Rerun failed tests

This should help stabilize the test set by making sure the test actually
fails twice. This'll still leave the output in the log so that the
problems that only occur on the first run can be eventually fixed.

maxscale-system-test/mdbci/run_test.sh

@@ -83,6 +83,7 @@ else
     eval "arguments=(${test_set})"
     ctest -N "${arguments[@]}"
     ctest -VV "${arguments[@]}"
+    ctest --rerun-failed -VV
 fi
 cp core.* ${logs_publish_dir}
 ${script_dir}/copy_logs.sh

maxscale-system-test/mdbci/run_test_snapshot.sh

@@ -80,6 +80,7 @@ fi
 
 ulimit -c unlimited
 ctest $test_set -VV
+ctest --rerun-failed -VV
 cp core.* ${logs_publish_dir}
 ${script_dir}/copy_logs.sh