Fibre Channel (SAN)

Reply
Contributor
will.mowat1
Posts: 26
Registered: ‎06-21-2011

Implementing ISL using virtual switches

Hi,

Wonder if somebody coud shed some light on the following issue.

We have a DCX and a 48000 that are currently linked up using two trunk groups. Each trunk group consists of two physical links. So in total we have 4 links that are the ISL between the two switches.

We are wanting to do the following, and we are unsure whether it is configuration which would work:

1.Create a virtual switch on the 48000

2.Create a virtual switch on the DCX

3.The virtual switch on each switch will consist of 4 ports

4.Run cabling between the ports on each vswitch

5.Turn the 4 links into an ISL

The end goal is to get the new virtual switches to become the ISL for the physical switches. However, the virtual switches are seperate entities in their own right, so could we merge the virtual switches into the same fabric and then make them the ISL?

Valued Contributor
TechHelp24
Posts: 3,634
Registered: ‎02-23-2004

Re: Implementing ISL using virtual switches

Virtual Fabric is not available on 48K Plattform

TechHelp24
Contributor
will.mowat1
Posts: 26
Registered: ‎06-21-2011

Re: Implementing ISL using virtual switches

OK,

What if we created the virtual switch on the DCX and plugged the links into the 48k and then made it an ISL?

Valued Contributor
TechHelp24
Posts: 3,634
Registered: ‎02-23-2004

Re: Implementing ISL using virtual switches

since the 48K is already ISL-Trunking to DCX, from my point of view make no sense to create VF on DCX and then ISL the VF to 48K

TechHelp24
Contributor
will.mowat1
Posts: 26
Registered: ‎06-21-2011

Re: Implementing ISL using virtual switches

Thanks for the info, the reason I ask these questions is because this happened recently:

One of the SFPs one of the links that forms one of the ISLs failed. This reduced the TX/RX bandwidth on that ISL to 8gbps. This then created a bottleneck on the link. We would have expected the other trunk to take the workload, but it didn't. This affected service quite badly and a storage array sent some continuos access volumes into an error replication state.

So I suppose the question is, why didnt the other trunk ISL link take over the workload, when the other dropped a link?

If there are two seperate trunks forming two seperate ISLs, will one take the wokrload of the other, if the bandwidth drops to half on the faulting ISL?

Join the Community

Get quick and easy access to valuable resource designed to help you manage your Brocade Network.