Aruba & ProVision-based
cancel
Showing results for 
Search instead for 
Did you mean: 

VSF + LLDP MAD

 
Dr1zzt
Occasional Contributor

VSF + LLDP MAD

We have linked two Procurve 5406 switches (v3 modules, FW 16) with VSF, created the stack and everything is working as intended.

I now want to create the LLDP-MAD link to prevent a split situation but I'm not exactly sure how to do it.

I know I need an assist device to be linked (trunked) to both members and a simple command which shows the stack the IP address and the community string of the assist device.

HP-VSF-Switch(config)# vsf lldp-mad ipv4 10.0.0.1 v2c public

The question is, do I need to use a dedicated switch as an asset device? The VSF stack is linked with 13 switches with trunks, so using a second trunk in one of those (for the MAD link) could possibly lead to a loop?

2 REPLIES
parnassus
Honored Contributor

Re: VSF + LLDP MAD

Interesting question. Personally I don't know even if I think it could be a best practice to use a separate ad-hoc switch for this task since MAD "assist" device (apart other requirements specified on Paragraph 19.19.6, see below) and VSF devices pair must be all on a common IPv4 Subnet.

If so, it would also be interesting to know if - actually (latest Firmware) - a little HPE 1920 series switch (Comware 5) could act as MAD "assist" device.

Also will be interesting to know the relationship between LACP-MAD and LLDP-MAD since both techniques play a role (and are cited) when a VSF stack is deployed.

As reference to VSF and LACP+MAD and LLDP+MAD see Paragraphs 13 (LACP-MAD) and 19.19 (LLDP-MAD) of "HPE ArubaOS Switch Management and Configuration Guide for K/KA/KB.16.01" (March 2016).

parnassus
Honored Contributor

Re: VSF + LLDP MAD


parnassus wrote: If so, it would also be interesting to know if - actually (latest Firmware) - a little HPE 1920 series switch (Comware 5) could act as MAD "assist" device.

After digging a little bit I found that it looks like the HPE OfficeConnect 1920 Switch series doesn't support MAD passhthrough...

Edit: finally found an interesting Configuration Guide recently published about VSF. Details here.

On page 19 there is a nice "VSF HA topology with LLDP-MAD" configuration example (an Aruba 2920 is used as MAD).

Two statements I didn't understood very well when reading the paragraph "Best practices and configuration notes" on Page 23, it's written:

  • For High Availability applications, it is recommended to trunk ports across VSF members on different modules (so the guide suggests to terminate each incoming Trunk - here I mean BAGG - within the same unit - no matter if it is Commander or Standby - but do so using ports located on different unit's modules <- this is to prevent disruption on module fault).
  • VSF is not compatible with Distributed Trunking and Meshing (so the guide states that an incoming Trunk - again here I mean BAGG - can't be splitted between two units, so between Commander and Standby).

Does this really mean that an access Switch connected to the VSF through a BAGG (made of two or more aggregated ports) must terminate all its BAGG members' links on one and only one VSF member without the chance of distributing/sharing BAGG's member links to both VSF members?

The only permitted scenario is when a BAGG terminates inside the same VSF member (eventually on different Modules, as suggested).