.\" $NetBSD: ccd.4,v 1.3 1994/11/30 16:22:06 jtc Exp $ .\" .\" Copyright (c) 1994 Jason Downs .\" Copyright (c) 1994 Jason R. Thorpe .\" All rights reserved. .\" .\" Redistribution and use in source and binary forms, with or without .\" modification, are permitted provided that the following conditions .\" are met: .\" 1. Redistributions of source code must retain the above copyright .\" notice, this list of conditions and the following disclaimer. .\" 2. Redistributions in binary form must reproduce the above copyright .\" notice, this list of conditions and the following disclaimer in the .\" documentation and/or other materials provided with the distribution. .\" 3. All advertising materials mentioning features or use of this software .\" must display the following acknowledgement: .\" This product includes software developed by Jason Downs and .\" Jason R. Thorpe. .\" 4. Neither the name of the author nor the names of its contributors .\" may be used to endorse or promote products derived from this software .\" without specific prior written permission. .\" .\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND .\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE .\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR BE LIABLE .\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL .\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS .\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF .\" SUCH DAMAGE. .\" .Dd August 3, 1994 .Dt CCD 4 .Os BSD 4 .Sh NAME .Nm ccd .Nd procedure for configuring concatenated disk devices .Sh DESCRIPTION The .Nm driver provides the capability of `concatenating' two or more disks/partitions into one virtual disk. .Pp This document assumes that you're familiar with how to generate kernels, and how to properly set up a system configuration file. .Pp First, you need to add the disks that you wish to concatenate to your kernel configuration. They should probably have `hard-coded' id numbers as opposed to wildcards, since you don't want the autoconfig sequence to move these disks around. For a hp300, your config might look something like this: .Pp .Bd -unfilled -offset indent master hpib0 at scode7 master hpib1 at scode? master hpib2 at scode? disk rd0 at hpib1 slave 0 disk rd1 at hpib1 slave 1 disk rd2 at hpib1 slave 2 disk rd3 at hpib2 slave 0 disk rd4 at hpib2 slave 1 tape ct0 at hpib0 slave ? .Ed .Pp In this example, the disks rd3 and rd4 will be concatenated. .Pp The `c' partitions of the disks .Pa should not be concatenated. Rather, you should use some other partition, which is defined in the individual labels of each disk, to use the entire disk .Pa except the first cylinder. This is to avoid corrupting the labels of component disks within the concatenated disk. In this example, the `h' partition will be used. .Pp The next piece is the pseudo-device line in your kernel configuration file that actually configures the .Nm device. It should look something like: .Bd -unfilled -offset indent pseudo-device ccd0 on rd3h and rd4h interleave 8192 .Ed .Pp This will configure a virtual disk using rd3h and rd4h, interleaving the access at every 8192 blocks. An interleave is recommended, but not required; if it is not specified, the driver will use each disk in sequence, which isn't as efficient as interleaving the access. .Pp In this example, the disks to be concatenated are HP 7958s, which are approximately 120 megabytes in size. The label on each of the example disks reads as follows: .Bd -unfilled -offset indent # /dev/rrd3c: type: HP-IB disk: label: flags: bytes/sector: 512 sectors/track: 36 tracks/sector: 7 sectors/cylinder: 252 cylinders: 1013 rpm: 3600 interleave: 1 trackskew: 0 cylinderskew: 0 headswitch: 0 # milliseconds track-to-track seek: 0 # milliseconds drivedata: 0 .Pp 2 partitions: # size offset fstype [fsize bsize cpg] c: 255276 0 boot 1024 4096 16 # (Cyl. 0 - 1012) h: 255024 252 4.2BSD 1024 4096 15 # (Cyl. 1 - 1012) .Ed .Pp Of course, this label will not work for any other type of disk, so you will have to create your own. In the simple case, make the `h' (or whatever partition you specified in your ccd pseudo-device specification) the size of partition `c', minus the number of sectors/cylinder, and set the offset to the same number as your sectors/cylinder. This leaves an extra cylinder for the label to reside in. .Pp Once you have given each disk an appropriate label, you should reboot with the newly generated, ccd-capable kernel. .Pp At boot time, shortly after the autoconfig sequence, you should see a message like the following: .Bd -unfilled -offset indent ccd0: 2 components (rd3h, rd4h), 507904 blocks interleaved at 8192 ccd0 configured .Ed .Pp If you do not see this message, then the .Nm device is .Pa NOT configured. .Pp Now it is time to make the device nodes for the concatenated disk, if you have not done so already. You can do this with the .Xr MAKEDEV 8 script, as follows: .Bd -unfilled -offset indent # cd /dev # sh MAKEDEV ccd0 .Ed .Pp Now you need to create .Xr disktab 5 entry for the virtual disk, since it does not support labels within itself. The following is appropriate for this example: .Bd -unfilled -offset indent rd7958-2:\\ :ty=winchester:ns#36:nt#7:nc#2024:\\ :pc#507904:bc#4096:fc#1024 .Ed .Pp All values are copied from the `standard' entry, taking into account any differences from your label(s), except: .Bd -unfilled -offset indent nc#4048 This is the total number of cylinders. This should be the total number of cylinders from all of the partitions you defined your ccd to encompass. .Pp pc#507904 This is the total size, which is not the sum of all of the partitions configured as the ccd. Use the number reported by the autoconfig sequence. .Ed .Pp If your .Xr disktab 5 entry is set up correctly, you can now use .Xr newfs 8 to create a file system on the virtual disk: .Bd -unfilled -offset indent # newfs /dev/rccd0c rd7958-2 .Ed .Pp where `/dev/rccd0c' is the raw device of the .Nm device, and `rd7958-2' is the name of your .Xr disktab 5 entry. .Pp You should now have a working file system on the concatenated disk, which you can .Xr fsck 8 and .Xr mount 8 just like any other, normal, disk. .Sh WARNINGS Presumably, you must use disks of all of the same type. The results of using disks with different geometries would be unpredictable, at best. .Pp If just one (or more) of the disks in a concatenated disk fails, the entire file system will be lost. .Pp There is a kernel-level limit of eight devices per one .Nm device. .Sh BUGS Currently, the .Nm driver works only on the hp300. It should be fairly easy to add support for other architectures using .Xr config 8 , but support for architectures using .Xr config.new 8 is more difficult. .Sh FILES /etc/disktab - disk description file .Pp .Sh SEE ALSO .Xr MAKEDEV 8 , .Xr disktab 5 , .Xr newfs 8 , .Xr fsck 8 , .Xr mount 8 , .Xr config 8 , .Xr config.new 8