It doesn't seem that way to me.
The only time I've had issues with cacti with that kind of problem were
(1) attempting to run cacti on a Windows machine, and (2) running the
CLI-based poller.
My experience was that a Windows machine just required a whole lot more
horsepower to do the same work as a Linux-based machine (like an order
of magnitude). The CLI-based poller also required a lot more resources
to do the same work.
Check your cacti log to see how long your polling cycles are taking. If
it's > 300 seconds that would explain a lot of your issues.
bp
On 11/4/2014 9:38 AM, Garrett Fosmark via Af wrote:
Is it just me or is cacti just unstable? It seems like at least once
every two weeks devices will either disappear, the poller cache needs
to be rebuilt or different ping methods need to be selected in order
to get devices to start graphing again. That or graphs will say Nan
for a day and then go back to normal.
-Garrett
