January 6th., 2006: New Year, same problems. Eventually, HP sent me a pdf relating to the storage tower, a Compaq ProLiant 3092, that I picked up for a song from Retro Computers. It is a hot plug SCSI based piece dating from 1998. Neat if I can find drives with trays, and then have it work as my main storage repository. It took a while for HP to find the file, so, should you need it, here it is. The 3092 should easily connect to one of my two Compaq ProLiant 800 machines. It will take most drives, including Fast-Wide SCSI-2, even high speed later types, assuming the drive cage is compatible. I have been looking for 18.2 and 36.4GB refurbished items. Not that I can afford them yet; there exists an Excel file detailing all of the updates, extras and such that I need. Especially important is finding a P4, with a suitable video card, that will enable me to use Google Earth, and to add my two 40GB spares to its storage capacity.

I took two Compaq drives, nominal 9.1GB SCSI Ultra3, from the stack being held for my other 800 (I shall hunt for Ultra3 SCSI 36.4GB drives for that machine). These were put on the AHA2940 68-pin internal connector, and now that Ghost has performed its magic, they have replaced the older 50-pin drives that were in the box. One of the older, an 8.4GB SCSI-2 drive, has been moved to my HP PA-RISC pizza box, ready for a Debian installation, now that I have downloaded the network ISO and burnt it to CD. All I need now is an external SCSI CD-ROM, and then we'll see how that will run.

Oxford is now running smoothly with these two Wide Ultra3 SCSI drives sitting on the floor of the case and humming quietly to themselves. My other main machine, the Compaq DeskPro EN6530, has started working properly since I removed the Promise Ultra 100TX2 card. This had been used to load two 40GB drives which are too large for the Compaq BIOS, even with the latest update. Nearly every day I had had to hard reboot the machine since the drives, seemingly in unnatural combination with the Promise card, had in some way locked up the box. Only thing to do was to shut down the computer. Now that I have removed the card, I am still waiting for a lock up, but it so far has confounded my expectations. And , now that I have cleaned up all the unneeded services, that still were operating after I had demoted it from a domain controller, it appears to be running quicker. It does need more RAM.

January 8th., 2006: MyHosting, which, as its name states, hosts sites, hosts mine. All three of them. All three are currently run under MS Windows Server 2003, and the web server used is Microsoft, IIS/6.0. My periodic problems with downloading my sites to my own network resurfaced. I had not downloaded them since the change from NetNation to MyHosting last year: when one is using dial-up it can't be done in a reasonable time.

Now that my internet connection is 3Mb/sec adsl, and I have a great deal to upload and change on my sites, I thought that I should check downloading rates again. I found that they were slow once more, even if, that is, I was comparing by date stamp and not by file changes. I use FrontPage 2003 (Office 11) SP2, and regularly check for updates and apply them. Everything should work properly, hein?

Emailing MyHosting via my management page at their site, I discovered that Microsoft put out a fix more than a month ago. The techie couldn't find the link, and I have made three requests for it. Here is a Windows Server page at MS about IIS 6.0 and so forth: complex indeed.

Now then, the rates were improved on each of my sites, and that's fine, but what is the fix, and why won't they tell me, and why can't I find it on a Support search on the MS site? Does it change security levels? It is a registry fix, I do know that. It is, therefore, a real bug, and although the techie said it was a recognised intermittent problem, well, it occurred on all three of my domains, so it can't be that rare. Googling shows that a lot of people have had the same problem.

January 9th., 2006: Well, it would seem I was correct about the Promise Ultra 100tx2 being the source of Rutland seizing frequently. I inserted an Adaptec AHA2940AU in its place: added a small SCSI-2 hard drive and connected the ScanJet IIc to it. Scanner works fine, and the hard drive runs smoothly. And, it hasn't seized. So, don't use a Promise card if you can avoid it. Especially since it uses an EZ-drive, whatever that might be. At the very least it prevents Partition Magic from working.

I was on the MyHosting chat with us over the internet support line: type your question into a box, etc. It appears that there is a bug in FrontPage on the server side, on the hosting server. Something to do with server extensions, IIS that is, and Windows Server 2003. The problem is accepted by Microsoft, and there is a hot fix, but it has not been disseminated. There is a line in the registry that fixes the slow download problem, but what the bug actually is hasn't been properly explained to me. Searching on the Microsoft site proves fruitless. Too much time wasted, because one doesn't know which parameters to use.

January 14th., 2006: At last, a reasonable reason for the download of sites to computers on my network. MyHosting technical support finally sent me a link to this MS page: Event ID 1000 warning messages are logged on the server each time that you perform an authoring action on a Web site. It is somewhat laughable, however, that the associated link on the page one reaches about FrontPage Server Extensions does not exist! Oh, well, at least an answer, for which I am grateful. I have posted this information on a couple of sites, just for information purposes. People should be made aware that their problems may have a real cause, hein?

And, one should recognize that this fix is simply not permanent. The download problem has not disappeared, and it seems ridiculous to inform technical support every time that it occurs. The rate has slowed when I simply added today's data, from a recent 18-20KBps to 6-8KBps today. Rats!

January 16th., 2006: Strange that even though I changed the video card to a newer ATi Rage XL, and downloaded and installed new drivers, the ProLiant 800 still freezes periodically. Event viewer showed that there was a problem with the older ATi card and I hoped this change would stop that. To no avail, and hard reboot causes yet another rebuild of the RAID-5 array. It appears to happen under FireFox, and not when using IE. This is a dual-processor machine, and maybe that has something to do with it, too. Who knows?

January 24th., 2006: Not all NICs work faultlessly. I was wondering why this machine frequently lost contact with the Web. I had to constantly check the diagnostics of the switch, thinking that it was dropping the connection. When I checked the other machines connected to the Web, it was revealed that there was something else wrong. They were always connected. So, I swapped the NICs in this machine, Oxford. The 3Com was made the local network NIC, and the NetGear became the DHCP NIC, with nothing but TCP/IP enabled. Voila, everything works as advertised.

At least, now that we have a Thatcher clone elected in Canada, I can be glad that my internet connection is cured and I can pursue the idiocies of this new Prime Minister. His underwear is fundamentalist, and his outerwear and glowering mien show the pressure this brings.

January 26th., 2006: Is it the dual processor, or is it something else? If I use FireFox 1.5 or IE 6+ then after a while I close the programmes and that does nothing. But if I reopen them then the computer, the ProLiant 800, seizes solid, permanently. Then it is necessary to press the power switch, and after reboot to wait whilst the RAID-5 regenerates. Every bloody time. The FireFox browser expects to send error reports, but it can't if it can't remember what happened. I wonder if it is related to Java applets or Macromedia stuff on certain sites. Probably the latter. Whatever it is I can't determine the cause because there is no way to gracefully restart the machine.

January 29th., 2006: Of course, it was very silly to try and send an error report to Adobe about the ProLiant seizing up by using the ProLiant. The page about Macromedia Flash Player was, itself, running flash, so there was a blue screen of death as I tried to complete the entries. I went to another machine to input a report as I waited for the ProLiant to reboot and regenerate. There were errors relating to event viewer messages that arose because of this, for which see the images above. Not particularly useful, at least for me, since they are somewhat cryptic.

Consequently, when things had quietened down, the machine was gracefully restarted. I have heard that an acquaintance's Intergraph box is constantly crashing. It too is a dual processor, PII, box. Perhaps for the same reasons? But he's the fool who persists in using internet messaging, so who knows?

January 30th., 2006: Thinking about this constant problem with my oldish, but still eminently usable, ProLiant, I wonder if I shall hear about more and more disasters with the increasingly more numerous sales of dual processor and dual-core processor based computers. Given, for one example, that HP sells a dual-core AMD based media PC, then anyone buying one and finding Macromedia Flash screws up their surfing will be irate. I do hope that I hear from Adobe soon, with a cure.

February 3rd., 2006: It may be that my problem with seizing/blue screen on Gloucester has been cured. Macromedia/Adobe sent me an email (I have corrected the grammatical error, slight as it may have been):

Should problems persist, try the following steps:
1. Delete the Flash folder located in C:\Windows\System32\Macromed\ (For Windows 2000 I believe the directory should be C:\Winnt\system32\Macromed)
2. Run the Flash MSI installer via this link http://www.macromedia.com/go/full_flashplayer_win_msi

Test if Flash works, by going to http://www.macromedia.com/shockwave/welcome

Now then, it wasn't quite as simple as that, given where the root directory on Gloucester actually lies. But, it was simple enough to find and to delete the folder. Then, when I clicked the link in the email, it took me to the file to download and allowed me to install the saved Active X executable.

When I had finished installation, I went to the Flash test page, whereupon I was asked if I wanted to download the latest version, which is 10.1 or similar. I had wondered why the test graphics hadn't appeared before this offering appeared. So, having agreed, it started to install, and also asked me to add Yahoo's tool bar, which I declined.

Finally, having finished the installation, I went to a couple of sites, including Adobe and Macromedia, both of which use Flash, and, so far, nothing untoward has happened. Time will tell.

March 18th., 2006: Nothing helps if one fails to read the instructions. Well, in this case, it doesn't work because it's not an Ultra2/3 SCSI drive, but an Ultra160. The machine I was upgrading kept telling me that the disk had failed, and no wonder. Since it was the second ProLiant 800 that I have, and I was bringing it up to workable condition, I wanted an 18GB boot drive. The used drive I bought was, I thought, suitable. Not so, because I never read the label underneath it.

The computer is, however, working now, with somewhat less RAM than I expected. I had thought I had obtained another box, for cannibalistic purposes, that had all of the necessary parts. It had one 256MB and one 64MB, not the two 256MB DIMMs I had been told it contained. Also, with the 18GB drive proving unworkable (the Ultra160), there were only four 9.1GB Ultra2 10k spinners available for use. Therefore, unlike the other ProLiant 800, I used JBOD, not RAID-5. But, it works well enough and will become another AD box when I feel up to it. RAM and another drive will be added when I can find the money and the pieces.

March 25th., 2006: Hunting for another monitor, given that one of my main 21"beasties is starting to give horizontal "tearing" which is indicative of capacitors wearing out in its power supply. I think I'll go down to a 17", because if it takes 1600x1200 that's all I need. I have that setting on a 17" attached to one of my ProLiant 800s, and that works fine. Only $25 second-hand at Retro. And I found another Ultra160 drive. I will have to find a proper adapter now.

April 7th., 2006: Today, in London, or wherever she may be, my elder daughter reaches 35. There was a comedy skit about that by someone, quite a long time ago, and I've been trying to remember the source. Something like, "thirty-five years" in a long, low growl. No idea how Sarah feels at this stage of her life. I do know that she was in India recently, and had a revealing time there.

Now, back to work. my sole Compaq DeskPro EN has been problematic. Signs of the monitor failing and occasional seizing of the system. I had had an Adaptec 2940AU inserted, to cope with my ScanJet IIc and an attendant small SCSI hard drive. The wires for the hard drive had to be routed via a shallow groove on a part of the frame. This meant that when rolling on the side panel, it grabbed the signal/power cable and put pressure on the connectors. Result meant I had to lean the side panel against the computer. Not clever, given what transpired. Yesterday evening, I took an unused Adaptec 2906 in another IBM box, and replaced the SCSI adapter in the EN. I also placed the hard drive in an external box. Sorting out the cables and ensuring the terminator was on the correct connector has enabled me to screw on the side panel. Everything is now running quite well. The heating is normal, and the monitor seems, at this stage, to have stopped producing horizontal streaming of text and such.

May 10th., 2006: Watched the UEFA Cup game. Sevilla winners by 4-0 over Middlesbrough. Thrashed unmercifully, even if the referee should have given the losers a clear penalty when it could have mattered.

Now back to the EN mentioned above. A friend gave me an external SCSI hard drive box. This was used to remove the hard drive from within the EN. I had enough spare cables to connect it to the ScanJet and have them both work. Now that that was done, it was possible to properly close the EN case. This promptly cured the monitor problem and the machine has not seized once.

I eventually removed the smaller of the installed hard drives and replaced them with 20GB and 30GB AT examples from Cornwall. Having copied over the files and renamed the drives before installing them, when the computer was restarted it worked properly immediately. That was a real shock, having had problems with similar operations previously. Now it has three larger drives and works better, now there is no swap drive snafu.

Cornwall has been demoted, since it is a desktop IBM PII 400MHz and one cannot add enough peripherals or drives. I will probably install Debian for X86 on it, using the two 9.54GB AT drives I took from the EN box, Rutland. It will be renamed, and then connected to the network. Having installed MS SFU (services for UNIX), on my other Windows 2k servers, it should be possible to fully implement the uses I plan for it

Not too long ago, I was moving some stuff around and rehurt my hip. So, I was not cured as I had thought I was. Sod it! However, having bought another ProLiant 800, I have built it up as a replacement for Cornwall, and it is now running well with a JBOD formation. Five drives in total, and one, 18.2GB, has been used for the OS, Windows 2k SP4. All three of my ProLiant 800s have the maximum RAM, four sticks of 256MB ECC in each. One needs a monitor, and is running headless at this time just as a unit to copy data to. One thing of interest to me is that of the three machines one has one drive that equates to the volume of all four of the drives in another. Most of the SCSI drives are 9.1GB and they all run at 10k. There are two 18.2GB drives, and then there is one 36.4GB. Can you see how they are related? Multiples of discs seemingly producing their basic capacity. Most are Compaq drives but there are others in the mix, all compatible.

The other thing that quit was my D-Link switch. That had replaced my NetGear switch, that had died last summer. Don't know why, but oddly enough just powering up the old NetGear switch proved it still worked. Why on earth it had seemingly broken down before is not known. So, my network still runs, but I shall have to find another switch to keep as a spare. It was a real horror to think that everything was halted because of the failure of one piece of equipment. Now that the NetGear FS308 has been resurrected I can continue working on PHP and allied self-learning. But, my fingers are permanently crossed.

May 24th., 2006: Lots of problems lately. Mainly, because AD needed cleaning up, two computers were incorrectly denied access to the network, and, therefore, couldn't be logged on. After using several tools: and finding out that several were outdated, I went to Daniel Petri's site again. That had a clear procedure for using ntdsutil to remove some redundant settings. The instructions on Microsoft's site are unclear, as is the help file for the actual utility. Then, having had trouble using adsiedit.msc, I discovered that there are Support Tools available for Win2k SP4. That helped: and that's an understatement. Now, having had to reinstall Win2kAS again on one of my ProLiant 800s, it is now clear that the network works. What a relief.

And, given that a friend has obtained an Adaptec 29160N SCSI card for me from eBay, I shall soon be able to install the two spare Ultra160 drives in hand. The other thing is that I was able to find a second-hand IBM eServer. It's a Series 220 model with a three-port hot swap cage. Now to find used converged IBM drives that will fit. That box has an AT 40GB drive for the OS, allowing me to have the three SCSI drives for storage. The ease of updating the BIOS was also a plus. It had version 1.04, and I downloaded the 1.07 update. The executable allowed me to run it in Windows and reboot to finish. It succeeded, cured a SCSI seeprom error, and now works smoothly. Good stuff: somewhat different from earlier methods of updating any BIOS.

Here is the answer to a clue in a recent Guardian Weekly crossword written by Paul (not me): The Arnolfini Portrait, 1434, Jan van Eyck:



May 30th., 2006: Oddments: at Retro Computers on Bronson today I found a CD external drive, SCSI, for my HP 9000 712/60 PA-RISC machine. Exactly the correct model, plus cables, for a really good price. Simply need to find a power module. But, they are gnats teeth in availability, it seems. And, they are expensive.

I was down there because the 30GB AT drive, that I had removed from the eServer x220 because I had a larger one handy, was in denial. What happened was that it gave me a Disk I/O error: status = 00001001: I tried everything to try to fix this, because although the drive was recognised in the BIOS on two machines, it was invisible to the OS even when using the Recovery Console to try to run fixboot or fixmbr. So, Brian told me that the server came from the RCMP, and that they often installed a programme to screw any attempt to use the drive to install anything: OS installation or NTFS formatting to use as a slave. I expect to hear that Brian has fixed it somehow, or to be given a replacement.

June 5th., 2006: Monday before the Friday that starts the games for WC2006. The Ultra160 drives that I had were placed in a server and now work. One of the drives was unusable, and I was able to exchange it after going through a few machines at Retro. Now, I have two 18.2GB 10K drives: one of them runs the OS: Win2kAS SP4 + Roll Up. Nice and smooth and enables me to work on other machines now that one has six drives of decent size. I am not a gamer, or someone who downloads movies, so I don't need space for those items: merely for web pages and text. And the databases.

June 6th., 2006: Of course, the credulous idiots are claiming that today is the day of Satan: a combination of 6 and 6 and 6. As if the calendar was accurate. How many times has it been changed, and which calendar of which culture is to be recognised? And, when was, in the Christian world, the actual birth? Back to reality: what a difference having the correct Ultra160 cable makes. The SCSI BIOS woke up with 160 instead of 40, which was the megabyte rate prior to installing the twisted cable. Nice to see how things move along now.

July 1st., 2006: England knocked out of the FIFA World Cup, by two horrendous decisions: one by the referee, one by the now unemployable England incompetent manager. Not that the result was exactly unexpected. Also, Brasil was knocked out by France. Huzza!! Zidane and pals were excellent and nice to watch. Unlike the English penalty kick twerps. Only one scored by England was the previously unacknowledged Owen Hargreaves, who had a tremendous game.


Click above to see the photo of my father at almost 87, and my daughter of eleven and some months. The presence of my sister, visiting from Auckland, New Zealand, with her digital camera, has enabled me to have a photo placed here of today's date.

Now then: Retro Computer has come almost good again. Brian, the owner, told me that the hot-swap drives in a Dell PowerEdge 2400 should fit in my IBM x220. Not so, they were out by about a thirtysecondth of an inch. However, after playing with the placing of the drives in the hot swap cage on the 2400, I attached an Ultra160 capable cable to the backplane and to the other two Ultra 160 drives I took from a ProLiant 800. This gave all of the drives in the machine the nominal 160MB/s transfer rate. It really made a difference when using the computer at all times.

The reason I had to rearrange the drives in the cage was the ordering of the SCSI buses on start up. These were altered when I updated the BIOS and the motherboard firmware. The IDs on the hot-swap drive cage are numbered. Thus, ID0 took control for the boot device for the Win2kAS start up. This was not desired since the Adaptec 29160N was attached to the OS drive on ID0, and ID1 was given to the other Ultra 160 drive. So, to have it all work with the current setup, I moved the two hot-swap drives in the bays marked ID0 and ID1 to the empty bays ID4 and ID5. Then, when I booted up the machine, the IDs were properly allocated for all six Ultra 160 drives. Took me a time to sort it out, and I learned something about having multiple SCSI controllers in one machine. This will have to be watched whenever I install a RAID adapter.

As I have mentioned, I'm looking for an Adaptec Ultra 160 RAID card. If this proves possible to obtain cheaply, then I can attach the four Ultra 160 hot-swap drives to it and convert them to RAID 5. This will make them truly hot-swappable, which can't be done with the Adaptec 29160n card that they're attached to at the moment. And, with the RAID card, I can find two more drives to fill the bay and create more storage.

The other thing is that the IBM x220 now has a 133MHz cable connection for the two 40GB AT drives. They now work at the proper connection rate. Took me a while to sort out the positioning of the cables in the machine: snakepit.

July 2nd., 2006: Every time I start up the Dell PowerEdge 2400, but only after I updated the BIOS, etc., (and installed SP4 Rollup and all the other stuff that is required) I receive a new hardware message. This is for the Dell 1x6 backplane on the AIC-7890 onboard SCSI. The hardware is installed and, as the second image shows, is working properly. This is a pain, and searching Dell is fruitless, as is the Windows 2000 support on Microsoft's site.

Above are the two relevant images, click to enlarge. The first shows the start up screen after logging onto the computer. The second shows the properties page for the device. It states it is working properly. Well, it has to be, otherwise the machine wouldn't be working, would it? Daft situation, but I can't cure it so far. Dell has a DPA04.exe producing an .inf file that is supposed to remove the yellow query in hardware. It does not succeed. There are pages in Microsoft support relating to unsigned drivers, but no real help, even using sigverif, in the Resource Kit. Whatever, it is minor, but distracting.

July 3rd., 2006: Apropos the Dell 1x6 backplane: one thing the hardware tells me is that there is an Adaptec AHA-2940U2/U2W in the computer!! Windows 2000 Advanced Server SP4 + Roll Up, plus all the other stuff that updates produce. It should be an Adaptec AIC-7880. This after the updating of the Adaptec drivers, too. No wonder the installation bug arises. Something doesn't know what it should be doing.

But, something has just occurred to me. Given that the backplane is there, that the drives are attached to it, but that the Adaptec 29160N has the cable attachment and not the motherboard, perhaps if I reattached the drives to that the error would disappear. That could be the answer because the BIOS would then be loaded, which it currently is not for the Adaptec AIC-7890 that held the hot-swap drives in the original state. Not likely, however, that I shall bother if I obtain a RAID adapter. I'd rather do that than mess with the on-board RAID, which needs a key that I don't have anyway.

July 24th., 2006: Odd date, just that it means I am 64 and one third years old. So, one wonders, how does one gain employment in these semi-ageist times?

Now then, anything that happens when one watches the Tour de France is astonishing, at times. Floyd Landis won after the most amazing recovery of anyone's memory to regain prominence and the ultimate yellow jersey in Paris last Sunday. Brilliant and nerve-wracking at once.

Computers: the Dell 2400 now has an Adaptec 3200s RAID card. It, too, now that I have updated all of the SMOR, BIOS, NVRAM oddments, shows the 1x6 backplane working, but not installed, even if it is, when Windows starts. Never mind, everything works, and the RAID card has improved the computer. Having turned off the booting option, when I find two more hot-swap drives they can be installed and the RAID-5 setting expanded. Later, later. Compared with another machine, which is using Win2k software for the RAID set up in disk management, not a card, the Adaptec 3200s works instantly. No re-reading of the drives as happens at every start up for the other machine.

Retro Computer should have more SCSI based machines for sale soon. A friend, who has helped me obtain a few goodies from eBay, wants me to reserve him a rack-based computer, should one appear at the store. I might follow his lead, because it would enable me to sort out the space in my cell, and make things more efficient. Money is tight, of course. Additionally, the government bureaucracy is slow in mailing out GST refunds this quarter. After telling me I was late in submitting my return, which was not true, and then stating that they were overworked (!), I eventually received the cheque. I went after cashing it to complete the payments owing for my eBay purchases: Starbucks old chubby ceramic mugs and the Adaptec 3200s.

August 2nd., 2006: Yesterday, after having swapped an inadequate 3Com switch for an external, slow SCSI-2 Yamaha CD writer, I tried installing it into the Dell 2400. The installed SCSI card, an Adaptec 2920, would not show the device, whatever I tried. So, I took the chance and added it to the daisy chain on Oxford, which has an Adaptec 2940UW. This worked straightaway, after ensuring no ID conflict. Odd, what? Will try to move it to the next server I acquire. Thankfully, it runs and it works. Speed is not everything.

August 8th., 2006: It would be better to pay attention to what is written on the labels of my hard drives. What I have done in the past few hours is swap drives in two of my ProLiant 800 boxes. I found, because I was searching the web for information, that some of the drives I had thought were Wide Ultra 2 SCSI, were in fact Ultra160 items. This meant that I firstly rearranged the Dell box and removed an Adaptec 29160n card. I placed all of the drives on the two internal connections on the Adaptec 3200s RAID card. After ticking the boot up allowed configuration box, this allowed the Dell PowerEdge to start up and run as smoothly as ever.

Then, the drives in the ProLiants were arranged: Gloucester now has six Wide Ultra 2 drives, all 9.1GB, four in a RAID 5 disposition, and all running at 10k. Then, I placed all of the Ultra160 drives in Antrim, attached to the Adaptec 29160n card. The Dell's drives run at 160MB/s, as they damn well should, and so should the Antrim drives.

A problem with my Office 11 installation on Gloucester was cured by remembering to alter the drive letter. The drive I took out was set as D: and I had installed its smaller replacement and forgot to check the drive letter, which was C:. Once that was rectified, the normal Tuesday update of security and other matters from MS, was done successfully.

August 10th., 2006: What I discovered with Antrim is that the Compaq BIOS hides the SCSI Ctr-A option screen. That's no good: I therefore trotted down to my storage and recovered four 9.1GB Wide Ultra 2 drives. These I placed in Antrim after removing all four of the Ultra160 drives. Just a few adjustments and most things work as they should. Now I have 5 Ultra160 drives to place in the next server I find. That will be one to replace one of my aging Intel PII 350 boxes. One is the AD box, and is working satisfactorily: but only because the load on it is minor. Few machines for it to bother with currently.

August 30th., 2006: Interesting stuff lately, if you're interested. It started off with an XP Pro upgrade to Oxford, my W2kPro box. It worked in a way, except that I couldn't use Windows Update. I had straightaway activated the machine. I then, because there was nothing in the Networking applet, the icons in the quick start bar failed to appear, and other things on the computer, tried all of the tricks offered by Microsoft to allow me to do the update. Nothing worked concerning update rights. So, I tried re-installing/repairing XP Pro SP2. That for some reason cancelled the activation, which now failed every time I tried it.

Next, I tried the activation tips on Daniel Petri's site, with registry tweaks et al. Nada, nada, nada and then I decided to use the activation screen, phone option. It produced a set of numbers. I phoned the offered number and spoke with a robot which told me I had what appeared to be an improper installation. I then was passed to a guy in India. Eventually, after I had told him about the snafu, he gave me a set of numbers, and voilà, the machine is validated, once again. After a reboot, I still could not run update. So, I go to another page on MS, concerning users. Following the instructions, I check my users status, and, which is what I should have done in the first place, ensured that the administrator was included. All updates to other computers, non-XP, had never produced this situation. My normal user name and status had been downgraded to debugger. Yes, a real Microsoft user name, not what I am, so there!!

Anne-Louis Girodet-Trioson
Portrait du citoyen Belley, ex-représentant des Colonies, 1797,
Huile sur toile - 159 x 111 cm
Versailles, Musée national du château et de Trianon
I have always been interested in colour representation. Having seen that Montréal's Museum of Fine Art has an exhibition, Girodet: Romantic Rebel which had been advertised through an article in the Globe and Mail, November 18th., 2006, where there was a black and white copy of this portrait. I found two versions that look quite different. It's a pity I can't afford to visit the actual exhibition to see what it actually looks like. Given that my younger daughter, Katharine, visited me on the same day I read the article and she's milk chocolate, one subsequently wonders about a lot of things.

 

When I rebooted, it was a different ball game: all of the icons appeared, the networking applet was filled with the proper NICs, and Windows update worked. Did it ever, there were about 60MB of downloads to cope with. Having done that and found that everything sort of worked, I went to use Outlook and found it totally non-responsive, and also giving errors. So did a lot of other programmes.

The cause was simple, and I'll tell you now what it was, in case you have no SCSI interests. The cause of all problems was simply a rabbit-eared connector being a tiny bit loose on one of my external SCSI drives.

Firstly, I put in two more SCSI Ultra 2 Wide drives to try and increase the usable space by spanning them. Errors galore, glacially slow copying. So much so that I removed them and used one as the temp and swap space drive on another server. Next, Outlook still buggy and slow, and every time I had to reboot as things were updated, one of the drives was repeatedly check-disked. I lost some data, and nothing was working as well under XP Pro SP2 as it had under Win2kPro SP4. Recently, I had obtained a cheap external SCSI Yamaha CD-RW. This had stopped working and was removed from the chain because the Adaptec start up would hang. I should have known right then what the error was.

I had found that trying to install the Recovery Console stopped with the statement that the computer held newer files than the XP Pro share folder. This seemed wrong to me, so I downloaded XP SP2 and, after expanding into a folder, slip-streamed the XP Pro share. Then, I ran Recovery Console install and the operation caused an immediate check with MS and the downloading of newer files into the share. This means that the Recovery Console is up to date at this time. Neat, indeed.

Then I found the loose connection, because I wanted to add a CD-RW to another machine and thought I would reuse the Yamaha CD-RW to see if it would work on a different adapter. The cable from the Yamaha to the external hard-drive box was fractionally loose. After a week's work, I should have checked my cables prior to this: it almost certainly happened when I had moved another server which had had the Yamaha drive sitting on its top. I installed the Yamaha in a machine with an Adaptec 2906 card and it worked straight away.

I then found that my main internet machine was working far quicker. Outlook has started to work as fast as it did before, and I have errors reaching sites due to rapidity rather than slowness. The drives all seem healthy, and the racket I was worried about, caused by searching on the SCSI internal drives, is back to its own usual level.

A few of the links I have found useful lately have been Sysinternals, Daniel Petri and a couple of others. These quotes, taken from recent emails I sent to friends, should clarify this: