Subject: Re: NetBSD, apple fibre-channel card & 2.8TB Xserve-RAID
To: NetBSD Kernel Technical Discussion List <tech-kern@NetBSD.ORG>
From: Greg A. Woods <woods@weird.com>
List: tech-kern
Date: 12/03/2004 17:53:57
[ On Friday, December 3, 2004 at 20:39:57 (+0000), David Laight wrote: ]
> Subject: Re: NetBSD, apple fibre-channel card & 2.8TB Xserve-RAID
>
> The limit for FFSv1 is 2^31 fragments, with large fragments you could
> have > 2^32 sectors.

Duh, of course (and it seems to be just 2^31-1 sectors in the label too,
so large fragments do no good, at least no yet).  Thanks for the correction!

That's really what I meant to say but my fingers over-rode when they
shortened the representation to a "natural" power of 2  :-)

Here's my working label on the other 1.4TB LUN, and the FS within
already using 2048kb frags but still of course just one TB in size:

[console]<@> # disklabel sd5
# /dev/rsd5c:
type: SCSI
disk: Xserve RAID
label: fictitious
flags:
bytes/sector: 512
sectors/track: 128
tracks/cylinder: 128
sectors/cylinder: 16384
cylinders: 179526
total sectors: 2147483647
rpm: 3600
interleave: 1
trackskew: 0
cylinderskew: 0
headswitch: 0           # microseconds
track-to-track seek: 0  # microseconds
drivedata: 0 

3 partitions:
#        size    offset     fstype  [fsize bsize cpg/sgs]
 a: 2147483647         0     4.2BSD   2048 16384    27   # (Cyl.    0 - 131071*)
 c: 2147483647         0     unused      0     0         # (Cyl.    0 - 131071*)
[console]<@> # dumpfs /dev/rsd5a | sed 23q
file system: /dev/rsd5a
endian  little-endian
magic   11954   time    Mon Oct  4 22:57:49 2004
id      [ 0 0 ]
cylgrp  dynamic inodes  4.4BSD  fslevel 3       softdep disabled
nbfree  66865114        ndir    1       nifree  621436  nffree  16
ncg     4855    ncyl    131072  size    536870911       blocks  536754345
bsize   16384   shift   14      mask    0xffffc000
fsize   2048    shift   11      mask    0xfffff800
frag    8       shift   3       fsbtodb 2
cpg     27      bpg     13824   fpg     110592  ipg     128
minfree 5%      optim   time    maxcontig 4     maxbpg  4096
rotdelay 0ms    rps     60
ntrak   128     nsect   128     npsect  128     spc     16384
symlinklen 60   trackskew 0     interleave 1    contigsumsize 4
maxfilesize 0x000400400402ffff
nindir  4096    inopb   128     nspf    4
avgfilesize 16384       avgfpdir 64
sblkno  8       cblkno  16      iblkno  24      dblkno  32
sbsize  2048    cgsize  16384   offset  32      mask    0xffffff80
csaddr  32      cssize  77824   shift   10      mask    0xfffffc00
cgrotor 0       fmod    0       ronly   0       clean   0x01
blocks available in each of 1 rotational positions


> However both FFSv1 and FFSv2 are limited by the label - Current might
> have enough patches to support 2^32-1 sectors.

Not as far as I can tell at the moment, though I've not tried everything
yet either ;-)

I need to rebuild my -current tools and libs first before I can build a
new disklabel binary for the alpha for testing....  Maybe that'll work
flawlessly to completion while I go do the shopping!  ;-)

-- 
						Greg A. Woods

+1 416 218-0098                  VE3TCP            RoboHack <woods@robohack.ca>
Planix, Inc. <woods@planix.com>          Secrets of the Weird <woods@weird.com>