=======================================================
Linux: ID world wide ID from a lun
=======================================================
:Title:        Linux: ID world wide ID from a lun
:Author:       Douglas O'Leary <dkoleary@olearycomputers.com>
:Description:  How to ID the WWID from a given LUN
:Date created: 05/2013
:Date updated: 06/21/2013
:Disclaimer:   Standard: Use the information that follows at your own risk.  If you screw up a system, don't blame it on me...

This little bit of ugliness came from trying to identify which disks 
are actually the same device in a multipath environment.  This isn't 
the whole solution, but it's a step in the right direction.

First method should be pretty reliable regardless of the environment.  
Basically, you identify the lun for which you want the WWID then use it as
as argument to the scsi_id command.  ``/sbin/scsi_id`` comes with the udev 
package so should be reliably avialable on any redhat related distro.

In my client's environment, we're using 3-par so I was able to use an 
inline script to identify WWIDs as follows:

::

    lsscsi | grep -i '3pardata vv' | awk '{print $NF}' | while read dev
    do
    w=$(scsi_id -g -s /block/${dev##*/})
    printf "%-10s %s\n" ${dev} ${w}
    done
    /dev/sdxd  350002ac0967f087a
    /dev/sdxe  350002ac09680087a
    /dev/sdxf  350002ac09681087a
    /dev/sdxg  350002ac09682087a
    /dev/sdxh  350002ac09683087a
    /dev/sdxi  350002ac09684087a
    /dev/sdxj  350002ac09685087a
    /dev/sdxk  350002ac09686087a

06/21/13:  There seems to be a fairly wide variation in the scsi_id command
arguments between the various versions.  For instance:

* rhel6.4: scsi_id --page=0x83 --whitelisted --device=/dev/${d}
* rhel5.9: scsi_id -p 0x83 -g -s /block/${d}
* rhel???: scsi_id -g -s /block/${d}
* rhel4:   scsi_id -g -s /block/${d} -d /dev/${d}

Nice to keep things consistent and easily scriptable... 

Another option which shows some promise is to use the ``multipath -ll`` 
command.  For instance, the display that follows shows that devs sdik and sdmd 
are pointing to the same gpfs volume with a WWID of 350002ac000200d1e.


::

    # multipath -ll | head
    gpfs167 (350002ac000200d1e) dm-212 3PARdata,VV
    [size=2.0T][features=0][hwhandler=0][rw]
    \_ round-robin 0 [prio=2][active]
    \_ 5:0:1:68  sdik 135:64  [active][ready]
    \_ 6:0:0:68  sdmd 69:336  [active][ready]
    gpfs152 (350002ac000110d1e) dm-194 3PARdata,VV
    [size=2.0T][features=0][hwhandler=0][rw]
    \_ round-robin 0 [prio=2][active]
    \_ 5:0:1:51  sdhp 133:240 [active][ready]
    \_ 6:0:0:51  sdld 67:432  [active][ready]

Both options should be scriptable; but, that's as far as I got before
getting pulled off to different projects.