*BSD News Article 81737


Return to BSD News archive

Path: euryale.cc.adfa.oz.au!newshost.carno.net.au!harbinger.cc.monash.edu.au!munnari.OZ.AU!news.Hawaii.Edu!news.uoregon.edu!hammer.uoregon.edu!arclight.uoregon.edu!dispatch.news.demon.net!demon!www.nntp.primenet.com!nntp.primenet.com!howland.erols.net!newsfeed.internetmci.com!news.bctel.net!news.axionet.com!amd40.wecs.org!amd40.wecs.org!duane
From: duane@amd40.wecs.org (Duane Eddingfield)
Newsgroups: comp.unix.bsd.freebsd.misc
Subject: CCD-config and disklabel Quetion
Date: 27 Oct 1996 20:41:35 GMT
Organization: A FreeBSD box.
Lines: 95
Message-ID: <550hdv$9ir@amd40.wecs.org>
NNTP-Posting-Host: mb130.axionet.com
X-Newsreader: TIN [version 1.2 PL2]



   OK thanks to Jordan and the rest of FreeBSD creators, I am just
about to go into production with my FreeBSD based news-server.

   I have 4 Quantum Fireball 3067 MB hardisks organized as follows:

First 1024 cylinders of sd0 has FreeBSD 2.1.5R system  and swap.

# /dev/rsd0s1c:
disk: sd0s1

8 partitions:
#        size   offset    fstype   [fsize bsize bps/cpg]
  a:   131072        0    4.2BSD        0     0     0 	# (Cyl.    0 - 63)
  b:    65536   131072      swap                    	# (Cyl.   64 - 95)
  c:  1079264        0    unused        0     0       	# (Cyl.    0 - 526*)
  e:   882656   196608    4.2BSD        0     0     0 	# (Cyl.   96 - 526*)

Remaining cylinders of sd0 are in a second slice:

# /dev/rsd0s2c:
disk: sd0s2

8 partitions:
#        size   offset    fstype   [fsize bsize bps/cpg]
  c:  5201920        0    unused        0     0       	# (Cyl.    0 - 2539)
  e:  5199872     2048    4.2BSD        0     0     0 	# (Cyl.    1 - 2539)

Other three physical drives are all the same as below:

# /dev/rsd1c:
disk: sd1s1, sd2s1, sd3s1

8 partitions:
#        size   offset    fstype   [fsize bsize bps/cpg]
  c:  6281184        0    unused        0     0       	# (Cyl.    0 - 3066*)
  e:  6279136     2048    4.2BSD        0     0     0 	# (Cyl.    1 - 3066*)

  I did not run newfs on that last three drives but just left them the way
sysinstall set them up, sysinstall was not instructed to newfs them
either.

#
# /etc/ccd.conf
# Configuration file for concatenated disk devices.
#
# ccd	ileave	flags	component devices.
ccd0    32      none    /dev/sd0s2e /dev/sd1s1e /dev/sd2s1e /dev/sd3s1e

 So now I've got:

# /dev/rccd0c:
type: CCD
disk: ccd
label: default label
flags:
bytes/sector: 512
sectors/track: 2048
tracks/cylinder: 1
sectors/cylinder: 2048
cylinders: 11736
sectors/unit: 24037152

3 partitions:
#        size   offset    fstype   [fsize bsize bps/cpg]
  c: 24037152        0    unused        0     0       	# (Cyl.    0 - 11736*)

   I can mount it as ccd0c under /var/news and I have the mongo news spool
that Jordon suggested :-) Seems to work fine and sure is convenient, I
don't have to calculate which groups to put where and the hardware
should be used quite efficiently.

   My problem is that 'disklabel -e -r ccd0' won't let me create the 'e'
partition?? I tried using a few less cylinders and adding the 'a' and 'b'
partitions but the reslult is always the same:

   'line 23, bad partition name', where line 23 is the 'e' partition
   right under the 'c' partition.

   Am I not clobbering the superblocks etc. by mounting and writing
on ccd0c ?? If I were to have a disk-full condition, (almost a sure
thing with a newbie news admin at the helm) would not the ccd0 label
be overwritten too ??

   After hunting around in the mailing-list archive I found the URL
where Satosi says the ccd0 avoids the first 16 blocks of the
compent drives which is somewhat conforting but I am still not
comfortable with this setup. I just need the last tidbit of info
to get this working, or to get my understanding of it right :-)

Thanks for any enlightenment,

Duane Eddingfield. TVINET Internet Services.