To facilitate discussions and sharing of code samples for DacFx and SqlPackage, a GitHub repository has been made available. In recent updates for SqlPackage, 18.7 (March 2021) and 18.7.1 (June 2021), functionality was added specific to big data in Azure Synapse Analytics, options were added to improve import performance in some scenarios, and changes were made to make the troubleshooting and general use of SqlPackage easier. With version 18.7 and above of SqlPackage, Azure Synapse Analytics gains improved portability through the Parquet flat file format. Through additional command line properties on the Extract operation (see documentation) data is written to flat files in Azure Blob Storage while the schema is stored in a Dacpac file. Flat files storing data can be published to a database along with a Dacpac through the Publish operation with similar command line properties. Both Extract and Publish operations have large performance improvements when writing to Azure Blob Storage for large database sizes. In some SqlPackage import scenarios indexes have a large impact on the performance, whether it is an index assisting with data insertion speed or rebuilding indexes hampering other operations. To allow further tuning SqlPackage’s behavior in these cases, the command line properties “RebuildIndexesOfflineForDataPhase” and “DisableIndexesForDataPhase” were added to the Import action in SqlPackage 18.7. If you are looking to better understand the SqlPackage operations for troubleshooting or performance improvements, there are some essential steps to capture more information. Enabling a diagnostics file with the parameter “/DiagnosticsFile:” is fundamental and in the latest SqlPackage release the version and architecture were standardized at the top of the file to assist in troubleshooting version differences. Trace data including timestamps for performance can be included by setting an environment variable “DACFX_PERF_TRACE” to true.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |