I notice that over time, rcapd can consume a fair amount of cpu, even though at any given time, it uses a reasonable enough amount.
Is there any way FEN could be used to make it more efficient? Like using FEN to monitor its config file(s) rather than repeatedly stat() or whatever? Or if FEN could be used /proc in such a way as to avoid other reasons for rcapd possibly having to loop over a set of process related info? It just strikes me that in in a perfect world, rcapd would be purely event-driven, and passive unless an event informed it of the need to intervene or to update its stats. As such, it's CPU usage would increase only with churn in the resources it controlled, but would otherwise grow very little with larger numbers of projects, zones, and ultimately processes to watch; which, although I haven't looked at it operating on different scales, I suspect isn't the case now. This message posted from opensolaris.org _______________________________________________ opensolaris-discuss mailing list [email protected]
