• Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 826
  • Last Modified:

Connecting two switches to bypass the core

So I was reading through the HP Lefthand SAN best practices and they mention connecting the two iSCSI switches together. I was thinking through this in my head and this is going to create a loop and cause spanning tree to kick in a block one of the connections somewhere, right? The configuration I have is two Cisco 2960Gs with a mirrored config, both switches are setup the same. Half the ports are on VLAN1 for basic network connectivity to the servers and the other half are on VLAN2 for iSCSI.

The link in question is the green one. Is there anyway I could prevent STP from cutting one of these links? It doesn't seem like it would buy me anything if one switch looses it's connection to the core. Could I just remove VLAN2 from the trunks that run back to the core? I tried this in packet tracer and it still caused spanning tree to block a port but I never trust packet tracer. There is no need for VLAN2 to be accessible anywhere else Network diagram Network diagram
0
theconqueror
Asked:
theconqueror
  • 6
  • 6
1 Solution
 
jimmyray7Commented:
As long as you're not teaming NICs, you shouldn't have any loops.  Each NIC has one connection to the network via one switch.  I'm not familiar with the HP SAN setup, but I'm assuming that one of the controllers is on standby and only becomes active if the primary connection fails?
0
 
theconquerorAuthor Commented:
All of the HP NICs are teamed in bundles of two, but I don't see how that would have any affect on loops. Can you explain that a little more?

I know that there are going to be loops just from drawing it out. There are going to be three switches all connected to each other

As for whether the SAN is active active or active passive, I'm not totally sure, I was just told to setup the switches, but I'm pretty sure it's an active active setup
0
 
jimmyray7Commented:
Ignore my original comment, I was thinking of teaming at the OS level, not vSphere.

Take a look at this to see if it helps - http://www.vmware.com/files/pdf/virtual_networking_concepts.pdf

Is the link between the switches 1 cable or a LAG?  As long as it isn't multiple links, STP shouldn't be an issue.
0
Free Tool: IP Lookup

Get more info about an IP address or domain name, such as organization, abuse contacts and geolocation.

One of a set of tools we are providing to everyone as a way of saying thank you for being a part of the community.

 
theconquerorAuthor Commented:
I still don't see how teaming in either the OS or vSphere would have any effect. For the sake of the argument lets just say that there are no end nodes connected to these switches.

The link between the switches will be an etherchannel
0
 
jimmyray7Commented:
There shouldn't be any loops for STP.  There is only one connection between each NIC and each switch.  STP won't disable anything.
0
 
theconquerorAuthor Commented:
Are you saying that STP only affects links between servers and switches?
0
 
jimmyray7Commented:
I'm saying there aren't any loops in your diagram.  Each physical link is only connected to one switch.  Have you tried this setup yet?  It should work.

VMware's networking infrastructure eliminates any loops at the virtual switch level, so you don' t have to worry about the VMs, just the physical links.

If you have implemented this and are having STP problems, what link is being shut down?
0
 
theconquerorAuthor Commented:
Ok, so maybe this confusion is because I used the Visio's router picture instead of a switch. The router in this setup is a layer 3 switch. If that's the case, I apologize.

Yes I have set this up in a "lab" (packet tracer) and STP was blocking ports, either of the two uplinks that run back to the core
0
 
jimmyray7Commented:
Ah, I see what you mean now.  With one router/L3 switch at the top, STP will block one of the routes.  That shouldn't be an issue though, if one of the switches dies, STP should switch to the other connection.

In our environment, which is very similar, we avoided this issue because we have two routers in a high-availability situation.  I just plugged the primary router into one switch and the backup into the other.  If one of the switches dies, the backup router will take over.

Is your concern with one link bandwidth?  If so, can you create a LAG/etherchannel to the L3 switch?
0
 
theconquerorAuthor Commented:
My concern isn't really about link bandwidth, I just don't really see the point of it, but yet HP puts it in their best practices. It's just adding complexity, everything is already redundant without that link so, why bother?

That's when I started thinking, since this VLAN is pretty much only going to live on these two switches maybe I could create that link between the two if I could somehow isolate the VLAN and prevent it from going back to the core but I'm not so sure how to do that
0
 
jimmyray7Commented:
Well, depending on your ESX networking setup, you could have a lot of data traversing your network, and running traffic over the LAG is much preferred over running alll the way up to the router/L3 switch and back over one link.  Especially if you're doing vMotion, Fault Tolerance or any other high bandwidth stuff.

Does that make sense?
0
 
theconquerorAuthor Commented:
Didn't really get an answer but awarding points for the effort
0
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

Join & Write a Comment

Featured Post

Free Tool: IP Lookup

Get more info about an IP address or domain name, such as organization, abuse contacts and geolocation.

One of a set of tools we are providing to everyone as a way of saying thank you for being a part of the community.

  • 6
  • 6
Tackle projects and never again get stuck behind a technical roadblock.
Join Now