"Jason R. Mastaler" <[EMAIL PROTECTED]> writes: > Tim Legant <[EMAIL PROTECTED]> writes: > >> That could work. What would we name the files, then? If the >> directory is called, e.g., 'tmda-ofmipd', then repeating >> 'tmda-ofmipd' seems redundant (and tedious for typing). > > How about just port.pid, or if you add IP, ip:port.pid or something.
Adding the IP in addition to the port makes sense. I'll go with that. >> 'kill -9' is an abortion. It doesn't give the application a chance >> to clean up. So no, the pid file will still be there. > > It was more of a hypothetical question. In practice, people do use > SIGKILL to terminate applications even though that might be overkill. It's more than overkill. It's possible to cause file corruption, even though the program has written the data. I'm not sure what else can go wrong, but hey, "You pays your money, you takes your choice!" I can't stop people from firing guns at their feet; I can only point out that it's going to hurt. >>> Why is a warning in order? >> >> Because in all normal application end cases the pid file will have >> been deleted. If you ever receive the warning, it's an indication >> that something went wrong -- and that you should probably figure out >> what. > > Sure, but doesn't the fact that tmda-ofmipd is no longer running also > give you that same indication? The presence or lack of a pid > file doesn't provide any more information that would help you solve > the problem. Well, I'm happy with daemontools, which needs tmda-ofmipd in foreground mode, but I seem to recall hearing of other systems that re-start daemons (background mode), not needing the control that supervise wants over pipes between servers and logging programs. In those cases, you wouldn't necessarily know. The server would just restart without any notice. Maybe that's good? I dunno... >> So we're in agreement but for the warning. Does my explanation above >> change your thoughts on this? > > Not really, but issuing a warning doesn't change functionality, so it > really doesn't matter that much to me. I think if you keep the > warning short and sweet I'll be fine with it. If people start to > complain about this we can always revise later on. I'll condense it to just a line and see what others think. Maybe just a 'print' instead of using warning() (thus, no '****' surrounding it, much "quieter"). It'll be more of a note than a warning. Tim _________________________________________________ tmda-workers mailing list ([EMAIL PROTECTED]) http://tmda.net/lists/listinfo/tmda-workers
