fstest: make test server start fail fast when docker run fails

The start function was run inside an if condition, where bash ignores
errexit, so a failed docker run was not noticed. The script then
printed its connection details anyway and the test spent 100 seconds
trying to connect before failing with a message that hid the real
error.

Run start in a subshell with errexit on and check its status
explicitly so the failure is reported immediately with docker's
error message.

(cherry picked from commit 4cf1da0d4f)

fstest/testserver/init.d/run.bash

@@ -49,7 +49,15 @@ case "${1:-}" in
       if _is_running; then
         stop || true
       fi
-      if ! out="$(start)"; then
+      # Run start in a subshell with errexit on so a failing command
+      # (eg docker run) aborts it. The subshell's status is only
+      # honoured if it is not part of an if/|| condition, as bash
+      # ignores errexit inside those, so capture it separately.
+      set +e
+      out="$(set -e; start)"
+      start_rc=$?
+      set -e
+      if (( start_rc != 0 )); then
         echo "failed to start" >&2
         exit 1
       fi