[wplug] install question

[Jonathan S. Billings]

On Friday, March 28, 2003, at 08:39  AM, bgtrio at yahoo.com wrote:
>
> So typically the first hard drive in a system will be /dev/hda, the 
> second
> one /dev/hdb, and so on.  There will often be several partitions on 
> each
> drive, like a boot partition, a root partition, and a swap partition.
> These will be named hda1, hda2, hda3, etc. for the partitions on the 
> first
> drive, hdb1, hdb2, etc. for the second drive, etc.

This isn't exactly clear, but the device naming for IDE disks isn't 
necessarily by the order of the disks, but rather where they are on the 
IDE bus.  So, the primary IDE master is /dev/hda.  The Primary IDE 
slave is /dev/hdb.  The secondary IDE master is /dev/hdc and the 
secondary IDE slave is /dev/hdd.  Other IDE chips in the motherboard 
will be /dev/hde and /dev/hdf, and so on.

So, if you have two disks on the primary IDE channel, they will be a 
master and a slave, /dev/hda and /dev/hdb.  If you only have one disk 
in the primary master, and a cdrom as the secondary master (a common 
configuration), the disk will be /dev/hda, the cdrom will be /dev/hdc.  
In fact, if you only put a disk on the extra IDE channel on the machine 
(maybe it's faster and supports bigger disks) the disk will appear as 
/dev/hde.

The device naming for IDE reflects the location of the device on the 
channel.  This is in marked difference from SCSI, where the first disk 
recognised is /dev/sda, the second /dev/sdb, rather than having the 
names based on the SCSI ID of the device.

Why did they do it one way for SCSI, and another for IDE?  I am not 
certain.  But it does make life easier with IDE because you will know 
what it's device name will be BEFORE you boot up the PC.

Jonathan