_encryption_mode.

Fixed incorrect behavior when dropping and adding a multicolumn columnstore hash index on the same set of columns in a single alter.

Improved the error information when creating a database without specifying a workspace name in the syntax on a master aggregator on a cluster with multiple workspaces.

Improved the way bad memory allocations are handled during Wasm UDF execution.

Fixed an error when using certain DATE_FORMAT options via Tableau.

Improved the performance of downloading blob files for unlimited databases when SSL is enabled between SingleStoreDB and the blobstore.

Fixed an issue with backing up an unlimited database. Backup could fail if the blob gets evicted from the blob cache during backup.

Fixed an issue that may cause the DATA API to become unresponsive under a highly concurrent workload.

Fixed an issue where the incorrect backup size showed in mv_backup_history for incremental backups.

Added support for backticks ( ` ) in column names.

Improved the performance of hash joins on expressions.

Fixed a potential issue with transaction rolled back-mid query errors that may occur when a database name has a number suffix.

Fixed an issue when specifying all columns in the SELECT command resulted in columns being out of order rather than matching the table definition.

Fixed a crash due to using the LINES TERMINATED BY clause when using INFER PIPELINE.

Fixed an issue that may cause information_schema.processlist to show “Sanitizing . . .“ instead of the actual query text.

Reduce the memory used to track cached blob files and run scans over columnstore tables.

Fixed an issue that may cause a deadlock when using the query SELECT INTO <objectstore> when flexible parallelism is enabled.

Added support for forwarding SHOW RESULT TABLES from child aggregators to the master aggregator when DDL forwarding is enabled.

Fixed crashes that may occur when downloading blobs from Azure bottomless databases.

Fixed an issue where a syntax error would occur when running queries against reference tables with comments following the query text.

Added support for NOT NULL in persisted computed columns in CREATE TABLE statements.

Added support for 256-bit keys for AES_ENCRYPT and AES_DECRYPT functions.

Fixed status so Pipelines running START PIPELINE FOREGROUND, TEST PIPELINE, and PROFILE PIPELINE will now show "Running Foreground", "Testing", and "Profiling" in information_schema.pipelines.state, rather than "Stopped".

DROP DATABASE is now only allowed from a workspace with the database attached as writable.

Disabled ability to ATTACH DATABASE as READ ONLY.