For those of you who are still stuck in the dark ages on that decrepit SQL Server 2014*, there’s a new Cumulative Update out that puts a little lipstick on that pig. Fixes include:

• Query performance improvement for columnstore
• Crashes with an access violation when you use TRY/CATCH for bulk copies
• Increased CPU with CHANGETABLE
• SSAS crashes when you CREATE SESSION CUBE
• Size of PDF file is larger than it used to be
• Deadlocks when you execute a nested loop join between two hash joins working in batch mode
• CHECKDB didn’t work with files in Azure Premium Storage
• Couldn’t uninstall SPs after installing a CU
• Memory leak with Azure Storage
• db_name() and db_id() failed to trigger permissions checks

* – SQL Server 2014 is neither decrepit nor old. This is what we call humor. Also, we should stick to our day jobs.

A-one, and a-two – here’s Cumulative Update 2, or as they say in Mexico, DOS. The fix list is pretty long, and here’s some of the greatest hits:

• DML statements unexpectedly replicated to subscribers
• Query returns incorrect results from nonclustered columnstore index
• Incorrect results are returned when query contains a GROUP BY
• Standard Edition’s Analysis Services ignores the 16GB memory limit
• Query Store returns unusual characters (like Erik Darling) when a JOIN operator is used
• Retention policy doesn’t work when you use Managed Backup to Azure
• Updating while compression is in progress can lead to nonclustered columnstore index corruption
• Access violation occurs when you use TDE and BPE
• Queries that run against Always On Availability Group secondaries always get recompiled
• Memory leak occurs when you use Azure Storage
• Access violation occurs when query contains many COUNT DISTINCT operations
• Access violation occurs if trace flag 4139 is enabled
• High CPU usage after you install SQL Server
• SQL Server doesn’t start after you configure TempDB to use a very small log file
• Access violation on secondary replica when there are too many parallel redo operations
• Query runs slowly when SQL Server uses a hash aggregate in the plan

And many, many more. Go check it out.

Build 11.0.6544 brings us new fixes including:

• SQL 2012 crashes with an access violation when you use TRY…CATCH for bulk copy
• Assertion failures when you query database-state information
• CPU usage increases when you use CHANGETABLE
• High CPU usage in general
• Queries that run against Always On secondaries always get recompiled
• SQL Server crashes when it reads data from an Extended Events file target
• SQL Server doesn’t start after you configure TempDB with a small log file
• Operating system error 32 when you restore a database

In other news, there’s a new SQL Server Management Studio build out (16.4), but the first comment on this blog post suggests that the ALTER TABLE scripting is broken, so maybe hold off on that one. That’s kind of a bummer, because 16.3 had some really ugly bugs around memory leaks and running SELECT @@TRANCOUNT after every query.

Last night, Microsoft unveiled 13.0.2149.0, and it’s a big one. It includes pages and pages of fixes like:

• Query Store automatic data cleanup was failing on Standard, Express, Eval Editions
• Fatal error when you query sys.sysindexes if you have memory-optimized tables
• Memory leak when DATA_CONSISTENCY_CHECK is run on a temporal table
• Memory leak on AG replicas when change tracking is enabled
• Temporal table data flush tasks consume 100% CPU
• Temporal table query plans with a view on an identity field do infinite recompiles
• SEEDING_MODE setting ignored when you add an AG replica
• Cannot create/delete a table or index when another DDL transaction is running
• Can run multiple UPDATE STATISTICS against a single table
• READ_ONLY and AUTOGROW_ALL_FILES not preserved after a database restart
• Fixed error 3456 that suspended AG replicas
• Incorrect results when you use LIKE with an “ss” wildcard
• Incorrect results with full text search
• Crash when running a query with HASHBYTES
• Non-yielding scheduler on queries with a UNION
• Data loss when you alter a column on a large memory-optimized table
• Data loss when you alter an in-memory temporal table
• Data loss or incorrect results when you use sp_settriggerorder
• Tons of Master Data Services fixes
• FileTables unavailable after a failover in an AG
• Availability databases not fully started after restarting an istance
• Error 1478 when you add a database back into an AG
• Online index operations were blocking on clustered columnstore indexes
• All data was going to deltastores when you bulk loaded a clustered columnstore index under memory pressure
• CHECKDB false errors on non-clustered columnstore
• .NET Framework 3.5 required for log shipping secondaries
• Cannot install non-English version on Windows Server Core
• Canceling a backup task crashes SQL Server

Check out the full list of fixes and download the Cumulative Update.

Wanna know when new updates come out? Subscribe on our home page.

SQL Server 2012 SP3 CU4 is out, and if you’re still on SP2 there’s a CU13.

SP3 CU4’s improvements include:

• Less crashing when you run a remote query in a stored proc with an invalid user name
• Faster backups to a network share
• Windows dedupe jobs don’t corrupt FILESTREAM data
• CHECKDB raises more alarms about invalid filegroups
• Fixed a memory leak in AlwaysOn with change tracking
• Log backup failures on secondaries stop preventing subsequent backups
• And more.

Doesn’t look like any performance improvements in this one. (You know things have gotten great in the SQL Server community when it’s an unusual CU that doesn’t include performance or diagnostic improvements.)

It was a dark and stormy night. The next morning, SQL Server 2014 SP2 arrived, chock full of fixes like:

• New DBCC CLONEDATABASE command to generate a statistics-only copy of the database for CE troubleshooting
• Faster performance for large databases with a large number of partitions
• More AlwaysOn (note the lack of space and AG in the headline) extended events and Perfmon counters
• Memory grant warnings in XML query plans
• Plans include trace flags
• Per-operator query execution statistics in query plans
• New min_grant_percent and max_grant_percent query hints
• Faster spatial queries
• Memory grant fields in sys.dm_exec_query_stats

And much more! Lots of good stuff for performance tuners in here. Go get ’em, tiger.

Microsoft just unveiled the latest fashions for their summer lineup. All the best-dressed SQL Server 2014s will be running the new Cumulative Update 7, which includes hotfixes for stuff like:

• High CPU usage and PAGELATCH_* waits on SP1
• Non-yielding scheduler errors on queries with a large number of OR clauses
• Log shipping retries can write over a successful backup, breaking the log chain
• Creating tables takes 10 minutes if you have 2-4TB memory
• Service Broker leaks memory if you use TABLOCK hints in an AG
• DBCC CHECKTABLE returns an incorrect result on rebuilt clustered columnstore indexes
• Better query plans for columnstore indexes
• Access violation when you combine TDE and BPE
• MERGE fails when CDC is enabled (seriously, stop using MERGE)
• Incorrect results when you use LIKE with “ss” against German data
• Incorrect page numbers in SSRS exports
• Change tracking cleanup fails if another database is offline
• SQL doesn’t start if you set TempDB’s log file to be smaller than 248KB (look, that one is your own dang fault)

Read the SP1 CU7 details, or if you’re dead set against installing SP1 for some surely oddball reason, like you still like pleated pants, check out RTM CU14.