SSIS Deployment Strategies

I am considering several deployment strategies for SSIS and wondering which one is the easiest to support. I tend to use SQL Server configurations to host the connection strings, and then run proc after I imported the package from the development server to change the connection strings. However, I have 75 packages and it seems a bit tedious. Can anyone suggest a good deployment strategy?

I have Dev, Stage and some live servers for deployment.

+3
source share
3 answers

We use SQL configurations for most packages created by my team. We decided to solve the migration problem by adding a second configuration based on an environment variable that tells the package which configuration database to use. This may not be a good option for everyone, but it works well for our installation.

More details:

  • Our packages always run from agent jobs.
  • Each of our environments is on a separate machine (we do not use named instances except the sandbox.)
  • We have a copy of our DB configuration in the default SQL instance for each machine, and it uses the same db name and schema in each environment.
  • The package looks at the machine name environment variable to indicate which machine is running the package.
  • -, .

, SQL . , .

, , , , . , . , , , .

+2

XML , . , ( Manifest Deployment). 75 , .

0

, . SSIS. - , - SSIS_CONFIG. .

, . ConfigurationFilter , ConfiguredValue .

SSIS_CONFIG. , ( SSIS ).

SSIS_CONFIG XML, . SQL Server, SSIS_CONFIG , .

, , , , , , .

. - ConfiguredValue. , DEV DEV, QA QA.

Editing packages as they move between environments makes testing invalid. This approach allows us to promote packages without affecting them. The design is also extremely flexible, which greatly facilitates development and testing.

We can use this approach with several instances running on the same computer, using this as a guideline: http://www.sqlservercentral.com/articles/Integration+Services+(SSIS)/69739/

0
source

Source: https://habr.com/ru/post/1796306/


All Articles