Hp 1920 24G Stack and Link Aggregation

[underscore99]

hi,

I´m conecting two 1920 swich, and i configure a Stack, and it works then i try tu set up a Link Aggregation between the two swich but this don't work and the stack stop wotking too.

The Stack configuration an the Link Aggregation can be use toogether?

Thanks for the help.

[Vince-Whirlwind]

That's an interesting question.

Personally, for me link aggregation would be the higher priority so I would do that first.

Is it worth trying to setup the network topology first and then afterwards try "stacking" the switches?

[parnassus · Accepted solution]

Hi AFAIK, the HPE OfficeConnect 1920-24G switch, see HP 1920-24G (JG924A), doesn't support "Stacking" in the sense of creating a single logical switch made up of a cluster of single physical switches (as can be done, as example, when stacking two, or more, HPE 2920 switches, each one with its dedicated Stacking Module + dedicated Stacking Cables, connected in a Ring or Chain stacking topology): that's the real Stacking known also as "Backplane Stacking" (Stacking at Backplane's level).

The HPE 1920 Series indeed just support "stacking (switches) for management purposes" (Stack Management) only and so you will be able to connect various switches togheter and then elect one specific member as a "Master device" (the Commander switch): that "Master device" will be able to monitor and to configure (centrally) connected "Slave devices" that were added as members of its Management Stack.

Trunking to other HPE 1920 Series switches happens usually through the definition of a LAG (Link Aggregation Group) using up to dedicated 4 ports per switch and let you configure Static (No Protocol) or Dynamic (LACP) trunking mode: this is supported and it has nothing to do with "stacking (switches) for management purposes". At least it should be independent by presence or absence of Management Stack feature discussed above.

The single port used (for Stack Management purpose) to connect to other other HPE 1920 Series switches can't be used for other purposes (Type: Trunk/Access)...so, basically, you should be able to define and add a LAG (on each involved Switch) with the same number of similar ports (1G) and so have LAG Trunking *and* Stack Management - at the same time - but not with the same involved ports.

If I'm not mistaken, No Loops (between LAG members ports and the Stack Management port on each involved switch) should form since Stack Management port defined on each Switch will be used only for management purposes (it's a lost prot for traffic purposes). Hope to be right here.

AFAIK a question can arise in the end: what's about first setting up LAG Trunks (using at least 2, 3 or at most 4 aggregated ports per switch) between switch pairs and then enabling Stack Management feature?

Will the Stack Management feature, once enabled in the Master device, require another intra-switch cable - used only for management purposes - between it and, as example, the first "Slave device" connected to it?

I think it will.