04-14-2023 03:31 PM
We recently deployed some VSP 4900-48P pairs for Top of Rack use. They have been configured, IS-IS is working, vIST is working, there are adjacencies with the VSP Core switches (7432's), remote management is working and we can SSH to the switches.
The problem is that when we connect an end device (access port, no trunk ports thus far), it simply is not passing traffic. We get a mac address on the 4900 and on the core, the end device get's its IP address, but is unable to pass any traffic.
I have connected my laptop to the switch, set a static IP in the subnet for the vlan on that particular port and I can see the mac address on the 4900 and on the core, but cannot ping.
I do need to add that the network is in a state of transition, so the L2 traffic is being handled by the Extreme hardware, but the L3 still resides on the Cisco core. However, we are also seeing the mac address for the end device and my laptop there as well and as mentioned, do get an IP address. We have tried clearing arp entry with no change. There is nothing in the log file indicating any issues
I am not near the hardware currently, and am still very new to Extreme, but thought I would post here and see if I could get some pointers for troubleshooting the issue for when I am back at it next week. The local Extreme Networks Engineer did look over the configuration and did not see any issues and did try troubleshooting the issue as well but neither of us could find anything out of the ordinary or that would point to a possible cause.
I have configured and successfully migrated other servers/end devices in another data center but that was using 7400's as the ToR switches, so I am not sure if there is some odd configuration parameter missing because it is a 4900, or if its a hardware issue (which does not seem likely), etc.
Any help or pointers would be much appreciated.
04-15-2023 11:15 AM
Hi RedDirt,
Just for verification, did you create in the same way your VLAN/L2VSN services on each vIST cluster's member ?
Regards.
Jave
04-15-2023 07:16 PM
Hello @Jave
Thank you for your reply. It is just a two switch cluster, but to answer your question, yes. The switches have a NNI to the core switch, the vIST between the two switches, and a NNI link to another two switch deployment in another rack for Top of Rack use. The vlans on each switch are not always identical, that is to say, vlan 100 for example may not be created on each switch as the vlans created depend on the end devices, but the vlans are all created in the same way. For example:
vlan create 100 name "GenericVlan" type port-mstprstp 1
vlan members 100 1/10,1/12,1/19 portmember
vlan i-sid 100 123456
slpp vid 100
The vlans created all follow that syntax, though the i-sid and name will be different as above is just a sanitized example. Then the ports would be configured in a manner like below:
interface gigabitethernet 1/10
name "Generic Access Port"
no shutdown
slpp packet-rx
slpp packet-rx-threshold 50
no spanning-tree mstp force-port-state enable
y
no spanning-tree mstp msti 1 force-port-state enable
exit
04-16-2023 01:06 AM - edited 04-16-2023 01:09 AM
Hi @RedDirt
I'm not sure to well understand your setup, but in all case, VLAN/I-SID database MUST be the same on two members of a vIST cluster, otherwise you'll probably have traffic issues. Indeed, some packets toward an UNI device attached on one cluster's member can pass through the other one, depending on SPBm unicast-tree: if VLAN/I-SID service does not exist on it, traffic is unable to reach UNI device through vIST link.
Regards,
Jave