On 19 Jun 2009, at 09:58, Karsten Müller wrote: >> Are the servers time sync'd? That would cause problems with working >> out rate of change. > > I don't understand the time issue. Do you think about master/slave > times not synced (they are)? After looking through the plugins code it > seems to me that the local time difference is used for calculating the > rate. Or did I miss something?
True. Scratch that off the list. >> Which was the previous version? High counters was not introduced >> until >> version 3.1 (from memory). > > The previous version was 3.2. As far as I see high counters are only > relevant for interfaces with speeds > 1Gbps. Perhaps it would be > easier to have a CLI switch for using high counters or not? No. We decided it was easier to just poll for a HC counter and use that if found, otherwise fall back to the normal counters. Setting a command line option would be more overhead on an administrator (and we want Opsview to be as easy as possible!). Are you sure there is no wrapping? > We use two checks to get separately the total amount of volume and the > ratio. Is it possible that these tow checks for one interface > interfere via sqlite DB? Yes, that's possible. We designed for concurrency between checks at the host and interface level, but not necessarily that there would be two service checks trying for the same interface at the same time. > We are about to implement a "line speed" CLI parameter in order to use > it as a divider for rate calculation instead of $link_speed. Do you > think this would be could be integrated in opsview? We'd be happy to look at a patch for that. I'm all for submissions! Ton _______________________________________________ Opsview-users mailing list [email protected] http://lists.opsview.org/listinfo/opsview-users
