I'm quite new to SSIS (actually SSDT with VS2015), and have a .sln
that uses 3 packages:
Main.dtsx
ArchiveFiles.dtsx
FTP.dtsx
The 'entry' point (is this the correct way of thinking of it?) is Main.dtsx
, which will execute the 'child' packages (no. 2 and no. 3) at various points in the flow.
To deploy, I simply just copied the wizard (the .ispac
file in /bin/Development
) which seemed to recreate my project within the Integration Services Catalogs
folder on SQL Server.
I have been able to execute the SSIS solution by right clicking and executing the Main.dtsx
package from SQL Server. Same with creating a scheduled job (just tell it to execute the Main.dtsx
package and everything works).
I'm deploying to SQL Server 2016 and SQL Server 2012.
Questions:
-
Since the project seems to have been copied verbatim to SQL Server, and the code is just XML, could I edit the deployed packages in the future? (I'm using Git for source control so I would NOT want anyone editing the package directly and then someone else deploying from the source control)
-
Is my deployment method correct and best practice? Seems like this is too simple and intuitive for the Microsoft world… (considering the headache and pain involved in working with SSIS in general)
Best Answer
Could I edit the deployed packages in the future?
You can import from GIT, open the package in SSDT and deploy again. Overwriting the existing deployed package.
You can also export using SSMS save and open in SSDT. Once you make your changes deploy again and update your source control.
As far as not letting anybody else modify or deploy packages this document will give you a good start. A SSISDB is considered another user database.
Is my deployment method correct and best practice?
As you are using datatools to create your solution I would follow this method (assuming you have the privilege).
You can also do the same using SSMS. Details here.
In case you are not aware that you can deploy package from SSMS using tsql, details here.