[SOLVED]Cacti poller interval, cron, and rrd file presets

[eholz1]

Hello All-

Just in case someone reads this now SOLVED. I set the cron to 1 minute, poller to 30 sec, and deleted all Polling Interval graphs and CPU usage graphs,
deleted all rrd files - and let them be recreated with the new settings for poller and cron.
The graphs now show true taffic flow.

Cacti 1.2.9, RHEL 7.0
This post is related to a post i put up about a year ago. I have discovered that running the cacti poller.
I have had server set for cron at 5 minutes, poller interval at 5 minutes, and an rrd preset for 30 seconds.

I have 477 devices, cisco switches and routers. I get no poller table not empty warnings.
I have also discovered that this configuration cuts the graph data at about dthe 100MB to 110MB level.
Some of our interfaces use up the the 1GB capacity - this shows only when we look at graph in realtime.

I changed the poller interval to 30sec, with cron still at 5 minutes. This produced graphs with correct traffic flow, BUT
gaps in CPU and Poller response graphs. The traffic graphs have no gaps.

So - are poller table not empty warnings unavoidable? It there a way to run cacti with a a 30 second interval, and create graphs without gaps, and without poller table not empty warnings?

I have seen other postings where a 1 minute polling cycle is configured, but with a 1 minute cron, and a custom preset. Is it possible to used the existing
1 minute preset?

I am open to suggestions.

This issue WAS quite frustrating.
eholz1

[abhisekp34]

HI Bro,

In You Case it is 400+ Devices.

I have 5000+ devices in Cacti. I have also lots of Cuts in Mostly Graph. We have tried by updating SNMP Time Out & Heartbeat on RRDFiles. But changes did not work.

Polling value is 1 minute minimum. Cant make less than that.

Do Have any Suggestion how to resolve ?

[TheWitness]

Post your Cacti.log output.