cisco switch config help

hello,

i've been tasked with installing a card swipe access system. Each door has a switch/controller combo which is wired to another door with a switch/controller combo then eventually to a cisco 3750 switch. So in essence, I have a series of loops all going to a main switch. I know this is not best practice, but unfortunately I have to work with the existing infrastructure. My issue is when a link in the loop goes down, I need to maintain access to the controller server.
I have 1 block of IP addresses to use, and there be approx. 200 of these switch/controller combos, one per door. Every IP device will have a static IP, so no DHCP. All switch/controllers will connect to a main cisco 3750 switch. The controller server is connected to this as well. Each controller communicates with the server.

I have a diagram to show how a portion of the network will look like to help describe the situation.

The switch/controller combo is a controller to manage door access and a switch to connect to the IP network. Ruggedcom switch I think.

The equipment was pre-purchased so I need to work with what I have.

I did use cisco packet tracer to "test" out this network using default configuration, and every time I broke a link in the middle of a loop, it took about 30 seconds to get a connection again. so my question is, how can i configure the cisco switch so that link from controller to server will be maintained via another "route" back to the main switch?

again....this is a physical loop...but i can't change it....sorry :(


network-layout.jpg
LVL 2
BrassMonkey26Asked:
Who is Participating?

[Webinar] Streamline your web hosting managementRegister Today

x
 
mstone76Connect With a Mentor Commented:
One more thing and I'll be done. Really need to get this all out the first time. Physical loops aren't bad because we have spanning tree. In fact they're an important part of design by creating a redundant path to the distribution devices.

Also there is no true way to get 100% of packets in the event of a failure. Even HSRP or VRRP between two routers has it's limits. Rapid spanning tree is your best option in this case. It will converge fast and most likely a person might have to scan their badge twice if a failure happens.
0
 
mstone76Commented:
What type of switches are at the doors? Sounds like something spanning tree could take care of automatically.
0
 
mstone76Commented:
Just noticed in your diagram your broken link. Thirty seconds is right on the money for STP. Try changing over to rapid spanning tree.. If the door switches are Cisco do the following in config.

spanning-tree mode rapid-pvst

That will give you rapid per vlan spanning tree. Convergence time is much lower. If this is still too long take a look at uplink fast. Might help a little.
0
The new generation of project management tools

With monday.com’s project management tool, you can see what everyone on your team is working in a single glance. Its intuitive dashboards are customizable, so you can create systems that work for you.

 
BrassMonkey26Author Commented:
mstone76:What type of switches are at the doors? Sounds like something spanning tree could take care of automatically.

The switches at the door are ruggedcom switches...specific model...not sure....and yah, i thought STP should do the trick here...

mstone76:   Just noticed in your diagram your broken link. Thirty seconds is right  on the money for STP. Try changing over to rapid spanning tree.. If the  door switches are Cisco do the following in config.

spanning-tree  mode rapid-pvst

That will give you rapid per vlan spanning tree.  Convergence time is much lower. If this is still too long take a look  at uplink fast. Might help a little.


Yah I thought RSTP would do the trick...but the problem is the ruggedcom switches....i don't think they have the capability of configuring them for RSTP....i'm on my way to work now...when I get there, I'll check the model number...see what these things can do...
0
 
BrassMonkey26Author Commented:
Ok, I got the model number for the switch/controller combo....it's a ruggedcom RS900. According to their website, the unit supports RSTP and Enchanced RSTP. I can test this on packet tracer since their "generic" switches still support RSTP...I'm going to check out the enchanced RSTP as well...I'm just trying to find a way to reduce downtime...I'll look at uplinkfast that you mentioned as well.

I'll keep you posted.
0
 
172pilotSteveCommented:
One other thing to make sure of is that the root of the STP is your core switch..  If you haven't set it, do so, otherwise if you're in a default configuration, it may be using a MAC address to decide which switch is root, which could be one of your remote switches...

Sorry - Not sure how to do this on RuggedCom..  :-(

-Steve
0
 
BrassMonkey26Author Commented:
yah I configured the core switch as root primary....enabled RSTP on all the switches...I did this because I verified that the ruggedcom supports RSTP as well....I tested the setup with a ping test from a controller to the server and would be down for about 4-6 seconds...I don't loose anything when I bring the link back up...I think this will be the best I can do based on this setup...

I just found out that I'll be adding another cisco switch for redundancy....I'll make the new core a root secondary. I would like to configure an 2-link etherchannel between the 2 cisco switches...

thoughts anyone?
0
 
mstone76Commented:
Port channels are great. They will give you a much faster fail over time than STP will because they're considered one link.
0
 
BrassMonkey26Author Commented:
with etherchannel, if I loose either of the physical links, will i loose the connection? Etherchannel is redundant, yes?
0
 
mstone76Commented:
For sure. You don't loose connection, just half the bandwidth. So if you have two 1 gig links between the switches then you logically have 2 gigs of bandwidth with a port channel. If you loose a link it just drops down to one gig but you're still up and running.
0
 
172pilotSteveCommented:
Definitely..  Sorry - Didn't realize that was an option!!  You'll likely only lose a ping or two in that case...  Just remember, your likely failure is going to be a remote switch going down, or someone accidentally cutting a wire, which breaks both connections in your etherchannel, so you're back to STP..  But, you're right, with the etherchannel, at least you're protected from a faulty port or cable...

-Steve
0
 
BrassMonkey26Author Commented:
yah the ruggedcom switches will be the faults as there is a lot of renovations happening in this building...but i think we can deal with 6 sec drop with RSTP....thanks a lot for the help :)
0
All Courses

From novice to tech pro — start learning today.