> On Mon, May 12, 2003 at 11:01:58PM -0500, Glenn Crocker wrote: > > > > > > It's supposed to. If things get _really_ bad, RNFs can happen. > > > > > > Well, if things are really bad - then it is hawk's own fault for not > > > dampening incoming requests effectively enough. > > > > If things get really bad, the node should restart itself. > > > > For example, a node that at one point had lots of traffic (was well > > integrated) and has none for some amount of time (hours) can reasonably > > conclude that Something Has Gone Wrong and that it might as > well restart. > > No, this is not a well-reasoned approach, but yes, it is the > right thing to > > do. > > No, it is a workaround for, primarily, buggy JVMs.
Yep, "JVM bugs" would fall in the realm of "Something Has Gone Wrong". It's hard to fix all the bugs and it's easy to make the node cope better (by having it restart). Add a checkbox so power users can disable restarts and help track down weird bugs and everyone would be happy. -glenn _______________________________________________ devl mailing list devl at freenetproject.org http://hawk.freenetproject.org:8080/cgi-bin/mailman/listinfo/devl
