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

Re: we need to start setting up for a new computer

On Thu, 4 Dec 1997, Fuzzy wrote:

> On Thu, 4 Dec 1997, Daniel Martin wrote:
> > (about installing the kernel-source package, etc.)
> I think the CDs have kernel 2.0.30? we've heard that 0.29 and 0.32
> were stable but 30 was a problem, but than again, .deb has local
> patches that maybe stablized it? 

The CD uses kernel 2.0.29.  (and installs it) It also has on it packages
that will install the source tree for 2.0.29 or 2.0.30 or a pre-compiled

> is that on the source disk? or do I need to ftp a source tree 
> tar.gz or source package for kernel 2.0.32 (or 33 if thats out by the
> time we do the install)? we currently use 1.2.13 and the setup option
> for kernels have changed in 2.0.x... I'm guessing something like 
> make mrproper
> make config
> make clean
> make dep
> make zImage
The kernel-source2.0.29 and kernel-source2.0.30 packages are on the
"binary" CD - this is actually the proper place to put them, even though
it may be counter-intuitive for some intuitions.  Installing the package
does essentially the same thing as downloading the .tar.gz, unpacking it,
downloading various patches, unpacking them, and configuring the kernel to
the debian defaults.

The kernel-package package enables you to take any kernel source tree
(such as created when you install kernel-source2.x.xx, or created by
downloading the .tar.gz file yourself and unpacking it) into a debian
package of the kernel's binary image.  (i.e. it automates the
configuration, compilation, and installation of a new kernel)  Install it
and read the docs.

But yeah, what you have there will still work. It's what I used before
discovering kernel-package.

> would still work tho.... the 386sx8 is too slow to run X so never
> installed it here. the new system is capable of running it, but it
> will be server and X is a resource hog. of course for the tools and
> when not actually running as a server it would be nice to have
> installed. the 2meg video card and SVGA monitor is supposed to be able
> to do 1024x768x(what ever is the next step after 256 colours).
> hopefully superprobe or whatever will be able to figure it out :). 

Two things: 1) X setup is difficult.
2) It's easier than you think.

Stay away from SuperProbe if you know most of what's in your machine.  Use
> > As for PnP devices - most PnP bioses have places where you can tell them
> > what IRQ's and memory addresses are already used by non-PnP devices so
> > that there won't be a conflict.  After that, you can use (I do) the
> > isapnptools package to set the IRQs, memory addresses, DMAs, etc. on your
> > PnP devices, so that you know what they are when you tell your kernel
> > about them.
> that sounds simple enough, but that means the device drivers have to
> be loaded after isapnptools runs so the drivers use the addresses and
> IRQ as instructed. which means it has to be modular. sighs.
> well maybe modular works better now. when we last tried it 
> in the 0.99.x and 1.0.x kernels it didnt work at all. we still have a
> useless ftape device, (never got it to agree on kernel and module
> versions). 

Well, I've not had any problems, and my kernel is pretty modular.  I seem
to remember that sound did work with my PnP soundcard even when I had
sound compiled in directly, so you may not need to use as many modules as
you think.

> > > we'd prefer not to give the new machine psychiatric difficulties ;). 
> > > thats why we are not going to run slackware on it.
> > > (3)
> > > from what I've seen so far I'm guessing the floppy system
> > > installs a base system, then that is used to dselect the rest
> > > of the packages to make a runing system, is that correct?
> > > 
> > Yep; the base system is actually installed by the set of "base floppies"
> > (about 5 of them), or the base .tgz file on the CD.
> we saw the .tgz file on the CD... would one boot the rescue diskette
> then untar the base .tgz to the mounted target root file system
> to get it built? I guess then setup and run lilo to allow it to be the
> boot device, (tell fdisk to make /dev/hda1 (target /) the boot device
> and have lilo put its boot record in the partition boot record).  

No, no, no.  You're too used to slackware, and having to do everything
manually. The setup program will be able to find the .tgz file on the CD
if you just tell it where the CD drive is, and that will be that.  The
setup program even leads you through partitioning your drive and all that
fun stuff.  If your BIOS can boot from CD (and anything made in the past
year should be able to), just set it to do that, insert the CD and watch
the wonders of setup scripts.  It'll take care of lilo, too.

> hmms... we need to access /home[1-9] on the 386sx system linux
> to copy data to the new system diskdrive, can we use nfs over
> a 10base2 lan with 'local only' ip range for the 2 stations
> assuming that temporarily neither eth0 needs to reach the internet.

Once you get your new debian box set up, look at installing the 'nfs'
package.  As I don't do nfs, I can't help you there.
> that is, a private intranet to device share -- concurent with one host 
> being on the internet with a real IP address, with the other stations 
> on the intranet only seeable by stations on thier segment? do I need
> to tell bind (named) about the intranet or since its only 2 stations,
> can I use /etc/hosts to define them to themselves and each other?
> since the nfs devices will be mounted readonly to the importing system
> locking over nfs isnt a problem. we will be getting a 3bit subnet
> (/29) and so will have 6 usable address that are addressable to the
> world soon. currently there are are only 3 actually in use the
> existing slackware 3.0.0, the new debian 1.3.1 and a WfW (that we are
> unable to convince slackware to see) booting linux on that system work
> fine, so its not a hardware/wire problem its micro$loth problem.
> actually... all that we do on that system is talk to people with ICQ
> mIRC pIRCh and use netscape... I think we saw doc for an x-windows
> based ICQ app and a linux netscape app. is wine able to run either
> pIRCh or mIRC? would WABI? we saw a demo of WABI at the linux expo
> here in RTP north carolina last spring. it had a look of win95 and was
> a win31 simulator. I'm guessing that they used a window manager with a
> look of win95 as the x-windows window manager. if we could get the 
> few 16bit apps that dont have linux versions runing under a simulator
> that would obviate the need for WfW and we could run a real OS there
> too :). even 0.99.x ran this small machine better than Win31 did.
> the current WfW machine is a 486sx25 with 20meg ram and 1.2 eide hd
> svga monitor that runs 1024x768x256 acceptably and 800x600x16m fine.
> it would be an good workstation, has a new sound card. we were looking
> at getting a faster 486dx???? cpu for it. it has enough ram and
> diskspace just a slow (relatively speaking) cpu. the socket has 
> a row of empty pins around the square, we thiunk that means it an
> heandle a newer cpu. its a transition mb 3 vlb slots and 4 ISA/16
> and 1 ISA/8 slot, with 4 (2 banks) 30pin simms and 2 (2 banks)
> 72pins simms. currently its 4-1meg 30pins and 2 8meg 72pins.  

I don't know what ICQ is, but I do run the linux netscape.  There are
debian IRC packages, both text-based and X-windows based. 

WABI is an occasional item of discussion on the list - ask there; I
don't use any windows apps myself.

Why don't you just try the base install (that is, make the boot and root
disks, boot from the disk made with resc1440, and insert the disk made
with root.bin and follow the menus) on the new machine, try to put the nfs
package on and copy what you want, and then see?

TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-user-request@lists.debian.org . 
Trouble?  e-mail to templin@bucknell.edu .

Reply to: