• Printer Support
• X Windows
• Mail/WWW/News Tools
• Dos/Windows connectivity
• File Managers
• Graphics Manipulation
• X Multimedia Support
• Console Multimedia
• Networked Workstation
• NFS Server
• SMB Connectivity
• Anonymous FTP
• WWW Server
• Network Management Workstation
• C Development Package
• C++ Development Package
• Extra Documentation
• --Lastly, select individual packages, to add the following :
• Applications/Communications/EFax
• Applications/Networking/ArpWatch
• Daemons/Sendmail*
• Documentation/how-to-html
• Networking/biff
• Shells/pdksh
• Utilities/Console/Open
• Utilities/Files/Symlinks
• Utilities/System/procinfo
• X11/Applications/PDF
• X11/Utilities/xprocstat, xscreensaver, xsysinfo
• Click OK to install (log of /tmp/install.log)

• Choose Generic 2 button (PS/2).

• Doesnt work properly. Must do it manually
• Set for 4 Mb memory
• No Clock Chip Settings
• Try Probe
• Use default of 8bit color depth at 1024 x 768 resolution
• Take it from me, the Rage Pro Turbo chipset is a pain in the ass to setup. I tried for about 2 hours, using manual and auto settings (probed) to configure this card. Then I decided that it was faster to reload the whole O/S. It takes 30 minutes to load base Linux. That config settings file works ok, but upon any changes, goes out to lunch.

    207.86.37.4        - Static IP 
    255.255.255.0      - Subnet Mask
    207.86.37.2        - Gateway 
    207.86.37.2        - DNS Server 
    getthru.com        - Domain Name 
    linux.getthru.com  - Hostname
                       - Second DNS 
                       - Third DNS 

Setup XF86Setup if not done already. Save copies of /etc/X86Config to a *.orig type file, and editting it manually might be a better idea.

Change the "id" line of the /etc/inittab file from a runlevel of 3 to runlevel of 5. This autostarts the gui like all other modern O/S's.

Reboot.

login as root and check the routes.  They should be all setup, because the network is up now.

Get the hosts file from another machine, like solaris or comet.

Setup proxy in netscape.

Now for the standard stuff :

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

• Using Samba, setup a printer.
• First, bring up the control panel's printer manager
• Make sure to make it an SMB type of printer.
• Use a spool directory of /var/spool/samba.
• Use a printer name of hp4l or HPLaser or similar.
• Use a printer filter of /usr/lib/rhs/rhs-printfilters/smbprint because its gonna anyway !!!
• You will have to edit the smbprint file to work properly and do the CR->CR/LF conversion.
• You will have to edit the config file to work properly.
• And here is a copy of the /etc/printcap file for reference.

• ftp over the /usr/local/bb directory from comet, and setup on gemini.
• 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.

And ALWAYS make a backup.

Other info:

• Docs directory for httpd is /home/httpd/html.
• Check on umask and ulimit.
• I setup dfspace script, because one does not exist.
• I setup a machine type script, because one does not exist.
• Speaking of which, setup a root .profile script.
• No lpstat and lp commands, I aliased them.
• ps command is wierd, must alias it too.
• I unailias all that interactive crap in profile.
• Make sure to set root's login to be a ksh !!

I tried to get the 3c905b (cyclone) adapter to work in 100Mb and Full Duplex mode, but it just keeps defaulting to 10Mb and Half Duplex. I had read somewhere that using the dos diskette and diagnostics/configuration program will fix this but it didnt. I tried the following :

• Shutdown server, and insert a DOS boot floppy
• Boot off of the floppy and insert disk 1 of the 3com driver diskettes.
• Information in help/instruct.txt is where I learned the command line options for the command
• \utils\browse.com \help\instruct.txt --> to view the document
• Run the command line version of the configuration tool as follows :
• 3c90xcfg list --> to find out what cards are in the machine. Mine only has one.
• 3c90xcfg configure /TR:100BASE-TX /OPTIMIZE:NETWORK /FULLDUPLEX:ENABLED --> to set the cards options.
• Exit and reboot.
• At the point of the exiting from DOS, my machine was listed as 100Mb and Full Duplex, but when Linux configures its kernel on the way up, it switches back to 10Mb and Half Duplex

I had to configure the remaining disk space available the other day, so I could finally try out the Oracle for Linux. To do it, i did the following :

• I recommend having a backup before starting this whole process.
• Use the cfdisk command to show existing partitions, and make changes :
• cfdisk
• Go down to your free space line, using the down and up arrow keys
• Use the right and left arrow keys, to select from the options at the bottom of the screen.
• I had to choose new to create a new partition.
• Upon completion, use the write option to write changes and quit. In my case, this was the only drive in the system, so the re-read of the partition table fails. To complete this step, you must reboot the server.
• Upon reboot, begin the filesystem creation. I used the mkfs command to create my filesystems.
• mkfs -V -t ext2 /dev/hda11
• Then after it is created, we need to mount it, and set the options. You can do this thru the linuxconf tool, under filesystems, but I used the command line :
• mkdir /oracle
• mount -t ext2 -o exec,dev,suid,rw /dev/hda11 /oracle
• To be sure everything is ok, I do the following :
• mount --- to see that it is in fact mounted properly
• df or dfspace --- to see if the size is correct
• ls /etc > /oracle/test ---- to make sure the write access is ok
• cat /oracle/test --- to make sure the read access is ok
• At the point, everything is ready. I always reboot the server, to make sure that the startup option is set properly. So just reboot the box again :
• cd /
• shutdown -g0 -y -i6

I must confess that even though I have the older agent (as shipped with version 7.0 of BackupExec) working on my Caldera system (version 1.1 lite) using the iBCS and SCO idea, I could not get it working on my RedHat system (version 5.2). I finally got the Seagate (now Veritas) BackupExec agent to work. I tried to get this to work with all kinds of things, like :

• using iBCS emulation in the kernel
• loading the SCO agent
• recompiling every single agent
• extensively modifying the install script, actually "be_ident"

• The New Veritas BackupExec FTP Site has all kinds of stuff. I use the NT Enterprise solution, so this is the page for that.
• Look for the be__agnt.tar file.
• Simply place this file in the proper area, untar it, and install it. It will completely guide you thru the process.

I had finally decided to begin to do centralized management. I figured that for starters, I will use Outlook Express to get the mail from my root and oracle accounts on all servers, and check them each day. Since I already check log files daily, this should have been no challenge. But what I found was that I couldnt connect to the server, because I kept getting the following error :

Your server has unexpectedly terminated the connection. Possible causes for this include server problems, network problems or a long period of inactivity. Account: linux server: 207.86.37.4, Protocol: POP3, Port: 110, Secure(SSL): No, Error Number: 0x800CCC0F.

• I checked where the log files were on the server, from /etc/syslog.conf
• I checked all the log files on the server, and found nothing
• I checked the /etc/inetd.conf file to see if the pop daemon was being started automatically - it said it was
• I checked /etc/services to see that the service name matches from /etc/services to /etc/inetd.conf which it must, and it did
• I searched for the executable daemon that would be launched from /etc/inetd.conf using : "find / -name *pop* -print" and found nothing

• I checked the Red Hat CD's that were shipped in the box.
• I ran glint from my terminal, which is the graphical installation tool for Linux
• After inserting the first CD in the drive and mounting it, hit the configure button.
• I used a directory of "/mount/cdrom/RedHat/rpms" because I mount the cd under /mnt/cdrom.
• Click the available button, and let it search for all installable packages
• I found a package (rpm actually) on CD 1 and also on 2. (I used disk 1)
• The package I used was called "imap" (version 4.4-2) and was under the "Networking / Daemons" area.
• I queried the package to find out if it had pop as well as imap, and saw the pop daemon that the /etc/inetd.conf is calling for. That seemed to be the right choice.
• I clicked the install button, and it proceeded to install it without a hitch.

• edit the /etc/inetd.conf file, and make a change to the pop line :

#pop-3 stream tcp nowait root /usr/sbin/tcpd ipop3d
pop-3 stream tcp nowait root /usr/sbin/tcpd /usr/lib/linuxconf/lib/vpop3d


• then just send the HUP signal to inet to get the file to be reread ::

ps aux | grep inet
kill -HUP "PID from above command"

My latest problem involved using Oracle on RedHat. It doesnt come from a problem running the products together, it comes from the fact that it works so well, that my space requirements have grown by leaps and bounds. I am well over 600Mb of data now, and that is totally unexpected. So I needed to install another disk. I use CSC (Corporate Systems Center) for most of my orders, because they get great prices on some truly great products. So I purchased an 8.4Gb, EIDE, UDMA2 drive, for all my Oracle data. This is what I had to do to get it up and working :

• shutdown the server as is, and add the physical disk into the machine (in my case, it was the second/slave disk on the primary channel, so the name was /dev/hdb).
• Power up and make sure the BIOS recognizes the disk and is using the preferred access method (in my case the UDMA/2 method)
• Once passing that point, allow the machine to boot Linux.
• After Linux is up and running, you need to format/partition the drive.

fdksik /dev/hdb
d - 1 -> delete any current partition on "slice" 1
n -> create partition
p -> primary
1 -> partition number
1 -> first cylinder
1024 -> last cylinder
p -> print table
w -> write changes


• Then you need to build a filesystem. In my case, I am using the entire disk for a single partition. I know this is a "no-no" for some, but for my specific circumstances, it made more sense. I usually would partition it into smaller "chunks". By the way, I alway use the bad block checking routing, when doing this (-c option).

mkfs -v -t ext2 -c /dev/hdb1


• After that is complete, make sure the entry is shown properly in /etc/fstab.
• Then we must mount the new filesystem.

mount /dev/hdb1


• I always reboot to make sure everything starts up properly, and leave nothing to chance.

Performed first updates to this box in quite a long time today. The goal was to get many of the newer versions of the scanning, security, etc tools installed.

• I began attempt at loading new version of gcc, version 2.95.2, instead of the base 2.7.2.1 that comes with RedHat 5.2, and blammo - the box is dead. I was in the untarred directory, ran a "./configure", and the box died, and would no longer boot. I decided to power off again, and wait another 5 minutes, and now it seems to boot, but of course it had to go thru all of its fsck'ing. On this box, it takes quite a while, due to the 8Gb Oracle database. (I also think I will work from the console this time ....) Man did that little mess screw up my /var partition - now I have bad blocks and all kinds of issues that required manual intervention.

  Finished my "./configure", - unsuccessfully of course, and decided that it wasnt worth all the effort, so I am installing these tools on my Turbo Linux box instead. command.