On Wed, Jun 24, 2015 at 06:08:02PM -0700, Sabyasachi Sengupta wrote: > >It's a little unconventional for us to use a wall clock time for this. > >I'd be more inclined to report it as "N seconds ago" or "N ms ago". Any > >particular reason to use a wall clock time? > > I've seen that all BFD other outputs use "now -/+" convention, but just that > I thought wall clock time was more user readable, especially because last > flap could be hrs/days ago. If we imagine that this output (last flap time) > could be parsed through a remote script as to which link went down and when, > its probably easier for them to see the absolute time rather than having to > convert it (note that the notion of 'now' could be different in both > machines?). I'd think that it makes more sense for next/last TX times > (should continue to) be in milliseconds as that reflects a more ongoing > activity.
It's much easier for a script to parse a number than a date. I doubt that anyone in real life cares whether the last flap was yesterday or a week ago. If the output says 32 seconds, I look for problems; if it says 138923 seconds, that's no big deal. > I understand that this will be based on what ovs-vswitchd 'thinks' as to > when the last flap occurred. Its still not the accurate information as > ovs-vswitchd might itself have restarted in between the last flap and when > the user actually reads it. This probably means that we save this info in > ovsdb and read it through a separate CLI, but probably we can keep it simple > for now? I don't think there's a need to store it. I imagine that restarting OVS often causes a flap anyway. _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
