Byron Nevins wrote: > Wow. Thanks for the reply. This is really food for thought. > > I could change our code so that a "stop" on the server first checks > to see if the server is being managed by SMF. If it is -- then we go > through SMF *instead* of our typical "external" stop procedure.
If you're referring to the "stop" mechanism you hand users, yes. > Of course we will still have the problem if the JVM crashes or a kill > -9 is given by the user... If your service mysteriously disappears, SMF assumes it failed and restarts it. This is implemented with a stop followed by a start. If you fix your SMF stop method to return 0 when the service isn't running, this restart should be successful. Dave
