Very basic newbie question

Post general support questions here that do not specifically fall into the Linux or Windows categories.

Moderators: Developers, Moderators

Post Reply
jimmo
Posts: 1
Joined: Tue Oct 19, 2004 5:50 am

Very basic newbie question

Post by jimmo »

Hi All!

I am extremely new to Cacti (I haven't even installed it) and I have a very basic question. I am familiar with rrdtool, and I already use it to create things like Oracle table space usage and certain performance data.

From what I read, Cacti is a "frontend to RRDTool" and is essentially passive. I know about the Cactid which collects data via snmp and that you can feed data into Cacti using various data sources. But the bottom line is that Cacti does not go out and collect anything. There are no "agents" that sit on the various nodes that the Cacti server queries for information.

One reason I am asking is that we are using HP OpenView and it's performance agents. There are agents on each node that can be queried by the server to provide the data the agent collects. My boss insists that Cacti works like that and we can simply replace the HP performance agents 1:1 (more or less).

I say that's not the case. Regardless of how the data is collected, it still needs some *external* process to get it into Cacti. For example, snmp. The problem for us is that on some platforms, not all of the values are stored in the MIB that we want. Although I have not loaded the HOST-RESOURCES-MIB onto all of our platforms, what I can see is that it does not have everything we would want to report (e.g. sys cpu usage versus user cpu usage).

Thus we would need some way to get this information from the various nodes to the Cacti machine. Our own MIB branch is obviously one idea, but difficult to manage (as far as I can see). Alternatively we could use something like scp to copy larger sets of data at regular intervals (e.g. once an hour).

Alternatively, we are testing Nagios which can excute remote command and retrieve data. Parsing the data from Nagios is an "issue" (not a problem), but I still see it similar to snmp in that we have to poll each of 100 machines at regular intervals reading out a couple dozens values.

In the end, not matter what method we use to collect the data, it is still external to Cacti, and **we** still need to figure out how to get it into cacti.

Is that correct? Am I getting it?

Regards,

jimmo
Post Reply

Who is online

Users browsing this forum: No registered users and 4 guests