when you tail the zenperfsnmp.log I see that 100 devises show no errors, then there are about twelve devises that the log shows as snmpd agent down. And they are listed as failed to connect. Strangely, all 12 devises reslove to names that no longer exist in zenoss. They were updated to new names and in the UI, the devises with the correct names show no errors at all. These are remnants of older servers/devises that do not exist and or have been renamed. From the logs it looks like zenoss keeps trying to access these devises and then fails causing a very large cyle time in the logs like 700 seconds before failing. This is far beyond the UI setting of 300. Yet, before looking at the logs and taking the advice of changing the cycle time to beyond 300, I did the opposite and cut my cycle time in half. My graphs looked at lot better I think because the cyle time interuppted the errors and all the devises that are working fine could draw their graphs without waiting for these bogus entries to error out. When I changed the cycle time to 500 seconds, I got no data on the graphs at all. I think these cached names are causing very long polling times, and making the cycle time even longer makes the problem worse to the point where graphs don't work at all. I need to figure out a way to get rid of these devises that don't exist anymore, or have been renamed from showing up in the logs. They also show up as snmp down in the event logs.
-------------------- m2f -------------------- Read this topic online here: http://community.zenoss.com/forums/viewtopic.php?p=18844#18844 -------------------- m2f -------------------- _______________________________________________ zenoss-users mailing list [email protected] http://lists.zenoss.org/mailman/listinfo/zenoss-users
