Partitioning

[jackthebookie]

Hi,

My name is Jack and I'm a Linux virgin.

I have a couple of questions.

First, let me decribe the setup:

1u rack mount

Came with Win Server 2003 Evaluation, which I intend to entirely purge, I do not want a dual boot system here.

Dual 500mhz Pentium III

L440GX+ motherboard

386 RAM

4gb scsi hd - has ntfs

120 gb ide hd - has ntfs

No floppy

Currently a cd rom is attached where the 120gb would be attached, so I have a way to install. I cannot get both to work at the same time. I don't think this should be any big deal, it's a cable issue.

I want to install Fedora Core 1 on the 4gb drive and then use that istallation to partition and utilize the 120gb drive after the install.

I am wondering how I should partition?

I have win98 bootdisk on cd and am able to run fdisk.

Thanks.

[jackthebookie]

I did not mention the fact that I do not intend to have a cd rom attached to the machine after install. The 120gb hd will take it's place.

[bigrigdriver]

Google up a copy of 'Linux Partition HOWTO', sit back, and read. Partitioning is easily accomplished, the the strategy behind the partitioning takes some thought. What the home user needs will be different from what the business user needs. Backup considerations affect the partitioning strategy. Study this HOWTO, then repost if you still have questions.

[mdwatts]

Originally posted by jackthebookie

I want to install Fedora Core 1 on the 4gb drive and then use that istallation to partition and utilize the 120gb drive after the install.

I am wondering how I should partition?


That would depend on what you plan to store on the 120gb drive since the main Linux installation will be on the 4gb.

Are you planning to store mp3's, movies etc.?

[jackthebookie]

Thanks for the replies and the advice on finding that document.

I ended up not being able to use the scsi drive at this time. Apparently there are issues with the aic7xxx drivers that my motherboard uses.

Fedora didn't like the scsi drive at all and the install program crashed on every attempt (hung is a better word than crashed).

I ended up disconnecting the scsi and installing gentoo (stage 3 from basic live cd) on the 120gb drive.

It was a challenge and I'm rather pleased with myself for having gone that route, b/c I think I learned a bit more about how linux works than I might have with a more automated system.

I ended up making 3 primaries and 1 extended partition.

I made boot, swap, and a '/' on the first three (32mb, 512mb, ~50gb).

I made the rest of it extended and then broke that into 5 equal partiions that are currently sitting empty.



In any case, my ultimate goal with this sytem is to use it as a wenserver, I'm sorry I forgot to mention that.

This machine will ultimately (hopefully) be hosting the following;

Apache w/ php support,MySQL, Tomcat, JDK, and Cocoon. As well, I hope have a mailserver. I am asking too much?

Those will in turn be used to host about twenty personal/business websites.

In the meantime I am going to experiment with my current installation and have some fun learning about linux.

What I think I am going to do is finish configuring the server, test it on my lan, then wipe the hard drive and do it all over again, just for practice.

I want to use the scsi and will have to be a bit more knowledgeable to figure out those issues.

I imagine I'll post many more questions.

Thanks

[mdwatts]

Originally posted by jackthebookie

I ended up not being able to use the scsi drive at this time. Apparently there are issues with the aic7xxx drivers that my motherboard uses.

Fedora didn't like the scsi drive at all and the install program crashed on every attempt (hung is a better word than crashed).


I found the following as one of the many hits I got with a G4L search for 'redhat aic7xxx'.


This has been reported to both Mandrake and Redhat:
> >
> > http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=29555
> >
> > I've been trying to find out if there's a fix (if it's aic7xxx 6.1.13
> > that's great!), but Redhat seem to believe it's a 2.4 kernel PCI bug:
>
> Personally I still think its a BIOS bug in those Intel BIOS boards. Hopefully
> some of the VA folks will eventually have time to double check the BIOS
> $PIRQ routing table on these systems and verify if the kernel is parsing it
> wrong or if its wrong in the BIOS ROM

EXTREMELY unlikely. Under a 2.2 no-apic kernel, the aic7xxx card uses IRQ 11
and works. Under a 2.2 ioapic kernel, it uses high interrupts and works.
Under a 2.4 no-apic kernel it *thinks* it is using IRQ 11, but it doesn't work
(however, the card BIOS thinks it is using IRQ 11 as well, and it *does* work
until we boot the linux kernel). Under a 2.4 ioapic kernel it uses high
interrupts and works.

Now, the failure mode of the 2.4 no-apic kernel varies depending upon
situation. Specifically, if the eepro100 or e100 drivers are not loaded, then
IRQ 10 is not enabled by any device driver and loading the aic7xxx driver
simply results in infinite SCSI bus timeouts with nary a single interrupt on
pin 11. However, if the eepro100 or e100 driver is loaded, then IRQ 10 will
have an active interrupt handler on it (which implies the interrupt is enabled
in the PIC) and loading the aic7xxx driver will hard lock the machine (which
implies an unanswered interrupt storm, which is fatal to the machine in
no-apic mode, while an interrupt storm is merely "inconvenient" in ioapic
mode, which makes debugging so much harder in no-apic mode :-( The
combination of this information implies to me that the PCI code in the linux
kernel is *somehow* (and I don't know how because I haven't looked deep enough
to find out) fixing things up in such a way that the interrupt for the aic7xxx
card that used to go to IRQ 11 is now coming in on IRQ 10 and as a result,
things don't work.

Now, I'm all for blaming Intel BIOS bugs if they are the cause of the problem,
but I'm having a hard time seeing that here...


You could have tried adding

noapic

to the initial boot/install prompt. Another is to ensure PnP OS is disabled in the bios.

[jackthebookie]

Thansk for the quick reply md!

I did a great deal of searching over the weekend when I was trying the install and found lots of similar references to the one you posted. Bugs similar to mine have been posted for a wide variety of distros, not just redhat. The Gentoo disto I used had a similar, but not equivalent, problem when I first tried it (before I just went ahead and removed the scsi device).

I am confident that I will be able to the scsi device in the long run, I just figured that for right now I might be better off just learning about how it all fits together, rather than spend hours experimenting with installs. I hope that this will lead to my understanding the problem, as opposed to just overcoming it.

From my reading, and my admitted limited knowledge, I imagine it is kernel related, as it was not an issue I found being addressed by people with kernel numbers lower than 2.4.

Someday I will know! (and understand)

Thanks again,

Jack

[mdwatts]

Did you have a look at

/usr/src/linux/drivers/scsi/README.aic7xxx

as there are plenty of options you can pass to the loading of the module.

[jackthebookie]

MD,

No, where would I find that? On the install disk? Are you talking gentoo or fedora?

[mdwatts]

From the kernel source.

/usr/src/linux
or
/usr/src/linux-#.#.# (your kernel version)

Of course you need to have the kernel source installed. You may find a online copy of README.aic7xxx by searching www.google.com/linux