[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: Linux Install on AlphaServer 4X00 5/400 21164 400MHz



> > My guess is that yes, the install kernel never intended to support
> > serial console.  I'd call that an oversight if that's the case, but
> > one that probably wouldn't happen again.  We'll see what happens -
> > this shouldn't take too long...
> 
> Well, that part was correct - it didn't take too long at all.  The above
> code is what's breaking.  The interesting thing is what appears to be
> the cause - the devices used on the RH 6.0 CD.  It _appears_ that to
> save space, /dev/console used the device number pairing (4,0) meant for
> tty0, instead of that which is normally used (5,1).

Um, yeah.  I reported this a while ago and I think it even made it
into the faq.

> Does anyone who understands this code path a bit better than I know why
> I can get through this section of booting if all I do is change
> "/dev/console" to "/dev/ttyS0"?  Yeah, it chokes shortly thereafter, but
> this says something (I think).  Anyway, here's the error that I'm up to
> now (I'm using 2.3.10, as it's the latest stable kernel I have that
> actually builds)...

I've never done the *redhat* install over a serial link, but that's
how you would attempt it. All I've done is used my "magic ramdisk"
system cloner over the serial console, and I did exactly what you say
above:

unpack ramdisk and mount
cd /wherever/it/is
cd dev
rm console
mknod console c 5 1
cd /
unmount and repack, write to floppy etc.

> couldn't open /dev/tty4 for syslog -- still using /tmp/syslog

Hm, this is the redhat program trying to open the 4 virtual consoles
it normally uses...

> running install...
> failed to open /dev/tty1failed.

Hm. Apparently it requires that the 1st virtual console be
openable. So if you also change /dev/tty1 to be /dev/ttyS1, maybe
you'll be on your way?

-- g



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index] []