We're trying to setup monitoring of our Nokia IPSO devices for memory usage. Unfortunately, Nokia only provides the ability to monitor memory usage on a per-process basis (which will always be dynamic OID's), rather than providing an OID for mem free/used/swap. This has meant we have to use a script that walks the appropriate tree, and adds the total value of memory used.
Sample script output:
bash-2.05$ /var/cacti/scripts/ipso_memory.tcl
337832
We have followed the section in the documentation on "simplest method from script to graph", located at http://www.raxnet.net/downloads/cacti/d ... graph.html in the Cacti Manual.
We've tried this several times now, and each time it results in rrdtool going into a zombie mode and the system load increases. The most we can get is the initial value into the created graph (indicated by Cur: 337.83k on the graph generated).
Question:
1. Is the documentation in the manual correct?
2. Is it correct but inappropriate for our example script?
3. Are there any other example howto's that mirror what we're trying to achieve?
Additional thoughts:
We've found it really difficult to figure out the interelations between the different steps when we create a simple script-based graph. In addition, as we have numerous hosts we want to monitor, we'd ideally like to put this into a template - but the howto's we've read and the documentation on this doesn't seem that clear (or do not match what we're trying to achieve).
However, overall we've found Cacti to be an excellent tool and we're now monitoring both traffic and CPU utilisation on our IPSO systems. Keep up the good work!
