It seemed that it was set to format only certain partitions, even after marking all of the above ones properly. What I did was to go "back" and set the "end" section of the partions again. Then it said ok to format the proper partitions. I dont know what this little quirk is about, but I figured that if I didnt do it now, the install would be hell, because moving /home (under automout supervision) and /var (always a problem) later would have been virtually impossible.

    207.86.37.9        - Static IP 
    255.255.255.0      - Subnet Mask
    207.86.37.2        - Gateway 
    207.86.37.2        - Name Server 
    0.0.0.0            - Backup #1 
    caldera            - Hostname

It then re-lauches the GUI Lizard type of screen, which has now become the default bootup screen.

After all of that, there are some more setup options to deal with. I had my machine boot into run level 3, not 5, because the display was corrupt. I had to setup the graphics adapter settings that didnt work during install, and some other additional settings to set.  Here they are :

Run the "XF86Setup" program to setup the display, if not done already. Save copies of /etc/X86Config to a *.orig type file (it lives in /etc).

I said "no" to reusing the existing settings as a starting point.

I chose "Generic Trident 8900 Series" for the adapter type.

I set the 8 bits per pixel color depth.

I set the resolution to 1024x768 at 60Hz.

Then finally I ran the "init 5" command to take me back where I should have been in the first place - graphical mode.

Ran the "lilo" command to be sure the boot loader program had the freshest copy and was updated.

Reboot using "cd / ; shutdown -g0 -y -i6".

The first graphical login on this machine takes over 6 full minutes to lauch !!

Run "ifconfig -a" to see if the interface is recognized or not, mine was not.

I had looked in some of the system files (/etc/nsswitch.conf, /etc/hwprobe.config, /etc/system.cnf, /var/log/messages, etc) to see if the network was configured and failing, or not configured. I found that it was setup in all the above files, so it appeared to be configured and failing.

I decided to check if the kernel was configured for the networking or not.

I Ran the kernel configuration program.

I added the 3c509 module to the kernel.

I also set it for autoprobing.

Then it seemed to be in the kernel because it showed in the "configured" side.

I then configured the interface as up, with "ifconfig eth0 up".

That brought up the interface, as "ifconfig -a" showed, but the tcp/ip was still not up yet, just the interface itself.

So next I ran the tcp/ip configuration to check on the settings, and just changed one of the fields, then back again to its original value, and saved it.

After a period of about 5 minutes, I again ran the "ifconfig -a" command, and saw that it was now up and running, and had the correct values for tcp/ip configured.

Now that the network was up and running, of course it was time to setup Netscape.

I launched Netscape and configured the proxy settings, under "Edit | Advanced | Proxies | Manual".

Then it was time to setup Samba, because by default it was not setup, but the package (v2.0.5-1) was installed.

Go to the /etc/samba.d directory.

Either copy the smb.conf.sample file to smb.conf, or get a copy from another working machine.

I copied mine over from the Red Hat server I have in house, and modified any specific fields, like the shares and the comments field.

Make sure the file has read permissions on it, if transferred over from another box.

Modify the /etc/sysconfig/daemons/samba file, changing values from "no" to "yes" for boot and configured.

Run the "/etc/rc.d/rc3.d/S91samba start" command to get it going.

Check it via the "smbclient" program and preferably a windows client.

Setup login ability from anywhere for root (generally NOT a good idea, but for my secured site its ok)

Modify /etc/pam.d/login and /etc/pam.d/rlogin to comment out line about securetty stuff.

Now for the standard stuff :

I have to admit that first off, I could not figure out how to add a printer to this system. I need to make the Linux system use my NTserver's shared HP4L printer, so I am used to using Samba for this, but sometimes this presents a challenge. This machine does not have an "lpadmin" command, which I am used to, so I need another solution. After thinking it over for a little while, I decided to try this out :

Install a printer using COAS printer utility, which is under the COAS | Peripherals | Printer.

Add the printer, in my case an HP LaserJet 4L, but without that choice I chose HP 5L.

Use a device name of /dev/null.

Use a printer name of hp4l or HPLaser or similar, I chose "hp4l".

Then let it add the printer for you, and setup the spool directory.

Then after it was added, since I called it "hp4l" I did a search for all files with that information, hoping to find the "print filter".

Basically then, just run the command "find / -name hp4l* -print".

I found what I was looking for in /var/spool/lpd/hp4l, and the file was called "printfilter".

I then copied this file over to "printfilter.orig" to be safe.

You will have to edit the smbprint file to work properly and do the CR to CR/LF conversion.

Now copy this "smbprint" script to the "printfilter" script.

Lastly, you will need to have the configuration file present also. My smbprint script shows my config to belong in "/var/spool/lp/.config".

Please note that I have no /etc/printcap configured to make this work.

Then simply test by catting any file to the printer. For example "cat smbprint | lp" and see if it prints.

ftp over the /usr/local/bb directory from the Red Hat server, since as another Linux distribution, it will be closer than any of the other platforms that I am running.

• Had to edit the bb-hosts file, to include current hosts. Just took it from the comet server, which serves as the display server anyway.
• run the /usr/local/bb/etc/bbchkcfg.sh script to check variable definitions. I had to change a few.
• Check the bbdef.sh for any messages to be "ignored".
• Delete any garbage that may have accumulated, and restart using "/usr/local/bb/runbb.sh restart".
• Get the startup script for the /etc/rc.d/rc3.d directory, so that the BigBrother monitor will launch on startup.
• Lastly, modify the display server's bb-hosts file, so that it will query this new client for all the proper information.

ftp over the Seagate BackupExec agent, so that we can do nightly backups to the "enterprise backup server". If you dont have a copy, like I did, you can get it from Veritas' FTP Site, just look for the be__agnt.tar file.

• Create the /usr/local/Install/seagate/backup_exec/agent directory to store the ftp tar file.
• Create the /usr/local/seagate/backup_exec/agent destination directory, where the program will run from.
• Untar the file, so that it can be installed.
• Run the "INSTALL" script.
• Choose the /usr/local/seagate/backup_exec/agent directory for the destination.
• Setup the shares and backup server options, for the script.
• Run the "INSTALL" script.
• After the installation completes, startup the agent, by running "/etc/rc.d/rc3.d/S96agent.be start".
• Switch over to the backup server, and make sure it sees the new "caldera" unix agent.

And ALWAYS make a backup.

Other info:

I setup a link from /bin/bash to /bin/ksh, so that all the scripts that I have written to be cross platform, will still work, without major modification. Sure its a huge hack, but it will work in the short term.

Web Docs directory for httpd is /home/httpd/html.

Check on umask and ulimit.

I setup dfspace script on caldera, because one does not exist.

I setup a machine type script, because one does not exist.

Modified the root ".profile" script to have the alias for "l" the way I like it, and to include the paths of /usr/local/bin and /var/adm/bin.

I had to (regretfully) setup NFSd on this machine, because a friend of mine, wanting to use Oracle for images, used up all my space on that server. Since this server was built mostly to be a public file storage machine, it made sense. The reason I regret it is that it is sooooo slow, I think it may cause problems. So what I am doing is mounting on the other machine, a portion of this system, so that I can use if for exports/logs.

• First off, the kernel needs to be configured for nfsd, which mine was not.
• So add the "nfsd" module to the kernel.
• Setup /etc/exports to have the proper directory exported.
• Restart the server with "/etc/rc.d/init.d/nfs start"..
• My server didnt have the required files, and the documentation just told me to add the module in the kernel, but that clearly didnt work.
• Since I had another linux machine around, even though it was RedHat, I just copied the /etc/rc.d/init.d/nfs, /usr/sbin/rpc.nfsd, /usr/sbin/rpc.mountd files from RedHat to Caldera, and used same permissions.
• Then I could execute the "/etc/rc.d/init.d/nfs start" and get it working.
• Now at least the other machine is not saying "mount: RPC: Program not registered" anymore, now I just get a permission denied error.
• Well, I must have missed a file, becuase after configuriguring the /etc/hosts.allow file for access, the problem persists.......
• Well, a good nights sleep helps I guess, because now I have it working. Turns out that you have to have a domain name in the /etc/exports file. My line looks like this : "/oracle *.getthru.com(rw)" and that is all it took. I even removed my entries from /etc/hosts.allow that I tried, and it is all fine now.
• I will mention from my testing that when on the client machine, I copied a file as root, on the server machine the permissions were ok, but the user became "nobody". I was in need of using the "oracle" user, so after I got the UID's to match across the boxes, it works ok. I would think that it is a good security for the root ID to be nobody on the server, so apparently it does "root_squash" automatically.

I had to pickup a version of iostat based on the previous nfs issue. I needed to know what was going on, although I dont think it was too hard to figure out.

• First off, ftp to the sunsite.unc.edu site.
• In the /pub/linux/system/status directory is a file.
• Get the sysstat-2.1-1.i386.rpm file.
• Get the libc.so.6 version library or sysstat-2.1-1 wont install. You will get the following :

# rpm -i sysstat-2.1-1.i386.rpm
failed dependencies:
libc.so.6(GLIBC_2.1) is needed by sysstat-2.1-1
libc.so.6(GLIBC_2.0) is needed by sysstat-2.1-1


• You can try to get the libc.so.6, ( I always use FTP Search to find it) but it didnt help me at all.
• Try to install rpm version by typing "rpm -i sysstat-2.1-1.i386.rpm", but it wouldnt work for me, so I had to do the source/complie method.
• Get the sysstat-2.1-1.tar.gz file, then unzip and untar it.
• I had to modify the make file to remove the locales area.
• Compile it by typing "make".
• Install it by typing "make install".
• There were a few problems for me, once again, so I had to do some things manually. I had to place the "sar" and "iostat" executables into /var/adm/bin, and the man pages into /usr/man/man.8.
• You now have "iostat", and "sar".

Sidenote : Not that my system is too slow and old, but "top" reported 80% and above while compiling the sysstat package, all while taking abou 20% itself !!!!.