Gaps in my graphs

[tholmes]

Hi Folks:

I am experiencing a very frustrating problem -- I have a new install of version 0.8.6g on a newly created FC3 box. The computer is a pentium 4 2.0 HT with 512 memory and 30 gigs on the disk. Running the versions of MySQL, PHP and Apache that came with the base install of FC3. The machines in question are a collection of foundry switches a windows 2003 server and a linux (FC3) server - all have snmp properly set up and can be snmp walked. The problem is that i am only getting data about 10% of the time on the servers and there are some gaps in some of the graphs on the switches as well. I am using the php poller. I am kinda running out of ideas here, and could really use some help, as i need to get this working. If you have questions etc, please feel free to IM, or email as necessary. If someone needs me to (and can tell me now to on this forum) i can post a screen shot of the graphs

thanks

TIM HOLMES
Medina Christian Academy

[rony]

Have you looked at your cacti.log? Anything in there to give you and indication of an issue?

What is your snmp timeout setting for the hosts that you are having issues with? Or is it all hosts?

[tholmes]

I have been examining the cacti.log, and it indicates that the poller time is exceeding 296 seconds, so for debugging purposes i deleted all but two of the hosts (the core switch and the backup domain controller) and i am waiting for the first run to determine what happens then.

As far as the snmp timout, i have used the default times that cacti was set up for

TIM

[gandalf]

Please have a look at /etc/crontab and /etc/cron.d or the like. If you'll find a file named /etc/cron.d/cacti, you don't need a second poller entry in /etc/crontab. Make sure, that only one poller is running!
For more debugging, switch to DEBUG mode (use "Settings" -> Global") and have a look at cacti.log. It is not wise to run DEBUG mode for a long time!
Reinhard