Jul 13, 2011

Checkpoints in SSIS Packages

Checkpoints are a great tool in SSIS that many developers go years without even experimenting with. I hope to enlighten you on what Checkpoints are and why it is beneficial to use them. Also, I will walk you through a basic example package where they have been implemented.
What does it do?
With Checkpoints enabled on a package it will save the state of the package as it moves through each step or task and place it in a XML file upon failure of the package. If your package does fail you can correct the problem in your package and rerun from the point of the tasks that did not successfully run the first time. Once the package completes successfully the file is no longer needed and automatically discarded.
How does this benefit you?
Just imagine your package is loading a table with 10 million records. Your package passes the Data Flow that performs this huge load without any problem (Other than the fact that it took two hours to load). The next task in your package is a Send Mail Task and for some reason fails.
You correct the problem in the Send Mail Task, but without using Checkpoints your package would still have to run that Data Flow that loads the 10 million records again (taking another two hours) even though you’ve already done it once. If you had enable Checkpoints on this package you could simply correct the problem in the Send Mail Task and then run the package again starting at the Send Mail Task. Sounds great right?
How do I configure it?
This example will run you through very basic package using Checkpoints.
Example Overview

 

  •  Use Three Execute SQL Task using the AdventureWorks2009 (It can really be any database for this example) database as a connection manager.
  •  Configure the package to handle Checkpoints
  •  Configure the individual tasks to handle Checkpoints
Step 1: Configure Execute SQL Tasks
  •  Drag three Execute SQL Tasks on your Control Flow.
  •  Use any database for the Connection property on all three tasks
  •  Configure Execute SQL Task SQLStatement property: Select 1
  •  Configure Execute SQL Task 1 SQLStatement property: Select A (Set to intentionally fail)
  •  Configure Execute SQL Task 2 SQLStatement property: Select 1

Step 2: Configure Package to enable Checkpoints

  •  Open the properties menu at the package level (Just open properties in the Control Flow without any task or connection manager selected)
  •  Change the properties CheckpointFileName: c:\Checkpoint.xml (Feel free to use the .txt extension when naming the checkpoint if you want to open it in notepad and look at it!)
  •  Change the properties CheckpointUsage: IfExists
  •  Change the properties SaveCheckpoints: True
Step 3: Configure Each Task

  •  Select each task individually and open the properties menu at the task level (Just click the task once then hit F4)
  •  Change the FailPackageOnFailure property to True
Step 4: Run the Package

  •  Run the package and you will see the package fail on the second task
  •  This also created the file c:\Checkpoints.xml. Feel free to open it and take a look! I use the tool XML Notepad to view XML Files. It’s Free.
  •  You could also save this file with the.txt extension and just view in regular notepad and it still works as a Checkpoint.

•  If you run the package a second time it will skip the first task that was successful and start right at the second task
Step 5: Correct the Problem and Rerun Package

  •  Open the Execute SQL Task 2 and configure the SQLStatement property: Select 1
  •  The package has now completed and skipped the first step which already succeeded. Imagine if that first step would normally take two hours to run!

No comments:

Post a Comment

Hi,

Thanks for your visit to this blog.
We would be happy with your Queries/Suggestions.