Re: Accessing SCSI-Devices >2TB

On Wed, Jun 08, 2005 at 01:38:42PM -0400, Paul Mather wrote:
> On Wed, 2005-06-08 at 12:11 -0500, Matthew D. Fuller wrote:
> > On Wed, Jun 08, 2005 at 12:33:06PM -0400 I heard the voice of
> > Paul Mather, and lo! it spake thus:
> > > 
> > > Also, have you considered using geom_stripe for concatenation?  It
> > > postdates ccd.  See the gstripe(8) man page for details.
> > 
> > It probably bears mentioning that ccd will concatenate as well as
> > stripe, which may be far more useful in this case.  And that would
> > lead to gconcat(8) instead of gstripe(8), too.

According ccd(4) it is doing stripes which is a good idea for most
settings.
 
> I never realised until reading your message that there was a gconcat,
> but, sure enough, it's there (and geom_concat). :-)

Concat those two "logical drives" is the only reasonable solution for my
setting. Let me try to explain:

The RAID has 12 SATA-Discs in a 19" unit. Theese 12 discs are
configured RAID5+HotSpare. Its host interface is SCSI320. 
So my host only can see the RAID's logical drives.

The first idea was to have just one large logical drive (LD1) with 12 
physical discs (PD1 - PD12), where P1 is HotSpare. The RAID wants to talk 
a LBA64 dialect of SCSI AFAIK and FreeBSD isn't able to talk this with 
the RAID  --> no /dev/daX!

Second solution was (preconfigured): Having one logical drive (LG1)
partitioned into two chunks 50/50. Both partitions PRT1 and PRT2 are 
upon the logical drive LD1 therefore every access to each partition will
result in physical action of PD2-PD12!!

The partitions PRT1 and PRT2 are mapped to 2 LUNs 0 and 1 for SCSI ID 0.

FreeBSD detects this as

da1 at ahc1 bus 0 target 0 lun 0
da1: <IFT A12U-G2421 342D> Fixed Direct Access SCSI-3 device
da1: 160.000MB/s transfers (80.000MHz, offset 127, 16bit), Tagged Queueing Enabled
da1: 1310177MB (2683242496 512 byte sectors: 255H 63S/T 167024C)
da2 at ahc1 bus 0 target 0 lun 1
da2: <IFT A12U-G2421 342D> Fixed Direct Access SCSI-3 device
da2: 160.000MB/s transfers (80.000MHz, offset 127, 16bit), Tagged Queueing Enabled
da2: 1310177MB (2683242496 512 byte sectors: 255H 63S/T 167024C)

(da0 is another external RAID on another controller channel).

So, just to work around the LBA64 problem, the idea was to recombine
those two "drives" /dev/da1 and /dev/da2 into a RAID (in FreeBSD).

Two ways to do this:
a) striping
  Every access to the logical drive in FreeBSD (/dev/ccd0) will access
  both partitions PRT1 and PRT2. Remember: Every single access to ONE 
  of those two will result in physical disc action for PD2-PD12 and every 
  access to /dev/ccd0 therefore will result in two accesses for each
  physical disc.  --> Bad Idea.

b) concat
  Every access to the logical drive in FreeBSD /dev/concat/data (see
  gconcat(8)) will access /dev/da1 _OR_ /dev/da2 and will cause just ONE 
  physical drive access for each physical disc PD2-PD12.  --> better.

So gconcat will be my way to try it. Thanks for the hint!

So if I will have a ~2.5TB device accessible from FreeBSD I will try to
use the gpt thing upon this.

 
Regards
Raphael Becker