Zoning - node and port WWN's

[Matt_276]

I'm hoping someone can tell me why each device connected to a SAN switch (a MSA1000 2/8 SAN switch in this case) has two WWN's. I believe one is a port and the other a node WWN.

What is the difference? I assume the port WWN comes into use when a device has multiple ports, like in the case of a dual HBA card?

Which WWN should I be using, or is it both, if I'm using WWN-based zoning and why?

Thanks,
Matt

[florence mathon lermusi · Accepted solution]

Hi Matt,

this is it the WWNN is describing the Node itself (for exemple an Array), whereas the WWPN is describing a port of that node (for exemple a controller of that array).

You can use both of them in zoning configuration. The choice is a question of granularity. Using WWPN is the common use in most of the cases...

fl

[Ranjith_5]

Hi Matt,

My understanding HP-UX is not supporting wwn based zoning so we have to use port based zoning. You can find out the port statistics from the switch.

Logon to switch
portshow port_number

#portshow 0
portshow 0
portName:
portHealth: No License

Authentication: None
portDisableReason: None
portCFlags: 0x1
portFlags: 0x24b03 PRESENT ACTIVE F_PORT G_PORT U_PORT LOGICAL_ONLINE LOGIN NOELP LED ACCEPT
portType: 10.0
portState: 1 Online
portPhys: 6 In_Sync
portScn: 32 F_Port
port generation number: 60
portId: 010000
portIfId: 43020001
portWwn: 20:00:00:05:1e:35:3a:8e
portWwn of device(s) connected:
50:00:1f:e1:50:04:59:4c
Distance: normal
portSpeed: N2Gbps

Interrupts: 0 Link_failure: 43 Frjt: 0
Unknown: 0 Loss_of_sync: 8 Fbsy: 0
Lli: 204 Loss_of_sig: 13
Proc_rqrd: 0 Protocol_err: 0
Timed_out: 0 Invalid_word: 0
Rx_flushed: 0 Invalid_crc: 0
Tx_unavail: 0 Delim_err: 0
Free_buffer: 0 Address_err: 0
Overrun: 0 Lr_in: 91
Suspended: 0 Lr_out: 41
Parity_err: 0 Ols_in: 41
2_parity_err: 0 Ols_out: 90
CMI_bus_err: 0



Regards,
Syam

[Uwe Zessin]

Syam,
do you have a pointer to a document which says that HP-UX does not 'support' WWN-based zoning? It should be transparent for the OS as it only defines which ports can talk to each other (in case of hard zoning)/ are reported by the name server (in case of soft zoning).

[Ranjith_5]

Uwe, there is no supporting doc with me...this comment was passwd to me by an engineer from HP during the configuration of our EVA switch. Not full proof. Please correc tme if Im wrong.

Regards,
Syam

[SAKET_5]

Hi Syam,

I will be surprised if any OS level restrictions apply to how fabric zoning is implemented as you describe in the case of HP-UX. May be you mean HP-UX uses PIDs (switch port IDs) to map its OS level devices as opposed to most other popular OSs which do not use the PID??

I concur with Uwe's statements before that fabric zoning is there so that at times of FLOGI & PLOGI when a device logs into the fabric - authorised nodes can talk to other authorised nodes as specified in the fabric name server, alias server, etc.

Please enlighten us if we are missing something.

Just a note on WWNN (World Wide Node Name) Vs. WWPN (World Wide Port Name) zoning - if the node only contains one port - either WWN or WWPN zoning would do - if the node contains multiple ports which connect to different fabrics - you will opt for WWPN zoning. We typically use the WWNN for things like the EVAs in which case an alias like EVA1_ALIAS actually contains four WWPNs (FP1 & FP2 of both HSV110 Controllers or two WWNNs of two HSV110 controller which automatically encompasses two WWPNs each on the controllers)

Regards,
Saket.

[Bill Costigan]

I suspect the reason for recommending port zoning has nothing to do with HPUX (we use WWN zoning with HPUX all the time. Works fine) but how zoning is implimented in the switch.

In the old days, switches did port zoning in hardware and WWN zoning in software. There were some situations where a system could fool the software based zoning and get some information about ports not in its zone. (Not sure of the details)

In the last few years, switches can handle both port and WWN zoning in hardware, and I believe the holes that used to be around in WWN zones have been closed.

I've seen many things accepted as 'Best Practices' long after the reason for them being a best practice no longer exists nor can anyone remember why they were best practices in the first place.