Difference in RepairError behavior in K2.net 2003 and K2 blackpearl
I saw this response from Blake (http://k2underground.com/forums/thread/25502.aspx) and I thought it interesting that the RepairError behaviour has changed. I did some research on this and I thought it would be nice to share out my findings on the key changes in the RepairError bahaviour.
In K2.net 2003, the repair error would affect the code of the version that the process instance is running on. So any changes would impact any running instances which is on the same process version. Checking the “Repair All Instances” option was primarily to auto-retry all the error instances having the same error. Otherwise, I would have to retry each error instance individually, this could get tedious if there were a lot of similar errors.
It seems that the repair error behavior is different in K2 blackpearl as it creates a new version and re-runs the error instances on the new version.
Depending on what is specified for the “Repair All Instances” option.
- Checked – Deploys a new version and retries all instances with the same error on the new version. The new version is also set as the default.
- Unchecked – Deploys a new version and retries only the specific instance. The default process version is not changed.
So this option now becomes quite useful in different scenarios.
- Unchecked – I have a specific error instance which is caused by a specific data error. I could do some custom code to hard code the invalid values without impacting the other instances.
- Checked – I have a general error in the process which needs to be fixed across all running instances.
I hope this article helps you understand the differences in the options and also how they function when doing a RepairError.