Hi,
It seems there is a fairly large delay in some of our builds between cancelling a build and it actually completing and being marked as Cancelled.
My initial feeling is that it is due to some logic with trigger steps? Ex.
I have a Build that the Server issues a cancel command at
2024-02-06 14:07:20,071 WARN com.pmease.quickbuild.grid.GridTaskFuture - Job still exists on job node and cancel command is issued
The node running the trigger step issues a cancel command at
2024-02-06 14:07:20,137 WARN com.pmease.quickbuild.grid.GridTaskFuture - Job still exists on job node and cancel command is issued
The trigger step doesn't start running until
14:07:20,314 INFO - Running step...
14:07:20,129 INFO - Checking step execute condition...
14:07:20,129 INFO - Step execute condition satisfied, executing...
14:07:20,314 INFO - Executing pre-execute action...
14:07:20,314 INFO - Running step...
The entire build doesn't completely cancel until 14:11:11,200 which is almost 4 minutes after the initial cancel command was issued.
I have another build where the node in charge of the trigger step issues a cancel command at
2024-01-31 13:18:33,346 WARN com.pmease.quickbuild.grid.GridTaskFuture - Job still exists on job node and cancel command is issued
and the trigger step starts up after:
13:18:33,780 INFO - Checking step execute condition...
13:18:33,780 INFO - Step execute condition satisfied, executing...
13:18:34,038 INFO - Executing pre-execute action...
13:18:34,038 INFO - Running step...
Is there any way to prevent the trigger when the cancel command is issued prior to it running?
And, even with the build being triggered, why is there such a large delay between the build cancelling and for it to stop running?