New SQL Server 2017 features
Are you considering an upgrade to a more modern version of SQL Server? Are you choosing between SQL Server 2016 or SQL Server 2017? If so, then my advice is to upgrade to SQL Server 2017 as I explain in this post.
The release of SQL Server technology provides lots of interesting new features for SQL administrators and developers to ponder. The Community Technology Preview (CTP) 2.0 for SQL Server vNext (generally called SQL Server 2017) is no exception. Many updates have been implemented in the existing features and services of the application. In this blog post, I discuss what is new in the database engine of SQL Server 2017 from a database administrator (DBA) perspective.
Microsoft® SQL Server 2017 is now on the scene with more new features that offer faster processing, more flexibility of use, and greater cost savings as a result. SQL Server 2016 provided many improvements, which Microsoft termed as a big leap forward. However, if SQL Server 2016 was a big leap, then SQL Server 2017 promises all that and much more of what corporate customers need at all levels. Database performance has reached a new peak with adaptive query processing, new flexibility with cross-platform capabilities, new integrations for statistical and data science analysis, and SQL Server versions on Linux®, Ubuntu® operating systems, or Docker®. The new version adds solid technology with cost savings.
The changes introduced in SQL Server 2017 discussed here include the following features:
- SQL Server on Linux
- Resumable online index rebuild
- SQL Server machine learning services
- Query processing improvements
- Automatic database tuning
- TempDB file size improvements
- Smart differential backup
- Smart transaction log backup
- Improved SELECT INTO statement
- Distributed transaction support
- New availability groups functionality
- New dynamic management views
- In-memory enhancements
- Security enhancement
- High availability and disaster recovery
- Performance improvements
SQL Server is no longer just a windows-based relational database management system (RDBMS). You can run it on different flavors of the Linux operating systems. You can also develop applications with SQL Server on Linux, Windows, Ubuntu operating systems, or Docker and deploy them on these platforms.
This feature resumes an online index rebuild operation from where it stopped after events such as database failovers, running out of disk space, or pauses.
The following images show an example of this operation:
Guidelines for indexing
When you perform online index operations, the following guidelines apply:
- Clustered indexes must be created, rebuilt, or dropped offline when the underlying table contains image, ntext, and text large object (LOB) data types.
- Non-unique and non-clustered indexes can be created online when the table contains LOB data types but none of these columns are used in the index definition as either key or as non-key (included) columns.
- Indexes on local temporary tables cannot be created, rebuilt, or dropped online. This restriction does not apply to indexes on global temporary tables.
- You can perform concurrent online index data definition language (DDL) operations on the same table or view only when you are creating multiple new non-clustered indices, or reorganizing non-clustered indices. All other online index operations performed at the same time fail. For example, you cannot create a new index online while rebuilding an existing index online on the same table.
SQL Server 2016 integrated the R programming language, which can be run within the database server and can be embedded into a Transact-SQL (T-SQL) script, too. In SQL Server 2017, you can execute the Python script within the database server itself. Both R and Python are popular programming languages that provide extensive support for data analytics along with natural language processing capability.
SQL Server 2017 adapts optimization strategies to your application workload’s runtime conditions. It includes adaptive query processing features that you can use to improve query performance in SQL Server and SQL Database.
There are three new query improvements as shown in the following diagram:
- Batch mode memory grant feedback: This feedback technique recalculates required memory for the execution plan and grants it from cache.
- Batch mode adaptive joins: To execute the plan faster, this technique can use a hash join or a nested loop join. After scanning the first input of the execution plan, it decides which join to use to produce output at the fastest speed.
- Interleaved execution: Interleaved execution pauses optimization of an execution plan when it encounters multi-statement table-valued functions. Then, it calculates perfect cardinality and resumes optimization.
This feature notifies you whenever a potential performance issue is detected and enables you to apply corrective actions, or it enables the database engine to automatically fix performance issues caused by the SQL plan choice regressions. Thus, the database can dynamically adapt to your workload by finding what indexes and plans might improve performance of your workloads and what indexes affect your workloads. Based on these findings, the automatic tuning process applies actions that improve the workload performance. In addition, the database continuously monitors performance after any change made by automatic tuning to ensure that it improves the workload performance. Any action that doesn’t improve performance is automatically reverted.
SQL plan choice regression
The SQL Server database engine may use different SQL plans to execute the T-SQL queries. Query plans depend on the statistics, indexes, and other factors. In some cases, the new plan might not be better than the previous one, and the new plan might cause a performance regression. Whenever you notice a poor plan choice regression, you should find a previously used good plan and force it to be used instead of the current one by using the sp_query_store_force_plan procedure. The database engine in SQL Server 2017 (v. 14.x) provides information about regression plans and recommended corrective actions. Additionally, the database engine enables you to fully automate this process and let the database engine fix any problems related to the plan changes that are found.
Automatic plan correction
The automatic plan correction is shown in the following diagram:
The following automatic tuning features are available:
- Automatic plan correction (available in SQL Server 2017 v14.x and Azure SQL Database): It identifies problematic query execution plans and fixes the SQL plan performance problems. Automatic tuning is enabled using the following command:
- Automatic index management (available only in Azure SQL Database): It identifies indexes that should be added in your database and indexes that should be removed.
SQL Server 2017 setup now enables you to specify the initial TempDB file size up to 256 GB (262,144 MB) per file, with a warning if the file size is set greater than 1GB without instant file initialization (IFI) enabled. It is important to understand that, depending on the initial size of TempDB data file specified, not enabling IFI can cause setup time to increase exponentially.
A new column modified_extent_page_count is introduced in sys.dm_db_file_space_usage to track differential changes in each database file in the database. The new column modified_extent_page_count allows DBAs, the SQL community, and backup independent software vendors (ISVs) to build smart backup solutions, which perform differential backups if the percentage of changed pages in the database is below a threshold (approximately 70-80%). Otherwise, they perform a full database backup. With a large number of changes in the database, the cost and time to complete differential backups is similar to taking a full database backup, so there is no real benefit of taking differential backup in this case. However, it can certainly increase the restore time of database. By adding this intelligence to the backup solutions, you can now save on restore and recovery time by using differential backups.
A new Dynamic Management Function (DMF), sys.dm_db_log_stats (database_id), was released. This function exposes a new column, log_since_last_log_backup_mb, which empowers DBAs, the SQL community, and backup ISVs to build intelligent T-log backup solutions to take backups based on the transactional activity on the database. This T-log backup solution intelligence ensures that, if the T-log backup frequency is too low, the transaction log size doesn't grow due to a high burst of transactional activity in a short time. It also helps to avoid a situation where the scheduled transaction log backup creates too many T-log backup files even when there is no transactional activity on the server. If that happened, it would add unnecessarily to the storage, file management, and restore overheads.
In SQL Server 2017, you can provide the filegroup name on which to create a new table by using the ON keyword with the SELECT INTO statement. The table is created on the default filegroup of the user by default. This functionality was not available in previous versions.
SQL Server 2017 supports distributed transactions for databases in availability groups. This support includes databases on the same instance of SQL Server and databases on different instances of SQL Server. Distributed transactions are not supported for databases configured for database mirroring.
This functionality includes clusterless support, the Minimum Replica Commit Availability Groups setting, and Windows-Linux cross-OS migrations and testing.
This functionality includes the following features:
Availability groups can now be set up without an underlying cluster (Windows Server Failover Cluster or WSFC) and across mixed environments (instances on Windows and Linux or Docker).
The new Minimum Replica Commit setting enables you to dictate a certain number of secondary replicas. You must commit a transaction before committing on the primary.
The dynamic management views (DMVs) include the following elements:
- sys.dm_db_log_stats exposes summary level attributes and information on transaction log files and is helpful for monitoring transaction log health.
- sys.dm_tran_version_store_space_usage enables you to see the impact on version store usage, grouped by each database. As a result, you can use this to profile your workload in a test environment (before and after the change) and to monitor the impact over time–even if other databases are also using version store.
- sys.dm_db_log_info exposes virtual log file (VLF) information to monitor, alert, and avert potential transaction log issues.
- sys.dm_d_stats_histogram is a new dynamic management view for examining statistics, as shown in the following image:
- sys.dm_os_host_info exposes things like platform, distribution, service pack level, and language.
- sys.dm_os_sys_info was expanded, revealing CPU information (such as socket count, core count, and cores per socket).
The in-memory changes in SQL Server 2017 include the following enhancements:
- Computed column, and indexes on those columns, are now supported.
- CASE expressions, CROSS APPLY, and TOP (N) WITH TIES are now supported in natively-compiled modules.
- JSON commands are now fully supported in both check constraints and in natively-compiled modules.
- The system procedure sp_spaceused now properly reports space for memory-optimized tables.
- The system procedure sp_rename now works on in-memory tables and natively-compiled modules.
- The limitation of eight indexes on memory-optimized tables has been eliminated.
- Memory-optimized filegroup files can now be stored on Azure storage.
You can now grant, deny, or revoke permissions on database-scoped credentials such as CONTROL, ALTER, REFERENCES, TAKE OWNERSHIP, and VIEW DEFINITION permissions. Also, ADMINISTER DATABASE BULK OPERATIONS is now visible in sys.fn_builtin_permissions.
Gain mission-critical uptime, fast failover, easy setup, and load balancing of readable secondaries with enhanced Always On functionality in SQL Server 2017. This is a unified solution for high availability and disaster recovery on Linux and Windows. You can also put an asynchronous replica in an Azure virtual machine for hybrid high availability.
SQL Server 2017 introduces the following changes to the way queries and statistics are collected and displayed:
- A new DMV sys.dm_exec_query_statistics_xml allows you to correlate sessions to plans, as long as query profiling is enabled. The following image illustrates this:
- Showplan XML now includes information about the statistics used for a plan and, for actual plans, runtime metrics and the top 10 wait statistics experienced by that plan. These wait statistics are also now being tracked in the query store.
- A new dynamic management function sys.dm_db_stats_histogram enables you to access histogram information programmatically, without databases console commands (DBCC).
There are many changes in SQL Server 2017 that could help your implementation, and a wealth of information is out there to help you. Even though you might think SQL Server 2017 was “just a Linux port,” there are significant real-world improvements in the core database engine that benefit all platforms. Microsoft continues to add extra functionality to the product in SQL Server 2017 cumulative updates, and SQL Server 2017 will be fully supported by Microsoft for longer than SQL Server 2016.
Use the Feedback tab to make any comments or ask questions.