Trying to do up an old 486.

[geedee]

Right. Have got this old 486 off a mate and am trying to install slackware 8 on it. Am just going to mess around networking my machines together, setting up a gateway etc. It originally had a 66MHz cpu, 16Mb Ram and a 520Mb hard drive. I have now put 64Mb Ram and an 8Gb hard disk in it. I have split the disk into 4, 2Gb partitions in the hope that the bios would recognise it.

Anyway, have got slackware installed on one of the partitions and have tried installing lilo to the MBR and also to a floppy. Whenever I try to boot, it only gets as far as 'LI' in LILO.

I have managed to boot into linux using a boot disk I made earlier using the lowmem.i.

I may reinstall and create a proper bootdisk (like I should've done. I also just did a full install the 1st time as It was getting quite tedious. Only a 2x cdrom:-()

I could just put the old 520Mb hard disk back in and put a minimum install on. Then just use it as a router, but I would like to get X running if possible.

So...
Could the problems I'm having be due to the bios having problems with the larger hard disk? If so, is that unsolvable?

Any thoughts on how I could get lilo working?

Are there any disadvantages to booting from a floppy?

Any thoughts or suggestions would be gratefully received...

cheers,
giles.

[Blackknight]

The only real disadvantage of booting off floppy is it's a lot slower.

As for lilo, try running /sbin/lilo after you get booted up from the floppy.

If everything in /etc/lilo.conf is right, it should work with no problems.

[The Whizzard]

I have now put 64Mb Ram and an 8Gb hard disk in it. I have split the disk into 4, 2Gb partitions in the hope that the bios would recognise it.
Anyway, have got slackware installed on one of the partitions and have tried installing lilo to the MBR and also to a floppy. Whenever I try to boot, it only gets as far as 'LI' in LILO.
The 'LI' when booting, tells me it's the '1024th cylinder' problem. I'm not sure which version of LILO comes with Slack 8 but it should be one that overcomes this problem. You'll also need to have a SWAP partition of 128MB(double of your RAM). Just to make sure, wipe the first partition. Then create a 16MB partition a 128MB SWAP partition and partition the remaining space. The 16MB partition is to be mounted as /boot. Because the /boot partition is WAY below the 1024th cylinder, it should work. If it doesn't, the problem is elsewhere. Possibly the BIOS not recognizing the drive properly.

[Blackknight]

To fix the 1024 cylinder limit, add lba32 to the top of /etc/lilo.conf and run /sbin/lilo. That should fix it.

[geedee]

Thanks. Tried both methods and am still having problems with lilo, have tried installing it to a 16M partition mounted as /boot, adding lba32 to lilo.conf and various other methods. Have recompiled my kernel and have made a boot floppy which works fine. Am not too worried about having to boot from floppy.

Still having a couple of minor problems with X but am putting this down to the ancient graphics card. At least it does what i want it to do.

Am now trying to network my two machines together and am having a couple of problems. They are connected using a crossed patch cable. Have done a netconfig and have given the two machines ip addresses and a netmask 255.255.255.0
Have tried pinging the other machine with no luck. The packet is sent but not returned.

Any suggestions?

[mmccue]

Am now trying to network my two machines together and am having a couple of problems. They are connected using a crossed patch cable. Have done a netconfig and have given the two machines ip addresses and a netmask 255.255.255.0
Have tried pinging the other machine with no luck. The packet is sent but not returned.

Any suggestions?

I tried doing this when I first got my laptop and stores weren't open for me to go buy a hub. I found that if one of the cards is auto-sensing (10/100) and the other is not, that it won't work (the auto-sensing card won't be able to detect the network speed). Got the hub, and viola! Problem solved. Just sounds like you might be hitting the same problem (older machine, prolly a 10m card, newer machine, prolly a 10/100m)

Just my experience,
Mike

[geedee]

I was wondering if the network problems I was having could be related to this:

NE*000 ethercard probe at 0x300: 00 40 33 39 4c 6a
eth0: NE2000 found at 0x300, using IRQ 10.
Hw. address read/write mismap 0

which is part of the output on startup.

Also after I have tried to ping the other machine this is produced by dmesg:

eth0: trigger_send() called with the transmitter busy.
NETDEV WATCHDOG: eth0: transmit timed out
eth0: Tx timed out, excess collisions. TSR=0xf, ISR=0x0, t=115.
Hw. address read/write mismap 0

Have also got another network related problem.When I use netconfig to set up a machine and give it a static ip address, I am unable to connect to my ISP using kppp. The pppd daemon dies unexpectedly. I have to reconfigure the network to loopback. If I am going to set up a machine as a gateway, It is going to need a static ip address and also be assigned a dynamic ip address by the isp. How does this work?