Fibre Channel (SAN)

Reply
Member
Posts: 1
Registered: ‎02-28-2014

Replace core switches and WWPN defined zones

Hello,

 

we have a dual fabric cluster (fabric A and B) with edge-core topology.
The edge switches are bladecenter switches Brocade 5470 and Brocade 6547. The core switches are brocade 300 switches.
The zones are all 1:1 zones WWPN based.
The host OS  are ESX 5.x , Linux (SLES), AIX, Windows w2k8.


Now we want to replace the brocade 300 core switches with brocade 6510 switches.

Our plan is the following:
- ensure same firmware versions at old and new core switches
- configure new core switches (IP, domain ID’s)
- ISL the new core switches to the old one
- move the ISLs to the edge switches and the connections to the storage servers one by one to the new core switches.


Is that a legal migration path if all the zones are WWPN based?

Must there be any precautions taken before moving the cables (e.g. disable port / enable port)?

 

Regards

Carlos

Valued Contributor
Posts: 761
Registered: ‎06-11-2010

Re: Replace core switches and WWPN defined zones

Hi there,

 

It looks good to me. Personally, prior to disconnect a host/ISL I prefer to disable the port.

 

Rgds

Frequent Contributor
Posts: 107
Registered: ‎04-05-2011

Re: Replace core switches and WWPN defined zones

Hello Carlos,

 

It seems to be a good strategy.

 

How many Edge Switches ?

 

just ensure you have multipath installed and all paths LIVE/OPEN/HEALTHY on your hosts to prevent LoA to the LUNs.

 

Chek the architecture of the new core device, you have a great opportunity to populate your new core accordingly to the best practices ;)

 

Run a SanHealth before starting the migration, look for hang mem zones, check if the ISL fan in/out ratios are ok...

 

the hang mems is just to prevent a host is without its target, resulting in a lack of san redundancy.

Join the Community

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

Click to Register
Download FREE NVMe eBook