Robert Milkowski wrote:
> Hello James,
>
> Tuesday, February 26, 2008, 11:27:56 PM, you wrote:
>
>
> JCM> But that doesn't help Robert much, or anybody else who wants
> JCM> to be able to get the info without having to learn any mdb-fu.
>
> prtconv/mdb is good enough workaround.
>
>
> JCM> It shouldn't be too hard to hack up a utility to retrieve
> JCM> this info though. (note that I'm not committing to do so!)
>
> Well, maybe I will - I already did some time ago, before fcinfo for
> fc. Maybe I will just add mpt/sas... 
> http://milek.blogspot.com/2005/07/getting-luns-for-mpxio-devices.html
>
>
> However, I believe the proper place for getting lun info with mpxio
> devices should be mpathadm show lu - as someone else suggested, lun
> and target number associated with each path.
>   
As I indicated previously we can have mpathadm display LUN number with 
each path.

I think you are trying to get the t<#> in c<#>t<#>d<#> link for non 
mpxio devices but the target number is not
straightforward to handle because the transport drivers uses different 
schemes to generate it.

The transport drivers export property called 'target' for each pathinfo 
node but they are not always used as the target id.
For FC, the target prop contains AL_PA associated number for private 
loop devices and port id for fabric devices but
it is only used for the target number for private devices. port WWN is 
used for FC fabric devices.
I understand that iSCSI target prop contains the session oid but the 
instance number of the target is used as the target number.
and that the mpt target prop contains some arbitrary(?) number from 
hardware and use it as the target number.
Without the consistent scheme across transport drivers, I don't think 
mpathadm is a proper place to provide such info.

--Hyon

_______________________________________________
storage-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/storage-discuss

Reply via email to