cluther wrote: > to use the 64bit onces as the interfaces are 1Gb. > > It's completely normal for these numbers to differ. These numbers are > the total number of bytes that have passed through the interface since > the device was powered on. However, the 32bit counter has likely > rolled over (meaning starting over at 0 once it reached the maximum > number storable in a 32bit integer) more times than the 64bit counter. > This is what leads to the discrepency. > > Zenoss is only recording the delta (difference between the value at > last check and current check) and not the total number. > _______________________________________________ > zenoss-users mailing list > [email protected] > http://lists.zenoss.org/mailman/listinfo/zenoss-users
Ok, I've monkeyed with the ethernet-csmacd template, and managed to add the 64 bit OIDs to that template. (BTW they also graph just fine.) The Performance Reports only seem to 'pay attention' to the 32 bit counters, which of course roll over and skew all the results. If I play 'hack' and delete the 32 bit Octet in/out stuff from the template, all I get is N/A on the performance report screen. I tumbled to this because the utilization of several gig links approaches 70% util at night during backups. The OS graphing shows the high traffic, but the reports NEVER indicate that that particular interface was busy. Any way to make the reports pay attention to JUST the 64 bit counters? -------------------- m2f -------------------- Read this topic online here: http://forums.zenoss.com/viewtopic.php?p=29594#29594 -------------------- m2f -------------------- _______________________________________________ zenoss-users mailing list [email protected] http://lists.zenoss.org/mailman/listinfo/zenoss-users
