Yes, I have also seen instances were stacking switches introduces an additional logical single point of failure. It's tricky, because stacking switches has many other advantages (like troubleshooting, reducing management overhead, and reducing the number of connections back to the core). But indeed, I've seen bugs where something in the management plane fails and then causes the dataplane to progressively fail, and in a stack scenario that failure is stack-wide.
I'm surprised that we are still managing switches the way we are in this day and age. In the wireless arena, everything is now centrally managed in some kind of controller or virtual controller based system. For a long time now, deploying a new access point has been plug it in, let it get a DHCP address and find the controller, then you provision it in the controller and you're pretty much done. Why this has not been the case for switching is beyond me.
In any case, standalone switches would still need to be physically swapped and the users still down in the mean time. I think the best strategy we're leaning toward is to deploy everystack with an N+1 switch that we just manually TREAT as a spare. We'll down all the ports so it can't be used, and in the event of a single switch physically failing we can SSH into the stack, make that member the member number that failed and let it pick up the config. Then we can tell someone on site to physically move cables 1:1 from the dead switch to the spare and at that point we're back up and able to wait until we can find time to come an properly remediate.
The wireless backup option for the endpoints is a good idea... but again, I think the additional wireless noise would kill the entire network. We'd need some way to keep the wireless NICs off and on standby until needed, but even then if the density is high the WLAN might not support all those wired clients in addition to the normal load of wireless clients.
Your switch vendor should have some type of central control that ou can use to maange all the swithes. If not there are 3rd party tools you cna use to backup configs, monitor uptime, and perform upgrades.
The rest, there are many ways to go about this, just depends on what your management will approve. I say offer them 2-3 options and let them make the choice, that way its on their heads. Except the central control of devices, that they should approve.
Central cloud managed switching has been arround for a number of years. But you need to step away from the Tier 1 vendors and look at things like Ubnt, Draytek, Meraki, etc..
For the situations where you have Quirky equipment. Configure specific ports on every switch for those items. if they don't support MDI-X Auto negotiation, then it's not going to support twin ethernet connections and LACP.
Click Accept as Solution to acknowledge that the answer to your question has been provided.
The button appears next to the replies on topics you’ve started. The member who gave the solution and all future visitors to this topic will appreciate it!
These simple actions take just seconds of your time, but go a long way in showing appreciation for community members and the Live Community as a whole!
The Live Community thanks you for your participation!