problem with the default memory usag template on a Debian VM

[DeeL]

Hi guys,

I've got a problem on my 0.8.7g running on freshly built debian vm (vmware workstation 7.0.1).

I've had the same problem with 0.8.7e and 0.8.7g, a funny behaviour of the "localhost memory-usage template".
free memory and swap keep going down. Once rebooted the graphs return to real values but as time passes by it just goes down...

The problem only occurs with this template, maybe the problem is not in cacti but in vmware, i will do a "top" soon to see if "top" values and cacti values are the same.

I join a graph image (jpg) and the technical support result page(pdf).

Regards,

Deel

[DeeL]

here is the result of a "top" command before rebooting

[DeeL]

here is the result of a "top" command after I rebooted.
This machine only runs cacti.

[DeeL]

And now the cacti graph. As i didn't wait long enough since the last reboot the difference is not really big but it's here and I don't get it.

[mxxcon]

i could be wrong, but don't "free memory" and "swap" are stacked on each other?
as you keep using your computer, your free memory decreases and so the graph will slide down..
if you'll look carefully you'll see that "swap" part of the graph remains the same but it's just "free memory" that is shrinking.

(for my servers i switched from using default load and memory graphs to these http://hans.fugal.net/blog/2009/01/10/s ... ith-cacti/)

[DeeL]

I hope you're wrong, it would mean that any pc is forced to reboot every week or so ^^, as i said I'm not supposed to use more and more memory as it's only running cacti with the same number of oids to check and as far as i know this template shows the memory usage i.e the way the box is taking and releasing memory.

No, i suppose it's a vm related issue, i'll first try to move the vm on an esxi server(if it fixes the problem then VmWare has a problem :p).

this new graph is based on the same template (0.8.7e) on another debian vm box running under Esxi, memory usage has been shown quite correctly for several weeks now.

[mxxcon]

wait a second, how much memory did you give that server?!
it has 1megs of memory free and 1megs of swap?
Attached is what my cacti server looks like

[DeeL]

After some re - thinking, yes the VM has 1.2M kB i.e 1GB Ram.

[knuwanga91]

I was wondering if you found a solution to this issue. I'm having the same issue with my cacti server running on a debian server

[camerabob]

A bit off topic, but I would change that graph template to display the swap at the bottom. That way only the other memory graph would change.Would make a quick peek easier to digest.