Wired Intelligent Edge

Hi all,

I am deaing with a situation where two 3810M "spine" switches are uplinking to Cisco. The spine(s) have a number of VLANs configured which are serving only downstream clients of the switches (ports 1-15, tagged on all VLANS).

Downstream, we have three 3810M(s), each uplinked with two ports (1/15,1/16) to the spine switch with all VLANs tagged on those two ports and the remaining ports on each switch untagged for client devices.

Port 16 on each spine switch is uplinking to a Cisco switch of unspecified type and it is set as untagged on only the VLAN which needs to uplink.

Aruba network is functioning perfectly. All clients on downstream switches maintain connectivity.

On the uplinks, For years, no collisions, no problems and connecitvity could be established through either spine. That is to say, if one spine went down or was powered off, connectivity would switch gracefully to the other spine.

All of a sudden, after a sequential reboot of both spine switches, colissions, dropped packets and loss of connecticity between Cisco and Aruba. What could cause this?

Here is the relevent config for the switches, each is set with a decending Spanning-Tree priority, with the Spine 0&1 and the client 2-4:

vlan 10

   name "Uplink"

   untagged 16

   tagged 1-15

   no ip address

   exit

spanning-tree

spanning-tree priority 0

---

@unsichtbarre wrote: Downstream, we have three 3810M(s), each uplinked with two ports (1/15,1/16) to the spine switch with all VLANs tagged on those two ports and the remaining ports on each switch untagged for client devices

Is each leaf Aruba 3810M connectect concurrently to both Spine Aruba 3810M? I mean - from what I've understood reading the text above - it seems that each leaf Aruba 3810M is concurrently connected with a two-links uplinks to the very same Spine Aruba 3810M switch...and not to both each Spine Aruba 3810M switches (let me say Leaf SW1 port 1 to port 1 Spine SW1 and Leaf SW1 port 2 to port 1 Spine SW2).

---

@unsichtbarre wrote: Here is the relevent config for the switches, each is set with a decending Spanning-Tree priority, with the Spine 0&1 and the client 2-4:

vlan 10

   name "Uplink"

   untagged 16

   tagged 1-15

   no ip address

   exit

spanning-tree

spanning-tree priority 0

---

OK but that piece of configuration just shows us that port 16 in Untagged member of VLAN 10 and port 1-15 are Tagged members of the very same VLAN 10...and the Switch where it is applied has STP Priority equal to zero (highest)...so that Switch was set with the role of STP Root...all in all it says almost nothing about your overall network topology and configuration...

Hi and thanks!

Turns out that root guard was not set upstream. We had been told it was. Issue is now resolved.

THX