I'm builing boxes for a friend to help get his company off the ground. He was leaning more towards windows but I convinced him Slack 10 was the way to go however if I don't get these machines up and running I'll have to go crawling back to M$.

They are servers with N440BX boards. I but using the SCSI boot/root disk (always get them confused) and then the install disks. I finish up with the network.dsk to ge the on-board Intel 10/100 pro NIC going (have to specifiy that or the script installs EtherExpress 10/100pro which is just wrong). Assign an IP and try to ping the NFS server.

Currently they're connected through a Cisco switch but that's not really relevant because I've had them directly connected with a X-over and currently Ethereal shows the ARP request from the N440BXs as well as the response leaving the NFS server. And the next request and the next response .... et cetera, ad nauseum.

I can only assume with the limited tools on the install diskettes that the packets are getting dropped on the N440BX side and for what reason or how to fix I don't know. These machines don't have CD-ROM drives or USB ports (a header, but no port) and it would be a humungous pain in the arse to unmount each one from the rack, move the HDs to another machine, build, then move them back (there's 13 of them now, more coming later). Any assistance would be greatly appreciated

Tomaczar; you need all the HOWTO help u can get hold of.1st http://www.ibilio.org/mdw/HOWTO/ contents expand what u need. then try
my favorite: http://tldp.org/HOWTO/HOWTO-INDEX scroll thru contents section,subsection. do not forget the HOWTO/Hardware-HOWTO/index for
compatible/uncompatible devices. scroll scroll scroll u r boat till u find u r dock.( maritime pun) sorry carried away.

Penguin4,

I appreciate your reply, especially so soon after my post, but I'm somewhat confused. Obviously I need help, but what should I go searching through the HOWTO's to learn how to do?? My problem is regarding the specific situation of network troubleshooting in the Slackware bootdisk environment. While I did go to the second link you posted (the first link was dead) I saw no HOWTO on that subject or anything close. I don't have a full set of IP tools to diagnose the issue and so I am hoping that someone familiar with the Slack bootdisk environment might have an idea what's wrong or at least where to start. Thank you again, however, for the response and if I decide to scrub the Slack bootdisk setup and search for alternatives I'm sure your HOWTO links will come in great handy.

tomaczar; that first site if you scroll thru to each section that would pertain to what u r searching for and expand it it will give info needed. that is how my finding that info works for me. just dig in scroll & research.

well maybe i missed something but you didn't actually say WHAT the problem was specifically ( or did you *looks confused )

So the first thing I'll say is, to my knowledge, the limited tools used for the NFS install for slack does not contain anything to look up hosts. So you have to type the IP in, NOT the hostname. Anymore info as per whats happening where and when something is failing during the install?

[edit]
*reads it a fifth time* so your saying that you can't ping the other computer after bootup? Is the other computer a DHCP Server? If not, are you using ifconfig to bring up the ethernet device on the machine you want to do the install with? if so what are you issuing?

It's possible I gave too much extraneous information so here is a more ... svelt description.

1.) Multiple machines needing an OS
2.) 1 NFS Server already configured
3.) Machines booted using Slackware Install Boot Disk
4.) Machines are capable of sending ARP requests to NFS server (as proven by Ethereal) however after ARP replies are sent back, the machines continue to send ARP requests.

Assumptions on my part (just typing that gives me the chills)
1.) The cabling is correct (or ARP wouldn't get through)
2.) The network settings are correct (or ARP wouldn't get through)
3.) The eth0 adapter is UP (or ARP wouldn't get through)
4.) There is no default firewall (that would be insane, not to mention annoying)
5.) All hardware is compatible (the machines came used with RedHat, and they're just Intel 10/100 pro NICs)

I'm running NMAP now to see if they'll respond to any type of probing. Thanx all for the help.

Oh yes, the actual problem is I can't get these computers to talk, specifically the blank machines don't seem to be acknowledging the ARP response from my NFS server thereby completing the handshake and moving forward to the next phase of communication. Also there is NO firewall on the NFS server (iptables CLEAR) . Thanx again.

Let's see. some additional questions:
1. are the ip adresses of the NFS server and the machine to be installed in the same network ?
2. are the NFS versions compatible ?
3. is the exports file set up correctly on the NFS server ?

I didn't do a network install yet, because I will test the new Slackware version a few months on my desktop first before migrating the server too. So I don't know what the behaviour of the Slackware 10 NFS install is, but all former versions I tried it was a piece of cake installing off a Slackware NFS server.

Yes, all the machines are on the same network.
Yes, the NFS versions would be compatible.
Yes, the exports folder is correctly configured.

Important items to remember here, the ARP is unable to resolve the address of the NFS server on the machines. This means that NFS or any other network based ... anything will not work until this happens. Furthermore, packets from the machines are getting to the NFS server so the link is correcty configured.

UPDATE: NMAP scan turned up zilch (surprise surprise). I realize that I forgot to mention that there are two NICs per machine (one onboard, one PCI). I have recently unmounted one of the machines, removed the PCI NIC and am going to try the whole thing from scratch again, hopefully with better luck. Thanx all for the assistance.

WHAT is your NFS server? A slack box, redhat, etc?
The entry in your /etc/exports should look like this

/opt/mirror/slackware10 blankcomputer(ro)

Is the actual problem occuring when you try to premount the nfs directory before going into setup or is the problem happening when you enter the host information when choosing NFS as the Source?

your nmap results will come back positive I'm guessing. Don't assume just b/c theres the ARP flying back and forth that the eth0 was brought up right. How did you configure the eth0?

yeah actually that is a surprise which goes back to what did you do to configure the ethernet on the blankless machine

Gentlefolk,

Many many thanx to all that have taken the time to read/respond to my post. The dualing NICs was the issue, I've done it before but these two NICs were by the same manufacturer and a little too close for the bootdisk environment's comfort. I just hope that after the full install has taken place I'll still be able to use the two NICs (it's integral to the end functionality). Oh well, I'm out of this particular wooded area for right now. Lessons learned:

Take it slow, one NIC at a time.
The e100 is sometimes (always for me) mistaken for the ee100pro by the Slackware network.dsk script
Always read the HOWTOs. Maybe not the answer your looking for, but good information nonetheless

A Merry Slacking to all, and to all a good night.

Just to mention it, in slack 10, I found that only eth0 is configured to get IP by dhcp by default. So that was probably the problem with the two NICs situation.