Try to find a way to skip the rest of the server tests if the server goes down unexpectedly. Running all the tests when there is no server present creates a huge log which makes it hard to diagnose the particular issue.
We had a case where a nightly build of cudf broke the server healthcheck and the server failed on the first check after boot. But, all the tests continued to run creating a hug log of failed connections.
Try to find a way to skip the rest of the server tests if the server goes down unexpectedly. Running all the tests when there is no server present creates a huge log which makes it hard to diagnose the particular issue.
We had a case where a nightly build of cudf broke the server healthcheck and the server failed on the first check after boot. But, all the tests continued to run creating a hug log of failed connections.