[ale] Request for info on mknod -- or How I ran out of devicenames

[Greg]

It is just a home fileserver - with some mp3s and such on it.  I am still
trying to rebuild it from the RH 7.3 --> Libranet 2.8.1 migration.  It runs
samba and is used mainly as a software RAID 5 backup for all of the home
machines (wife has to have her game configs and Favorites backed up !!)

I guess that after I get it finished I could run some tests ... especially
since I suspect that my harddrives are not taking advantage of the UDMA 133
that they should be on (the dmesg states something like they both are at
UDMA 133 *and* pio) - so that is something to do after I get it finished.

Greg

> -----Original Message-----
> From: ale-bounces at ale.org [mailto:ale-bounces at ale.org]On Behalf Of Bjorn
> Dittmer-Roche
> Sent: Sunday, March 07, 2004 5:43 PM
> To: Atlanta Linux Enthusiasts
> Subject: Re: [ale] Request for info on mknod -- or How I ran out of
> devicenames
>
>
> On Sun, 7 Mar 2004, Danny Cox wrote:
>
> > Greg,
> >
> > On Sat, 2004-03-06 at 18:03, Greg wrote:
> > > I am trying to create 2 device files for my fileserver. It
> has 3 Maxtor IDE
> > > controller cards in it and the dmesg tells me that the 2
> drives that I am
> > > trying to use are called hdo and hdm - however they are not
> in /dev.  I
> > > tried MAKEDEV hdo and MAKEDEV hdm, but it only goes to hdk or
> so. This is
> > > true also on my Debian Woody Sparc64 box.  Someone on the
> Libranet list said
> > > that they can do a MAKEDEV hdo just fine, but I cannot.  I
> have Googled and
> > > know that I need a major and minor number for hard drives -
> is there a way
> > > to find this information on Libranet Debian 2.8.1 or on Linux
> in general ?
> > > I really need to use those 2 hard drives, but it looks like I
> have ran out
> > > of device names.
> > >
> > > Of course, if mknod is the wrong approach I would like to
> know what the
> > > right one is also ...
> >
> > 	That's exactly the problem we ran into at Qunatum building
> a 12 (IDE)
> > drive system (but only using the master side of each chipset).  That hit
> > (hmm, let me see, uh) hdv.  The problem is that due to how they're set
> > up, the kernel just doesn't recognize more than 10 hwifs (hardware
> > interfaces), unless you jigger the defines.  See
> > linux/include/asm-i386/ide.h (MAX_HWIFS) for that, plus you'll have to
> > jigger (perhaps) linux/drivers/ide/ide.c, function init_hwif_data(),
> > plus the array ide_hwif_to_major[] to map more drives.
>
>
> Isn't performance a problem with all those IDE drives?
>
> 	bjorn
> _______________________________________________
> Ale mailing list
> Ale at ale.org
> http://www.ale.org/mailman/listinfo/ale
>