Hi
I am reposting this topic as i am not able to solve this particular issue . I recently updated from 0.8.6i to 0.8.7a version and all my desired graphs r plotting fine except for the CPU Usage graph.i have added around 50 devices to my cacti database and around 47 r showing me all the graphs which i have created..The rest of the three servers having a similar setting and OS`s as the other servers r not plotting me this CPU Usage graph.Below r a few info required abt my servers.
OS version - Solaris 5.9
Cacti version - 0.8.7a
SNMP Version - 5.4
I manually tried walking these servers which r not plotting the graphs and i am getting the values, the problem is that these values r not being plotted...Someone kindly help..
Note: If there is any seperate template for CPU Usage in 0.8.7a , kindly upload the same
HELP NEEDED...CPU USAGE GRAPH NOT BEING PLOTTED
Moderators: Developers, Moderators
HELP NEEDED...CPU USAGE GRAPH NOT BEING PLOTTED
Regards
Mayur
Mayur
Be sure to skim Gandalf's URLs on debugging tips.
Some other things I sometimes do to figure out problems like this... If you go to the cacti settings and set the log level higher you can see in the log file the values the cacti poller is getting back before adding it to the rrd file. If I still don't see values it usually means I've done something silly like made a data template of type gauge instead of counter and the counter cacti is getting back is much larger than the max value specified in the template. I'll sometimes run a sniffer like tcpdump, snoop, or ethereal/wireshark and confirm that cacti is making the snmp queries I want it to and getting the results I expect it should.
Or, sometimes I've set the max value on a data source to be some value, then later (after the RRD file has already been created) tried to set it to something higher. But the first max value is already in the RRD file. I see this with CPU graphs sometimes where I think the machine I'm graphing cpu for is a 4 processor machine (so I set the max value on it's cpu related data sources to 400) then later discover it's really a quad-core machine and not a dual-core machine. Increasing the max values on the data sources doesn't change the max values specified in the RRD file. I usually smack myself on the forehead a few times, then export the RRD file to an XML file, edit the maximum values in the XML file, then use rrdtool restore to turn it back into an RRD file. Or if I'm lazy I'll just delete the RRD file and let cacti re-create it.
Some other things I sometimes do to figure out problems like this... If you go to the cacti settings and set the log level higher you can see in the log file the values the cacti poller is getting back before adding it to the rrd file. If I still don't see values it usually means I've done something silly like made a data template of type gauge instead of counter and the counter cacti is getting back is much larger than the max value specified in the template. I'll sometimes run a sniffer like tcpdump, snoop, or ethereal/wireshark and confirm that cacti is making the snmp queries I want it to and getting the results I expect it should.
Or, sometimes I've set the max value on a data source to be some value, then later (after the RRD file has already been created) tried to set it to something higher. But the first max value is already in the RRD file. I see this with CPU graphs sometimes where I think the machine I'm graphing cpu for is a 4 processor machine (so I set the max value on it's cpu related data sources to 400) then later discover it's really a quad-core machine and not a dual-core machine. Increasing the max values on the data sources doesn't change the max values specified in the RRD file. I usually smack myself on the forehead a few times, then export the RRD file to an XML file, edit the maximum values in the XML file, then use rrdtool restore to turn it back into an RRD file. Or if I'm lazy I'll just delete the RRD file and let cacti re-create it.
Who is online
Users browsing this forum: No registered users and 0 guests