Unknown Pundit: "The shortest distance between two points is always under construction."

In the Trenches with LAROKE

Konsultant's Log, Cyberdate 06.06.1998 (Old Blue becomes the Old Guard Part I)

SITREP: 12:11 P.M. 5/4/98 "Old Blue" is a shell of his former self at this point. All data has been relocated to the new file server/web server, "Johnny Mnemonic" and only system software and the trial version of the Cheyenne backup software remains. The time has come to start reconfiguring Old Blue for his new role as the company's first communications server.

This is fitting as Old Blue has always been a "pioneer". He was our first IBM AT purchased back in the early eighties, and later he became our first file server, then our first intranet Web server.

The game plan is to remove the Stacker compressed drive volumes from the physical drive D: first. After drive D: is clean (and empty), the installation files for Windows 95 B OSR2 will be copied to it over the network from HAL's CD-ROM drive since Old Blue does not have a CD-ROM drive. Then Old Blues' drive C: will be reformatted with the FAT32 file system prior to reinstalling Windows 95.

TACAMO:

Step 1: Re-configure the hard drives

All files were deleted from the two Stacker Volumes on physical drive D: (re-designated by Stacker as drive F:). Using the "Uncompress drive" command from Stacker's "Tools" Menu, first compressed drive G: (F:\STACVOL.000) was removed, then compressed drive D: (F:\STACVOL.DSK) was removed. The Windows 95 installation files were copied to Old Blue's physical D: drive . . . I think, because the wrong drive light was flashing on the front of Old Blue's case!

3:23 P.M. 5/4/98 As an afterthought, I copied the Invisible LAN installation files to a subdirectory on Old Blue's physical drive D: to make network component installation easier once Windows 95 was reinstalled.

Next, Using the "Uncompress drive" command from Stacker's "Tools" Menu, I tried to remove the last Stacker compressed drive, C: (E:\STACKVOL.DSK) but the program choked with a message that CHKDSK (an ancient DOS command) had to be run. Old Blue was rebooted into MS-DOS mode and SCANDISK was run on all three hard drive volumes instead.

SCANDISK found and corrected problems on two of the drives, then I restarted Windows 95 to try un-compressing the last Stacker drive again . . . RATS! Stacker went into Defrag mode and shortly thereafter produced a "FATAL OUT-OF-MEMORY ERROR 101". I am all too familiar with these shennagigans, and have had a belly full by this point (see In the Trenches Cyberdate 09.13.1997 "A tune-up for Old Blue"). I decided to wipe the drive at this juncture, enuf pussyfootin' around.

A Windows 95B OSR2 system diskette was made for Old Blue (using another PC on the network that had this version of Windows) and the FAT32 versions of FDISK.EXE and FORMAT.COM copied to it. Old Blue was booted from this diskette and Old Blue's drive 1 was partitioned as a single FAT32 volume, then formatted with the "/S" (System) parameter to make it bootable. The drive was labeled "Charlie" at the end of the process (see drive naming conventions).

Step 2: Install Windows 95

Next, after re-booting Old Blue from the newly formatted drive C:, Windows 95 was installed from the installation files copied to drive D: earlier. A custom installation was selected and components to be installed were selected based on their suitability in a server environment. Most communication and utilities components were chosen while "dodads & gimmicks" suited to a user's workstation were not.

I turned on the external HP DataStor SCSI DAT tape drive and TI MicroLaser printer to give the Windows 95 Installer Wizard a better chance of "seeing" them. Old Blue's motherboard is not PnP so it's a toss-up whether they will be recognized or not.

After experiencing the speed of installation of Windows 95 on Johnny Mnemonic (Pentium II 233MHz with 32Mb memory), the installation on Old Blue nearly put me to sleep (486/66MHz with 16Mb memory). Near the end of the process I installed the generic HP LaserJet Series II printer drivers for the TI Microlaser which will emulate a HP printer. The "Generic/Text Only" printer drivers were also installed.

When the setup Wizard was finished, I noticed there was not a "Network Neighborhood" Icon on the desktop. I had instructed the setup Wizard not to install MS Networking components during the setup, only the "NE 2000 Compatible" and "Dial-Up" Adapters.

I was hoping I had not made a mistake here. I did find the "Network" Icon in the Windows 95 "Control Panel", however. I opened it and found the two adapter components there. I checked the properties for the "NE2000 Compatible" NIC and found IRQ3 and I/O address 300h-31F assigned resources. I had forgotten to check these settings before wiping Old Blue's hard drive but I had a copy of Old Blue's previous Windows 95 "System Resources Report" which told me these were the correct settings for the adapter.

The Invisible LAN "Protocol", "Client" and "Server" components were installed (see Step-by-Step "Installing Windows 95 Network Components"). Then I tried to exit the "Network" Dialog to complete the network installation but ran smack-dab into a "catch-22" situation. I was not allowed to leave the Dialog without setting up the computer identification parameters but since I had told the setup Wizard not to install MS Networking components, there was no "Identification" Tab in the Dialog and therefore, I could not enter the parameters! OK, BillG, you win. I installed the "MS Network" Client and "NetBEUI" Protocol components in order to conjure up the Identification Tab. That done and Identification parameters entered, Old Blue was restarted so I could now install the Invisible LAN components again.

The Invisible LAN components were configured (for the second time) using the settings on Johnny as a template and Old Blue was restarted again. The MS Networking components were now removed from the Network Dialog and Old Blue restarted again. OK. That's better. Now, "User Profiles" were implemented from the "Passwords" Dialog (reached from the Control Panel") and again Old Blue had to be restarted.

3:21 P.M. 5/5/98 Now that Windows 95 and the Invisible LAN components were installed, The WIN95 and LAN installation file subdirectories were moved from drive D: to drive C: in preparation for reformatting drive D: as a FAT32 volume. Old Blue was re-booted from the previously prepared A: drive system diskette and the OSR2 versions of FDISK.EXE and FORMAT.COM were used again to convert drive D: from a FAT16 to a FAT32 format. Old Blue was restarted in Windows 95 and Computer Associate's Cheyenne Backup for Windows 95 v2.0 was installed across the network using HAL's CD-ROM drive.

Step 3: Install Modem . . . ~!@#$%^ once again into the breach!

2:56 P.M. 5/6/98 Old Blue was ready for a modem. I had downloaded the Trial version of WinGate v2.1 Internet proxy server / fire wall software last night, and Old Blue needed a modem before it could be installed and evaluated. I decided to give Old Blue the SupraExpress 288i PnP internal modem currently installed in 4-Bits. This is a PnP modem but Old Blue's PnP capabilities are extremely limited.

4-Bits was shut-down and taken off-line for the modem removal. The modem was removed and replaced with an expansion slot cover plate (cover plates should be used when removing adapters so that the proper airflow through the computer case is maintained). 4-bits was restarted to make sure she was OK with this change, then shutdown again.

Taking Old Blue off-line always causes me anxiety. Sometimes he will not start up again . . . It's always a "crap shoot" and this time I would be rolling a run of "snake eyes". Old Blue was now shutdown and taken off-line to install the modem. The modem was installed in the only ISA/VLB slot left.

When it came time to start Old Blue up again, we had problems. At first, he would not start at all until I switched power cables with the DataStor tape drive. Why that should be an issue, I have no idea. Then, after the initial POST tests, he could not find the AdvanSCSI host adapter. All I could get was an "No VL SCSI Host adapter found - BIOS not installed! <press any key>" error message.

I shutdown Old Blue and removed the modem just in case its PnP feature was interfering with the host adapter's resources . . . No joy, the host adapter was still dead! Time to get the manual out. In the AdvanSCSI Silver User's Manual, in the troubleshooting section I found

"AdvanSCSI ROM BIOS does not appear when the system boots: * If the Advan SCSI BIOS does not appear during the boot process, it could be 'hidden' by another adapter . . . or the BIOS may be disabled. * In either case, boot from the AdvanWare Installation Diskette and follow all instructions. By the end of the program, the AdvanSCSI BIOS will be 'relocated' to avoid being hidden by another card (if necessary) and re-enabled (if disabled)."

I found the AdvanWare VL Ver 2.0 utility diskette and placed it in Old Blue's floppy drive. Old Blue was re-booted and after the "No VL SCSI Host adapter found - BIOS not installed! <press any key>" error message, booted off the utility diskette. I followed the instructions exactly including the toggling of a dip switch on the portion of the AdvanSCSI adapter exposed on the back panel of Old Blue.

Then came the "Catch-22": The last instruction in the program was the press the computer's "reset switch" to initiate the changes to the AdvanceSCSI adapter. Old Blue has an original IBM AT case, and they did not have reset switches then. The AdvanceWare utility had disabled the "three-finger-salute" method of warm-booting (CTRL-ALT-DEL), and cold-booting (turning the machine off, then on again) did not work!

Using the PC "HAL", I went to the Advance System Products, Inc. Web site to try to find a way around this dilemma. I did not find much technical info online and the e-mail support form was configured for newer products . . . Catch-22 again. I did make note of the support telephone number, and also managed to find a "utility" download for the AdvanSCSI Silver adapter. After the download I copied the file (an executable) to a system diskette which is what I interpreted the following somewhat cryptic directions to mean:

"These are the available Boot Diagnostic Utilities. To create Boot Diagnostic Disk you will need a blank floppy or the floppy you have created from Specific OS or Multi-OS Drivers"

Next I booted Old Blue with this floppy and tried to run the downloaded executable file. I got back the message that the program must be run with either an A: or B: drive parameter. I re-ran the program with the A: drive parameter and it proceeded to try to make the diskette into a special boot-loader utility diskette but failed when it tried to re-write the boot sector . . . Uhmm. I copied the executable to a new blank formatted DOS disk (but not a bootable system disk) and tried again. This time it worked but the end result was another copy of the AdvanceWare VL Ver 2.0 utility diskette . . . Full circle. Now I had two of them.

It was five in the afternoon by this time in Boca Raton, Florida where I was. It was early afternoon in San Jose, California, where Advance System Products, Inc. is located. I called the support number. After a relatively short time, I was shuttled to a voice-mailbox and asked to leave a message, which I did. I had no idea of when, or if, I would get a response to my query, so I struck out in another direction.

I got out the manual for Old Blue's motherboard, an ASUS VL/ISA-PB486PC model. Even though Old Blue's case did not have a reset switch, his motherboard had a reset switch connector. I figured maybe I could scavenge a reset switch from one of the depleted remains of the already cannibalized PC's "Krash" or "Christine".

I started with Krash. I found that I would have to get the front plastic faceplate off the chassis to get at the reset switch . . . problem was that after a few minutes I realized I could get at all the retaining screws except two, which were blocked by the motherboard. A bad design.

I was pondering this puzzle when a lady named "Kune" called me back from Advance Systems Support. I was pleasantly surprised at the quick response time. The two of us went through the "cold-boot / AdvanceWare Utility / dip switch toggle / cold-boot" process about ten times. I was starting to feel like Pavlov's dog. This process was not as easy as it sounds. Old Blue's location is almost as inaccessible as the computer "P2" in the reception area (see In the Trenches Cyberdate 02.23.1998 "Moving P2 is as much fun as pulling teeth."). Every time I had to toggle the ~!@#$% dip switch, I had to wedge myself into Old Blue's corner of the Krash Lab and lean him forward to get line-of-sight and hand access to the switch. After ten iterations all the fun had gone out of the procedure for me. Finally, Kune determined that I would have to send the adapter to her and they would reconfigure it or provide me with a new one. I thanked her for her help and told her I wanted to try my jerry-rigged reset switch idea first.

Back to Krash. I abandoned this effort after a few more minutes fiddling with the faceplate and breaking a small section of it off accidentally. Time to try getting the reset switch off the Christine machine instead. The reset switch on Christine turned out to be accessible but it was part of an integrated mini circuit board about ¾" high by 2" wide which also had the computer power LED and the hard drive LED incorporated.

I removed the whole assembly and connected the reset switch connector to Old Blue's motherboard. Old blue had been moved out of his corner to a work space next to the computer "Kato". With his case removed, he was connected to Kato's monitor and power cable. Christine's keyboard was dug out of the salvage yard and connected. I decided I wouldn't need a mouse for these operations.

I started Old Blue in this environment and tried the "cold-boot / AdvanceWare Utility / dip switch toggle / reset switch" process . . . The reset switch did not work, so Old Blue was shut down again.

My hunch was that the reset switch assembly needed a power source to work, so Old Blue's computer power LED connector was disconnected from the motherboard and the matching connector from the reset switch assembly substituted. Old Blue was restarted and the process was tried one more time . . . The good news is the reset switch worked this time. The bad news is that it didn't make a difference. The AdvanceSCSI host adapter still would not work. Curses! I'm near the end of my rope here, but not ready to throw in the towel yet.

There are three combination VLB/ISA slots on the Old Blue motherboard and the AdvanSCSI adapter is the only VLB adapter. In addition, the AdvanSCSI adapter needs a "bus-mastering" slot. The ASUS motherboard manual notes that two of the slots are bus-mastering, but neglects to indicate which slots they are. I made a guess when I first setup Old Blue a few years ago and, as far as I can determine, the AdvanceSCSI adapter has been working properly in the slot I picked. The slot I selected is the one farthest away from the power supply connecters and the CPU.

My last card trick here would be to shuffle the deck. The other two VLB/ISA slots were currently occupied by a generic Adaptec SCSI adapter for the HP DataStor DAT tape drive in the slot closest to the power supply and CPU, and the troublesome PnP modem that started this mess in the middle slot.

I removed the Adaptec SCSI card and the modem, leaving only the AdvanSCSI adapter in its original slot, and restarted Old Blue . . . no change occurred. Next, I moved the AdvanSCSI adapter to the middle VLB/ISA slot and tried again . . . Hurrah! The host adapter was found and the SCSI BIOS initialized. Old Blue booted right into windows like he was supposed to!

Old Blue was shut down again and the Adaptec SCSI tape drive card was returned to its original location. Old Blue was tested again, and both devices were found . . . two for two! The final test was critical. The modem was inserted into the slot farthest away from the power supply and CPU, the one vacated by the AdvanSCSI host. I estimated I had a fifty-fifty chance. Either the modem would screw the AdvanSCSI up again or it wouldn't and, if today's luck had changed, all three cards would work.

Old Blue booted into windows again and asked if I wanted to install drivers for the new modem Windows 95 had found on startup. I declined for the time being. It was 7:30 P.M. I decided to quit while I was slightly ahead. Old Blue was shutdown and his cover replaced. I would place him in his customary spot tomorrow.

MISREP: Enuf carnage for one episode. Time to go home and beat the bird . . . OOPs! I mean complain about my day to Wingnut. Next time we'll finish setting up the modem, fix some nasty hardware conflicts, and finetune the new Windows 95 environment prior to Installing the WinGate proxy server / fire wall software on Old Blue.

LAROKE Microcomputer Consultants
155 East Boca Raton Road
Boca Raton, Florida 33432
(561)368-0659 (Tel & Fax)

Issued Saturday June 6, 1998

copyright © 1996-1998 LAROKE Microcomputer Consultants all rights reserved