Wired Intelligent Edge

Hi all!

I'm having trouble trying to correctly configure Aruba NetEdit 2.0 (2.0.2 actually) to discover an ArubaOS-Switch device hosted on a subnet different from the one where the Aruba NetEdit appliance actually runs.

Aruba NetEdit 2.0 is on subnet A and it was able to discover VSX nodes hosted on its subnet...no issues here (a set of credentials was defined for Subnet A) but it logs as unreacheable any device hosted on a routed subnet B reacheable by subnet A (clearly using another set of credentials for that subnet: subnet B and ArubaOS-Switch SSH/SNMPv3 credentials).

From the Aruba NetEdit appliance console successful ping tests against the ArubaOS-Switch devices on subnet B provide good results...so those devices, from the Aruba NetEdit, shouldn't be exactly reported/logged as "unreacheable" IMHO.

Hi, 

we had the same issue.

I guess you are allready familiar with the seed address.(one ("the first") ArubaOS-CX device at the far end subnet)

Even that wasn't working in all situations for us.

What realy helped was a 1:1 device adding process

So at the device adding form:

Subnet: deviceIP/32 with your related credential

Seed Addresses: deviceIP of the particular device 

We repeated that procedure for each device. Takes some more time but works 100%.

(in some cases adding with a larger target subnet NetEdit added some of the devices but not all)

Regrads

Hi @Rene_D!

I'll for sure will give a try using your /32 approach! Great suggestion

Pointing the Aruba NetEdit directly against the host IP address using your suggested workaround (device IP/32 as Managed subnet + device IP as Seeder) could be a way...but  instead the Aruba NetEdit should correctly using the host's subnet (plus Seeder) otherwise the presence of Managed Subnet and Seed Address parameters (and the discovering concept they were introduced for) will become a little bit useless.

Personally I expect that defining various Managed Subnets should make the Aruba NetEdit capable of contacting hosted devices whitin that subnets (through their respective seed addresses)...and not that each single device must be manually defined...if so we are in front of a "manual targeting method" not a "automatic discovering method".

Just to make sure: 

if you configured your other managed subnets where your AOS-S devices are located to, with snmp credentials, it should discover them.

It looks like Aruba NetEdit is not able to discover them...my scenario is:

• NetEdit is on VLAN 2250 (10.255.250.250)
• A devices subset is hosted on the 1st Managed Subnet on the same VLAN 2250 (10.255.250.0/24) where also NetEdit resides.
• ArubaOS-CX devices on that 1st Managed Subnet are correctly discovered and managed.
• 2nd Managed Subnet is defined on a VLAN 3126 (10.140.126.0/24)
• VLANs are both routed by the very same core (ICMP Ping tests are good from/to switches hosted on those two VLANs)
• On 2nd Managed Subnet there are mostly ArubaOS-Switch NOS based devices
• Associated credentials were defined (SSH+SNMPv3) on both Managed Subnets (they are different)

If I add a device (device discovery) a switch on VLAN 3126 discovery process starts correctly with the log message "Discovery of device '10.140.126.117' has begun." (note that the 10.140.126.117 is the seeder IP Address of a device actually managed by HPE IMC via SNMPv3 so I'm pretty sure the SSH+SNMPv3 credentials defined on NetEdit are OK apart not using REST for ArubaOS-Switch based devices [*]) but immediately the message "Discovery classification failed for device '10.140.126.117' since the device is unreachable." is logged. Device addition fails because device discovery fails.

[*] hope the REST credentials lacking is not the cause here...

Just to make sure you are using ArubaOS Devices (meaning nonArubaOS-CX) like the 2900/2500/5400 on your 2nd MGMT subnet?

Because NetEdit works only with ArubaOS-CX Devices like 8300/8400 or the very new 6300/6400.

Hi @Rene_D! Yes, I'm using ArubaOS-Switch based devices (such as Aruba 5400R zl2 and Aruba 2540) on defined 2nd Management Subnet.

Are you really sure it works only with ArubaOS-CX based devices...I mean...it should at least recognize ArubaOS-Switch based devices...I don't pretend that Aruba NetEdit is able to manage or configure them...just recognize them and build relevant network topology.

See what I mean on some screenshots here and look at how ArubaOS-Switch based devices appear on the network topology build around the VSX.

Am I missing something?

Hi Rene,

With NetEdit 2.0, it is possible to add ArubaOS (29xx, 54xx) with SNMP for monitoring...

Has anyone gotten snmpv3 to work?  I've only gotten snmpv2c to work as expected.

Do you mean SNMPv3 against ArubaOS-Switch based switches or against ArubaOS-CX based switches? the latter works (!)...the former...I personally was unable to configure my Aruba NetEdit 2.0.2 to discover my SNMPv3 enabled ArubaOS-Switch based switches (never tried SNMPv2 because I don't use it).

Yes correct, SNMPv3 against ArubaOS-Switch based switches.  I switched it to SNMPv2c and then it discovered fine.  Otherwise it kept showing up as unreachable in the diagnostics tab.  This is a lab and not production.

That's great...then we're just hitting a bug (I didn't go down that route since all my switches use SNMPv3).

I also face some issues with discovery:

1# I have two 2930f on the same subnet; same snmp config; NetEdit discovers one of them, the other always stays at "unreachable" and pending discovery

2# NetEdit tries to connect to every single IP address it detects on the seed devices even if they are not members of any discovery subnet (ex: subnet = 10.1.1.1/32; seed = 10.1.1.1; NetEdit will try to connect to any IP it finds on 10.1.1.1, like  ex 10.10.10.10)

---

@ricardoduarte wrote: (ex: subnet = 10.1.1.1/32; seed = 10.1.1.1; NetEdit will try to connect to any IP it finds on 10.1.1.1, like  ex 10.10.10.10)

---

Well, that's the strange part.

If a Managed Subnet is defined with the net 10.1.1.1/32 IP address it means that that Managed Subnet is exactly equal to just one single host.

Then setting the 10.1.1.1 IP address as the Seed(er) address will mean that the Aruba NetEdit discovery process should start discovering that very host (the 10.1.1.1) only and nothing else at all...because it should be locked at looking for that single host only.

It looks like the discovering process is not honouring the subnet mask assigned to defined Managed Subnet.

Here is a slide that better explains some aspect: mgmt address provided in LLDP TLV is important item (it should be reachable)

Also, if you put a seed, topology will start from this seed.

Once the topology is discovered from this seed, there is no further "subnet scanning" inside the managed subnet. Discovery is made from neighboring discovery, not scanning.

Hi,

But I can see, under NetEdit logs, that it is trying to connect to all devices it finds under the "seed" device, even tho they are not part of the managed subnets.

Regards

Hi,

The SNMPv3 issue should be solved in next version of NetEdit 2.0.7

/Peter