Port speeds are often incorrect on small vendor’s implementations. I’m not sure why. These are reported by the device in ifSpeed and ifHighSpeed, and there’s not much we can do if the device is reporting the wrong speeds. You can ask the vendor about this, they should probably fix the ethernet ports. The wireless interfaces are more troublesome, since there’s the “theoretical” maximum of the technology, and the current negotiated speed, which changes frequently. I don’t have much of an opinion on what should be reported here, but you’ll find that many vendors will just report some placeholder.

 

It seems the data you’re referring to is being polled by some super old and sort of weird legacy code :

 

includes/polling/wifi.inc.php:    $wificlients1 = count(explode("\n",snmp_walk($device, "entCLInfoIndex", "-OUqnv", "SENAO-ENTERPRISE-INDOOR-AP-CB-MIB")));

 

It seems to be counting the number of entries returned in that table. This is really old contributed stuff, so I’m not sure who added it. I’m not familiar with SENAO stuff, so it’s also not immediately obvious what would be a better way to do it.

 

Adam.


 

From: observium <observium-bounces@observium.org> On Behalf Of Serge Caron
Sent: 09 May 2018 13:32
To: Observium <observium@observium.org>
Subject: Re: [Observium] SENAO (Engenius) support for dual band products

 

Hello Adam,

 

Thank you for your answer.

 

I don’t think things are as bleak as it looks. Below is an “inventory” of an ECB1750 access point serving 6 VLANS over two bands (2.4/5 GHz): observium is picking up all ports even if this device does way more than the old Engenius “hockey puck”.

 

As you point out, data needs to be interpreted correctly. In the case of dual band units, there will be two fields for wireless clients (one for each band) and all observium outputs/graphs needs to be adjusted accordingly. I overlooked this reality before asking my question.

 

In the case of this ECB1750 device, all ports are reported as “Ethernet 10 Mbps” even though the wired port is connected at 1 Gbps in the switch and each radio is “capable” (the quotes are from me as emphasis on shameless marketing) of 450 Mbps on the 2.4 GHz band and (OMG!) 1.2 Gbps on the 5 GHz band. It is these things that look “broken” in observium.

 

In essence, I am getting most of the stats I need as things are.

 

Regards,

 

Serge Caron

 

 

Inventory of sample ECB1750 device:

 

#

Description

Type

Status

Errors

Load

768

Processor

hrDeviceProcessor

running

0

12

1025

lo

hrDeviceNetwork

running

0

1026

ifb0

hrDeviceNetwork

down

0

1027

ifb1

hrDeviceNetwork

down

0

1028

eth0

hrDeviceNetwork

running

0

1031

br-1

hrDeviceNetwork

running

0

1032

eth0.401

hrDeviceNetwork

running

0

1033

br-2

hrDeviceNetwork

running

0

1034

eth0.402

hrDeviceNetwork

running

0

1035

br-3

hrDeviceNetwork

running

0

1036

eth0.403

hrDeviceNetwork

running

0

1037

br-4

hrDeviceNetwork

running

0

1038

eth0.404

hrDeviceNetwork

running

0

1039

br-5

hrDeviceNetwork

running

0

1040

eth0.600

hrDeviceNetwork

running

0

1041

br-99

hrDeviceNetwork

running

0

1042

br-lan

hrDeviceNetwork

running

0

1043

eth0.172

hrDeviceNetwork

running

0

1044

wifi0

hrDeviceNetwork

running

85513

1045

wifi1

hrDeviceNetwork

running

0

1046

ath1

hrDeviceNetwork

running

0

1047

ath11

hrDeviceNetwork

running

0

1048

ath12

hrDeviceNetwork

running

0

1049

ath13

hrDeviceNetwork

running

0

1050

ath14

hrDeviceNetwork

running

0

1051

ath15

hrDeviceNetwork

running

0

1052

ath0

hrDeviceNetwork

running

1956

1053

ath01

hrDeviceNetwork

running

0

1054

ath02

hrDeviceNetwork

running

0

1055

ath03

hrDeviceNetwork

running

0

1056

ath04

hrDeviceNetwork

running

0

1057

ath05

hrDeviceNetwork

running

1684

 

 

De : observium [mailto:observium-bounces@observium.org] De la part de Adam Armstrong
Envoyé : 8 mai 2018 17:25
À : 'Observium'
Cc : 'Observium'
Objet : Re: [Observium] SENAO (Engenius) support for dual band products

 

MIBs do not affect the data the device reports, they just make it easier to identify and work with the data.

MIBs are like DNS, all they really do is provide a hierarchy, a name<>number translation and sometimes some hints about the format of the data returned. They're not really useful by themselves in enabling support for new things, like a driver would be. They're rather more like the technical specification used to write the driver.

It's possible they have moved this data to a new location on newer MIBs, or added new OIDs with more useful data. It's strange they'd be returning erroneous data in old tables though, but tis difficult to say without examples.

Adding support for new things requires writing code and/or definitions for the new data. This isn't a zero-effort thing, and usually occurs based on greatest demand (or commercial sponsorship).

Mike/Tom: I think we should come up with a generic "new stuff" template form or something, and include the "mibs don't do shit" boilerplate in the mailing list sign-up process somewhere. I think this would go some way to mitigating these open ended things.

Adam.

Sent from BlueMail

On 8 May 2018, at 21:15, Serge Caron <scaron@pcevolution.com> wrote:

Greetings!

 

Quote from Adam :  “The archives died during a server move, and the mailing list software is voodoo, so they never came back”. Apologies for asking something that is probably old news.

 

The Community Edition ships with Senao (Engenius) MIBs dated 2013.11.25.

 

This is fine for older devices such as the EAP9550 and we get proper data (such as number of wireless clients).

 

However, it seems all modern dual band devices such as the (older) EAP600 and the (newer) ECB1750 always report 1 as the number of wireless clients (as well as other broken stats).

 

Are there updated MIBs compatible to observium for these products?

 

Regards,

 

Serge Caron

 


observium mailing list
observium@observium.org
http://postman.memetic.org/cgi-bin/mailman/listinfo/observium