This mirrors a change made in the Shepherd 0.8.0. Previously, upon
startup failure, we could have left processes behind.
* gnu/build/shepherd.scm (read-pid-file/container): Kill (- PID) instead
of PID.
(read-pid-file pid-file
#:max-delay max-delay)))
(#f
- (catch-system-error (kill pid SIGTERM))
+ ;; Send SIGTERM to the whole process group.
+ (catch-system-error (kill (- pid) SIGTERM))
#f)
((? integer? container-pid)
;; XXX: When COMMAND is started in a separate PID namespace, its