SNMP and OIDs

[Grogman]

Ok, I've increased the timeouts from 500ms to 5000, and the retry count from 1 to 5. Now let's see what happens...

However, I usually see entries in the poller_output table after each polling cycle, so I suppose that is not related with timeouts, am I rigth?

Thank you very much for your help!!!

[Grogman]

It didn't worked

I made the changes at 8:00, and as you can see in the attached graph, gaps still appear...

Thanks!

[Grogman]

Hi people

Finally I've fixed the gaps in my graphs!!!

After months of gaps and problems, yesterday we realized that the guy who installed cacti made a mistake. He inserted the line regarding the poller in crontab, but also in a file in the cron.d directory, so the poller was running twice for each polling cycle. Yesterday we discovered it, and after deleting the entry in cron.d directory, the gaps vanished...

Thanks to all people who helped me, specially The Witness

[gandalf]

Oops, this is a very well known mistake. RPM installs will introduce /etc/cron.d/cacti, and if you edit /etc/crontab, you have two pollers. That will not work. I've mentioned this already LOADS of times, but must have forgotten it here. Sorry for that
Reinhard