Fibre Channel path problems

[handsy]

I have a new architecture installed and before we go live, I have time to play.

Problem is I'm not getting to play much because the LUNs I'm creating on the NetApp store are not being 'seen' on my ESXi hosts :smileysad:

The ESXi sees the FC adapters fine:

~ # esxcli storage core adapter list

HBA Name  Driver       Link State  UID                                   Description

--------  -----------  ----------  ------------------------------------  -----------------------------------------------------------------------------

vmhba0    hpsa         link-n/a    sas.500143803358d420                  (0:3:0.0) Hewlett-Packard Company Smart Array P440ar

vmhba1    lpfc         link-up     fc.2000ecb1d7ffa274:1000ecb1d7ffa274  (0:5:0.0) Emulex Corporation Emulex LPe16000 16Gb PCIe Fibre Channel Adapter

vmhba2    lpfc         link-up     fc.2000ecb1d7ffa275:1000ecb1d7ffa275  (0:5:0.1) Emulex Corporation Emulex LPe16000 16Gb PCIe Fibre Channel Adapter

vmhba3    lpfc         link-up     fc.2000ecb1d7ff92dc:1000ecb1d7ff92dc  (0:11:0.0) Emulex Corporation Emulex LPe16000 16Gb PCIe Fibre Channel Adapter

vmhba4    lpfc         link-up     fc.2000ecb1d7ff92dd:1000ecb1d7ff92dd  (0:11:0.1) Emulex Corporation Emulex LPe16000 16Gb PCIe Fibre Channel Adapter

vmhba32   usb-storage  link-n/a    usb.vmhba32                           () USB

...but when I look for paths I only get a return on the USB HBA!!!!

~ # esxcli storage core path list

usb.vmhba32-usb.0:0-mpx.vmhba32:C0:T0:L0

   UID: usb.vmhba32-usb.0:0-mpx.vmhba32:C0:T0:L0

   Runtime Name: vmhba32:C0:T0:L0

   Device: mpx.vmhba32:C0:T0:L0

   Device Display Name: Local USB Direct-Access (mpx.vmhba32:C0:T0:L0)

   Adapter: vmhba32

   Channel: 0

   Target: 0

   LUN: 0

   Plugin: NMP

   State: active

   Transport: usb

   Adapter Identifier: usb.vmhba32

   Target Identifier: usb.0:0

   Adapter Transport Details: Unavailable or path is unclaimed

   Target Transport Details: Unavailable or path is unclaimed

   Maximum IO Size: 122880

What on earth is going on? Surely I should see an entry for each FC HBA in that last command even if there were no paths?

The LUN is configured fine, and I've even disabled zoning on my Brocade FC switch pair.

Any help appreciated.

[a_p_]

... and I've even disabled zoning on my Brocade FC switch pair.

Can you please clarify this. Is it a typo and you actually meant "enabled"? In case you disabled zoning on the FC switch (which is not a real good idea), you may need to check the switch defaults, i.e. whether the default-zone-access is set to "allaccess" or "noaccess".

André

[handsy]

As I said in my first post, this is a test network so I have power to do what I want right now.

Zoning is 'Default/All Access'.

[a_p_]

Ok, so let's start from the beginning. Do you see the ESXi host's initiators on the storage systems? Did you present the LUN to the ESXi hosts properly (which Host LUN IDs?).

Maybe not related to the issue, but in any case I'd suggest you create a proper single-initiator zoning rather than to allow all access on the switch.

André

[handsy]

Yes, all 16 host initiators are seen on the NetApp.

I think I may have found a problem......there are no NMP claimrule entries for my FC HBAs:

~ # esxcli storage core claimrule list

Rule Class   Rule  Class    Type       Plugin     Matches

----------  -----  -------  ---------  ---------  ---------------------------------

MP              0  runtime  transport  NMP        transport=usb

MP              1  runtime  transport  NMP        transport=sata

MP              2  runtime  transport  NMP        transport=ide

MP              3  runtime  transport  NMP        transport=block

MP              4  runtime  transport  NMP        transport=unknown

MP            101  runtime  vendor     MASK_PATH  vendor=DELL model=Universal Xport

MP            101  file     vendor     MASK_PATH  vendor=DELL model=Universal Xport

MP          65535  runtime  vendor     NMP        vendor=* model=*

This may be because the firmware on my FC HBAs is not on the VMware supported list??

The most recent HP service pack upgraded the firmware on the HBAs, so if I do need to downgrade does anyone know where to grab the last version?

[handsy]

The firmware issue was a red herring :smileysad:

For some reason my ESXi hosts are not seeing the paths to the FC storage volumes. The NetApp sees every WWPN initiator off the ESXi host servers, but ESXi itself sees nothing. This brings me right back to suspecting I need to add a NMP claimrule?? This seems strange though as VMware documents all say I shouldn't need to change anything in there.

Some help would be appreciated from the community on this please.

Thanks