Hi, I am hoping someone can help me figure out why TeamCity is re-running builds in our build chain when a suitable one already exists.
Any ideas welcome, I am out of ideas...
We have a extensive build chain setup in one of our projects which has multiple dependancy chains as show here.
The problem we are seeing is that there appears to be no stability in the reasoning for re-running dependent build configurations. We can run the "Git Hub Integrator" build several times in a row and each time get different results on what will be run. If a run of all dependancies succeeds, the next run should not re-run any of the dependancies in the chain unless there has been a code change.
Example: (No code changes in between runs, verified by github hash on each run)
Build 3735: A run of the GitHub Integrator configuration with the "rebuild all snapshot dependencies transitively" selected re-ran everything and each succedded
Build 3736: No code changes were committed, as shown in the dependancies image below, and the git hub revision hash between Build 3735 and Run 3736 is the same.
Notice that while the dependancies show no changes, some builds were re-ran. (Compile API, Smoke Schema, Full Automation Suite, Smoke API)
I have walked down the dependancy chain and verified that in each case for the above chain a snapshot dependancy exists and has the following two options checked. "Do not run new build if there is a suitable one", "Only use successful builds from suitable ones"