On Mon, 2005-09-05 at 15:50 -0400, Robert Story wrote: > DS> > - add host to the default list > DS> > DS> Globally, or on specific architectures? > > You mean there are still architectures which don't support it?
Almost certainly. Much of the HostRes code is just as O/S-specific as the MIB-II stuff. When I proposed adding "host" to the default list a couple of years ago, this was the main argument against doing so. > Sigh. Indeed. Sigh! > I'd argue > that it should be in the default list, and config_require used to weed out > bits that don't have OS support. That's probably too much work, so it might > just be easier to just do known good architectures. The trouble is that if we don't start pushing it onto "broken" systems, very little will happen. The HostResources implementation is some eight years old now, and development has been sporadic at best. SIGH! I'm not sure what timescale we're thinking of for 5.3, but it may well be a little late to push this now. So you're probably right: for 5.3, just enable it on known good architectures. But once that's out, we should perhaps consider enabling it by default on the main development line, and wait for the complaints (and fixes?) to come flooding in. The other thing that might help is a move towards greater use of the Hardware Abstraction Layer that I started to put in place a couple of months ago. It currently only covers CPU and memory (and only for Linux boxes), but it's a step in the right direction. But again, that's probably more of a 5.4.dev thing. > DS> > - add disman/event-mib to the default list > DS> But there are certain aspects of the current > DS> implementation that are not according to spec (IMO). > DS> And part of the rationale behind my recent > DS> implementation of the Schedule MIB was to lay the > DS> groundwork for a reworking of the Event MIB code. > > Would that include any functional changes that would affect users? In terms of snmpd.conf configuration - probably not. There are a few aspects of the existing directives that might need to be tweaked slightly, but I'd expect to keep things backward compatible. The most visible changes would probably affect the contents of the MIB tables that were produced as a result - particularly with the linkage between the trigger tables and event entries. Given that most people probably don't manipulate these directly, then I doubt this would be particularly significant. Dave ------------------------------------------------------- SF.Net email is Sponsored by the Better Software Conference & EXPO September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf _______________________________________________ Net-snmp-coders mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/net-snmp-coders
