Subject: Re: Why not change disk naming?
To: John Nemeth <>
From: Lars Heidieker <>
List: current-users
Date: 12/30/2001 16:26:34
At 07:38 AM 12/30/2001 -0800, John Nemeth wrote:
>On May 20, 10:02am, gabriel rosenkoetter wrote:
>} On Fri, Dec 28, 2001 at 08:41:03PM +0100, Julio Merino wrote:
>} > Maybe... but I don't know Solaris disk naming ehehe
>} /dev/dsk/cAtBdCsD, where A is the controller number, B is something
>} I always forget, C is the disk number, and D is the slice number.
>      B is "target" as in SCSI ID number.  I believe C is the SCSI LUN.
>Unfortunately, the only devices to which I have access with multiple
>LUNs (tape changers) are connected to non-Solaris production machines
>and can't be moved, so I can't test this.  I don't have any experience
>with Solaris on machines with IDE drives, so I don't know happens

it uses cAtBdCsD as well and leaves C to 0 allways as there are no luns on ide
B indicates 0 as master and 1 as slave drive on ide bus A

this numbering is really nice but I haven't spent any thoughts about how it 
fits in the partition slice concept.....

>} Note that Solaris also uses a different conception of what
>} "partition" and "slice" mean. (They use the SysV definition. We, for
>} obvious reasons, use the BSD definition.)
>      For most purposes, they are the same meanings (ignoring the
>FreeBSD specific meanings, which don't match other BSD systems).
>}-- End of excerpt from gabriel rosenkoetter