Netonix Forums

Would also love to see power use (total) !
thx

josh wrote:I just want to say that I ran this (tiny) MIB through every validator I could find, and it passed with flying colors.

Which, is something UBNT hasn't been able to do with 5 bug reports now :)

EXCELLENT WORK! =]

Sidenote: If you were to say, kick http://www.observium.org your MIB (and netonix logo) and maybe $1,000 or so, you could probably get them to code in support for your switches. It would be really nice to have support for them in a proven wide-spread project. You might even be able to advertise that feature.

How would one go about using this MIB with LibreNMS? I've been searching Google for days for a solution without any luck....

Thanks,
Sam

Samgoober wrote:
How would one go about using this MIB with LibreNMS? I've been searching Google for days for a solution without any luck....

Thanks,
Sam

Have you sent the MIB to a LibreNMS coder ?

FYI, I added support to LibreNMS for Netonix Switches last month. Please open an issue on their github if you have any problems.

Also, as of firmware 1.3.9 the version has a little bug where it is reported as "1.3.9.n.........".

tonym wrote:Also, as of firmware 1.3.9 the version has a little bug where it is reported as "1.3.9.n.........".

I'll fix this in the next release.

CPU Usage seems to have an issue as well.
In the gui usage fluctuates quite a bit, but it seems to stay at 20% according to SNMP.
UCD-SNMP-MIB::ssCpuUser.0 = INTEGER: 10
UCD-SNMP-MIB::ssCpuSystem.0 = INTEGER: 8
UCD-SNMP-MIB::ssCpuIdle.0 = INTEGER: 80

(and you guys already mentioned sysUpTime being wrong)

tonym wrote:CPU Usage seems to have an issue as well.

The CPU is only used to drive the UI/CLI and Daemons such as SNMP and stats collections. All packet switching is handled by the switch core which has no affect on the CPU utilization as it has its own specialized and dedicated CPUs.

Any time another user logs into the UI or CLI or the switch is dealing with SNMP walks, SMTP events or such the CPU utilization will go up.

The CPU is only a little utility MIPS 24K 416 MHz specifically to drive the UI/CLI/Daemons so it will fluctuate when being accessed (UI/CLI/SNMP/ETC) but as I said has nothing to do with packet switching.

You misread my comment. I'm well aware the CPU doesn't affect the packet forwarding or vice versa.

To be more explicit, the SNMP reporting of CPU usage seems to have an issue. It always reports the same values. Either it is a stub implementation, or it is using some sort of running average.

tonym wrote:You misread my comment. I'm well aware the CPU doesn't affect the packet forwarding or vice versa.

To be more explicit, the SNMP reporting of CPU usage seems to have an issue. It always reports the same values. Either it is a stub implementation, or it is using some sort of running average.

I pinged Eric, he should respond later today

I found a bug in the version of net-snmp we are using that prevents the CPU counters from rolling over when they get too large, this causes the CPU usage to get "stuck". It'll be fixed in the next release.