Redundant Switches

[Alberto2]

Hi all,

 I need to implement a couple of swithces 2530 in redundancy. I would like to assign to couple ports (23 and 24) for redundancy link. The servers and workstation implement the LACP on NIC ethernet ports. What is the suggested switch configuration for my scope?

Thanks in advance,

Alberto

[parnassus]

Port Trunking is the magic word.

If you need to interconnect your two Aruba 2530 Switches Port Trunking on both ends is required: using two (or more) physical ports aggregated together in a single LAGG (Link AGGregation Group) set up to use LACP should suffice...doing things that way you will be able to enhance the interconnection's resiliency (and also obtain a throughput enhacement too for particular inter-Switch traffic conditions).

Refers to "Port Trunking" on Aruba 2530 documentation.

If your Hosts (Servers/Clients) are yet using Teaming (that's also port trunking with/without LACP) you should pay attention that doing Port Trunking between your two Aruba 2530 doesn't mean you can terminate each (Server/Client) Team's member link into different Switches (to do that you should do something called Distributed Trunking)...but you can connect them to just one of two Switches.

This limitation is due to the fact that Port Trunking doesn't create a Virtual Switch as VSF (Virtual Switching Framework) does so every (Server/Client) Team's member shall co-terminate on one Switch (co-terminus) or the other.

[Alberto2]

Hi, 

 first of all thank you for the answer.

Here I detail my request, the servers and worsktation has been configured with the HP Network Fault Tolerant only (NFT), and here attached the network scenario that I had implemented.

So considering this the Port Trunking should be the solution that you suggest?

Best Regards

[parnassus]

If Servers/Workstations use [*] NFT (Network Fault Tolerance) it means that those hosts use just one Active port (called Primary) of the team to transmit/receive network traffic to/from the network, other remaining teamed ports (called Secondary/Non-Primary) are left in Standby mode, idle until the Primary port eventually fails.

Said so and considering that - probably - your Aruba 2530 Switches have (R)STP yet globally enabled (and so Fast Port Span is globally enabled too) the actual connections (Team's member ports that span to both Aruba 2530 Switches as per attached diagram) should work because each Secondary/Non-Primary link of every Team will be automatically and immediately [**] blocked (not causing a reconvergence calculation by the RSTP) by the RSTP until the Primary member ports eventually fail.

Port Trunking is the only option you have (apart the single link uplink) to interconnect/uplink your two Switches Aruba 2530 giving the interconnection link a basic resiliency (against port/cable failures) and enhanced throughput (in case of inter-Hosts traffic flows of type: many-to-many or many-to-one...from Hosts on Switch A to Hosts on Switch B).

[*] as per NFT related documentation:

NFT is deployed in environments that only require fault tolerance and do not require transmit or receive throughput greater than the capacity of the Primary Adapter (e.g., a server that requires fault tolerance in case of a network adapter malfunction, but does not have a demand for receiving or transmitting more than the capacity of the Primary adapter.)...Server's Primary port throughput will define and set the throughput limit due to the NFT adopton (there isn't balancing, just fault tolerance).

and also:

Team members can be split across more than one switch in order to achieve switch redundancy. However, all switch ports that are attached to members of the same Team must comprise a single broadcast domain (e.g, be on the same VLAN). Additionally, if problems exist after deploying a Team across more than one switch, reattach all Team members to the same switch. If the problems disappear, then the cause of the problem resides in the configuration of the switches and not in the configuration of the Team. If switch redundancy is required (Team members are attached to two different switches), then HP recommends that the switches be deployed with redundant links between them and (R)STP be enabled (or other Layer 2 redundancy mechanisms) on the ports that connect the switches. This helps prevent switch uplink failure scenarios that leave Team members in separate broadcast domains.

[**] because Fast Port Span feature disables STP’s blocking, listening and learning stages on involved ports.

Old but good: read here (even if it doesn't show the case you're facing...where Switch A and B are uplinked through Port Trunking LACP...which is correctly managed by the (R)STP). Probably you will be able to find something more updated...but the basics are there anyway.

[Alberto2]

Just to select better my situation, I confirm that all teamed NIC's serever has been connecteted in two separate switch and configured on server as NFT , the switched has been connected with a couple of link (23-24) and I had enable the "spanning-tree".

My question is I had to add some instruction on the configuration below? trunk? or something else?

Thanks so muche

Alb

hostname "SWITCH-A"
timesync sntp
sntp unicast
sntp 30
sntp server priority 1 ....
ip default-gateway ....
vlan 1
name "DEFAULT_VLAN"
untagged 1-28
ip address ....
exit
spanning-tree

hostname "SWITCH-A"
timesync sntp
sntp unicast
sntp 30
sntp server priority 1 ....
ip default-gateway ....
vlan 1
name "DEFAULT_VLAN"
untagged 1-28
ip address ....
exit
spanning-tree