I wanted to know what things should be taken into consideration when deciding whether to create an L2 trunk between two distribution switches in a routed distribution layer.
I have 2 server farm access switches, each SINGLE-homed to one of the CatOS distribution switches and each server will be dual-homed. The distro switches are legacy CatOS/MSFC 6513 architectures, each running SRM with dual SUP 720s.
So, I will, of course, build an L3, routed link between distro switches to support redundancy, route summarization/aggregation and HSRP. I am wondering if it would be worthwhile to also create an L2 connection between the distro switches. What I am thinking is that it will be useful in the event that the routing engine in one of the switches dies,(lets say its in distro switch 1). Switch 1 will then be able to switch the traffic over a vlan that spans the trunk over to distro switch 2 for processing. No? I guess this is less of an issue when youre running high availability SUPs plus SRM mode in each switch.
Here's how I would look at this. What do you have to gain or loose by adding a trunk betwen the two distribution switches.
The servers are dual homed to two server farm switches. The two server farm switches are connected to two separate distribution switches. Therefore the redundancy for the servers exist all the way to the distribution switch. I would assume the distribution switch probably has two uplinks each to two core switches. In this case there's no single point of failure that can disable communication to the servers.
On the downside adding a trunk between the two distribution switches add more potential for STP problems. My experience suggests troubleshooting STP problems can get very ugly and could make routing issues way easier to troubleshoot.
It all depends on whether your access-layer switches are connected to each other, how they connect to the distro switches and how your server NIC's are set up ie.
1) You have a server that is dual-honed to two access-layer switches. Are these access-layer switches connected to each other ?.
2) The dual-honed servers. Are they in fault tolerant mode where one NIC is active and one is passive.
If the answer to both questions is yes then you have a problem.
Server A is in vlan 10 and the active NIC is connected to access switch 1
Server B is in vlan 10 and the active NIC is connected to access switch 2.
A wants to send traffic to B and so it needs to be L2 switched traffic as they are both in the same subnet. But there is no layer 2 path between your access switches.
Access switch 1 could send the traffic to distro switch 1 but distro switch 1 has no way of switching the traffic across to distro 2 or down to access switch 2 and it can't route it because the traffic is contained within the same vlan.
If your access-layer switches work as a pair and are connected via a L2 trunk and you route to the distribution layer then you are fine.
If your access-layer switches are dual connected to both distro switches with L2 trunks then you will be fine as well although Narayan is spot on about HSRP then having to run across the access switch links.
But if your access switches do not connect to each other, are singly honed to one of the distribution switches via L2 trunks and you have no layer 2 trunk between the distro switches i think you will have problems.
VMware Trunk Port Group is supported from ACI version 2.1
VMM integration must be configured properly
ASA device package must be uploaded to APIC
ASAv version must be compatible with ACI and device package version
In the Previous articles of ACI Automation, we are using Postman/Newman as the Rest API tool to automate the ACI Configuration.
In this article I’m going to discuss on usin...
One of the first steps in building your ACI Fabric is to go through Fabric Discovery. While Fabric Discovery is usually a straightforward process, there are various issues that may prevent you from discovering an ACI switch. This article wil...