cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
1399
Views
0
Helpful
14
Replies

how many wwpn pools

Dragomir
Level 1
Level 1

i created one wwn pool. Is thee 2 wwpn per each wwn? if so do I need 2 wwpn pools?

1 Accepted Solution

Accepted Solutions

No, for the WWNNs create a single pool

20:11:00:25:B5:00:00:00-20:11:00:25:B5:00:00:FF - this is 256 addresses.

For the WWPNs create 2 pools; fabric-a and fabric-b

fabric-a

20:00:00:25:B5:11:AA:00-20:00:00:25:B5:11:AA:FF - 256

fabric-b

20:00:00:25:B5:11:BB:00-20:00:00:25:B5:11:BB:FF - 256

for vHBAs on fabric A select the fabric-a pool. For vHBAs on fabric B select the fabric B pool.

View solution in original post

14 Replies 14

abbharga
Level 4
Level 4

Hi Tony,

You can just configure one pool, both the hba call pull their wwpn from the same pool.

./Abhinav

Jeremy Waldrop
Level 4
Level 4

Another option is to create a WWPN pool per fabric to aid in zoning and management.

For example fabric A vHBAs would have a pool of

20:00:00:25:B5:11:AA:00-20:00:00:25:B5:11:AA:FF

Fabric B vHBAs would have a pool of

20:00:00:25:B5:11:BB:00-20:00:00:25:B5:11:BB:FF

As you can see the next to last octet defines the fabric the vHBA is on.

You can do the same for the MAC pools.

hi

can you explain why it would help with zoning?

I have 2 brocade swiches, so hba0 goes to 1, hba1 goes to 2.

Because you can look at a list of WWPNs on your switch and very easily pick out which ones are for UCS. Also helps to very that Fabric A is connected to FC Fabric A and not crossed into Fabric B.

so is it 2 wwn pools and 2 wwpns pools or 1 wwn pool and 2 wwpn pools?

each wwn has a wwpn value right?

1 WWNN, each server has 1 WWNN

2 WWPN, each server should have 2 WWPNs, one for each vHBA

I typically use this for WWNNs

20:11:00:25:B5:00:00:00-20:11:00:25:B5:00:00:FF

Where the "11" in the second octet is the site/cluster ID. Site 1 UCS cluster 1 would be "11" site 1 UCS cluster 2 would be "12". This is just a logical way to keep IDs unique.

Having the "11" in the second octet of the WWNN makes it easier for zoning as well because you know that you should never zone to any WWN with a "11" in the second octet because that is the WWNN and not the WWPN.

so I guess you are using wwpn for zoning? Or are you using both?

thanks

Zone the WWPN not the WWNN

should the value of the wwn and wwpn be the same?

wwnn: 20:11:00:25:B5:00:00:00-20:11:00:25:B5:00:00:FF

wwpn: 20:11:00:25:B5:00:00:00-20:11:00:25:B5:00:00:FF

so I guess you are saying to create one wwn pool and 2 wwpn pools per chassis?

if I have 2 chassis,

it will be 2 wwn pool and 4 wwpn pools?

No, for the WWNNs create a single pool

20:11:00:25:B5:00:00:00-20:11:00:25:B5:00:00:FF - this is 256 addresses.

For the WWPNs create 2 pools; fabric-a and fabric-b

fabric-a

20:00:00:25:B5:11:AA:00-20:00:00:25:B5:11:AA:FF - 256

fabric-b

20:00:00:25:B5:11:BB:00-20:00:00:25:B5:11:BB:FF - 256

for vHBAs on fabric A select the fabric-a pool. For vHBAs on fabric B select the fabric B pool.

Hi

Thanks a lot. I am asuming on fab-a you mean

20:11:.... and not 20:00...

right?

Also do you have an recommendation on naming mac pools on 2 chassis?

thanks

No, I like to use the second octet of the WWNN as the site/cluster ID. I also use the second octet for the site/cluster ID so that I do not mistakenly zone to the WWNN instead of the WWPN. If there is a 11 in the second octet I know not to zone that.

We do our pools like this.

WWNN

global

20:11:00:25:B5:00:00:00-20:11:00:25:B5:00:00:FF

WWPNs

fabric-a

20:00:00:25:B5:11:AA:00-20:00:00:25:B5:11:AA:FF

fabric-b

20:00:00:25:B5:11:BB:00-20:00:00:25:B5:11:BB:FF

You can use the same logic for the MAC and UUID pools.

do you use the 20:11... for all chassis?

or for the second chassis you use 20:12...

etc...

I only have one site with 2 chassis

thanks.

For all chassis on that cluster. The 20:12 ID would be for a second UCS cluster in the same site. A cluster is made up of a pair of Fabric Interconnects and 1-20 chassis.

With UCS it doesn't matter what chassis a service profile is currently in, it is all one unified system. That is one of the UCS features. Kind of like how it doesn't matter what host a VM is running on in a vSphere cluster.

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community:

Review Cisco Networking products for a $25 gift card