Ernst, is the reload vs. restart issue really related to this bug? This was about doing DNS checks, not a bug in the upstart script. Should there be a new bug report instead?
BTW, "reload" is not supposed to kill the daemon. For oneiric, "reload" should be fixed to actually work, instead of using "restart" instead. If that fix is easy, then it should be backported; if it's complex, then the "restart" workaround might be more appropriate for an SRU.
Ernst, is the reload vs. restart issue really related to this bug? This was about doing DNS checks, not a bug in the upstart script. Should there be a new bug report instead?
BTW, "reload" is not supposed to kill the daemon. For oneiric, "reload" should be fixed to actually work, instead of using "restart" instead. If that fix is easy, then it should be backported; if it's complex, then the "restart" workaround might be more appropriate for an SRU.