Web and Unmanaged
1755994 Members
3348 Online
108839 Solutions
New Discussion

Re: HP V1910-48G stacking configuration

 
Doomus
New Member

HP V1910-48G stacking configuration

Hi team.

I have three HP V1910-48G switches. I want configure stacking on this switches, but i have some questions.

- Could i connect stacking switches in ring, i.e. connect SW1 with SW2, SW2 with SW3 and SW3 with SW1?

- Which load-balance algorithm working on stack ports?

- Will be a data loss during stack activation?

1 REPLY 1
parnassus
Honored Contributor

Re: HP V1910-48G stacking configuration

Hi, I'm not going to speak about Stacking for (centralised) Management here, this just to be clear.

The Switch type you have doesn't stack and you can't create a single logical Switch made up of stacked members, instead you can interconnect them using trunking uplinks.

Speaking about trunking two (or more) Switches using two (or more) aggregated ports (member ports) to enhance resiliency and throughput of the link that interconnect them (it's called Trunking or Bridge Aggregation or Link Aggregation in HP jargon, EtherChannel in CISCO jargon) you should then avoid any loop.

If you try to develop a Ring Topology (so you deliberately create a physical Loop) with your three Switches you're going to experience severe broadcast storm and network traffic disruption (at least until you kick in something like MSTP instead of STP/RSTP).

If you need to interconnect two (or more) Switches using HP's LAG (Link Aggregation Groups, generally with up to 8 member ports per Group as the maximum threshold) and adopting Static (non-protocol) or LACP (IEEE802.3ad protocol) then you should plan either a Daisy-Chain Topology (where Switches are "cascated" togheter) or a Star Topology (where a core Switch collects all Trunks from your edge Switches), clearly the best choice is dependant on your scenario, your requirements and the number (and role) of involved Switches and the availability/type of Physical Links between them.

With similar Switches (from the same vendor, same models and with the same System Software version) a good choice is to use LACP Dynamic. It's pretty straightforward and runs very well.

If you're starting from three Switches yet inter-linked with single links then you need to first define LAG, Members and Static/LACP on each Switch, disable running links and connect physically (or enable if disabled) involved LAG ports. A minimum of disruption is going to happen.

If you're starting from three Switches not yet inter-linked then you shouldn't worry about any traffic disruption since there isn't any traffic between them yet.

See the Chapter "Configuring link aggregation and LACP" on page 198 here.


I'm not an HPE Employee
Kudos and Accepted Solution banner