30th March

This is the sort of thing that my little LCD panel can display - a real-time link into WinAmp and Motherboard Monitor courtesy of Cheeseman at IceHardware, one of the premier UK sites for things LCD. Mine doesn't do anything like this, as yet, but yesterday I finally managed to persuade it to display something useful, and I'm sure that I'll make it sing and dance eventually...

SonicWall have just made their enterprise-level firewall management suite, ViewPoint, freely available to users of the low-end firewall appliances like my little SOHO2. I've been using an evaluation copy of WebTrends Firewall Suite recently, and although it's very nice, it's also extremely expensive by home standards. The SonicWall offering seems to present most of the data that Webtrends does, but also interfaces directly with the firewall to provide real-time graphs of bandwidth and protocol usage via a web browser! It took a little work to install, as it needs a firmware upgrade for the firewall and an immediate patch to the software, but those graphs... <rubs hands together> Ah, now that's more like it... The ViewPoint software can be found on SonicWall's FTP site.

29th March

Today's bonus image - grab your red/blue 3D glasses and marvel at INFINITY's Gold Orb heatsinks in wonders of the patented Epicycle Surround-O-Vision!

Now, this is interesting - apparently Novell offered to help Microsoft in the anti-trust case if MS would collaborate on interoperability with Netware, but were refused. Of course, this was in February 2000, by which time Novell was seriously losing presence in both business and technical spheres, so I can see why Microsoft wouldn't be too impressed with the offer... but the war was over by then (lost by Novell's failures more than won by Microsoft's successes, in many ways) and it wouldn't have harmed MS to improve the lot of the hard-working techy with some better levels of compatibility... Maybe I shall write to Bill again...

This guy is either a balloon-head or way out in front of the pack. He's written a short article on optimal fan placement in a PC case, and it contradicts almost everything I've read over the last few years! According to him, I should ditch all the intake fans on INFINITY and just have a barrel-load of exhaust fans, creating an under-pressure condition in the case and sucking cool air in from outside. However, most pundits seem to favour an even balance between intake and exhaust, on the grounds that the best thing to do is to keep the air moving through the case - and if you push air, you can at least try to control the direction in which it moves. A low-pressure area created by pulling, on the other hand, will be fed from the entire surrounding volume - including air that has already been heated by other components.

I'm not sure, though... INFINITY is definitely far warmer than it should be, considering the hundreds of cubic feet per minute of air that I'm moving, and considering the price I pay in decibels... Maybe I'll suspend disbelief long enough to fiddle with the DigiDoc a little, and see what happens when I turn off all the intakes... Hmmm...

28th March

I think my immediate manager has been out in the sun too long...  He's normally very level-headed and sensible, but today he seems to have decided that he's qualified to specify an enterprise-level fibre-channel RAID array without consulting me... Given that he recently needed his hand held over the phone to insert a new tape into the backup library, I think he's being a little reckless, there... Last time he did this it turned out very badly indeed, and we've lived with that mistake (the current RAID array, as it happens, which can't extend the volume onto additional disks without a complete re-format!) for the last three years... I'm not in a much better position myself, I have to admit, as I've never even seen fibre-channel hardware, so presumably this is going to be one of the "challenges" that my recent appraisal mentioned... Oh, well...

More adventures with Compaq support today, too - today's engineer was rather more competent that last week's, but apparently under orders from his management to only replace the faulty disk controller if threatened with deadly force... I wasn't feeling that contentious, so instead he upgraded the controller's BIOS, which removed the "parity errors" warning and thus the remainder of the evidence... I'm still convinced that it was a hardware glitch, especially given the errors in the NT event log, so I have no particular confidence in the quality of the repair... But my management are in a hurry to get the server back online, and I wasn't even allowed time to do any soak testing before rebuilding the OS and handing it over to the DBA for the Oracle installation. I see nothing good coming of this, unfortunately, but at least I'm on record warning against it... <sigh>

One positive note, today, was finally getting the third of last week's server victims back up and running - I've been running some of my monitoring and management tools on that system, and have felt rather insulated from the network's heartbeat without it.

Ah... the weekend...

27th March

I've finally heard the full story of the Compaq engineer's visit last week, and his death-toll stands at three servers, not just the two I knew about! Apparently he decided that one of the CPUs had failed, and swapped it into our identical DR server in a remote office - and this server is now dead, too... As far as I can see, this is complete bollocks: a dead processor would have been clearly identified in the power-on self-tests, and is very unlikely to cause a blue-screen mid-way through NT's boot process instead! It's most likely that he killed the CPU in what was apparently a frenzied burst of random component swapping, and then tried to cover his tracks... And, of course, his theory doesn't explain the "Disk Controller Failure" entries that I recovered from the system event log today, or the fact that the server diagnostics clearly report parity errors in the controller's cache... Why he didn't run those diagnostics himself, as soon as he arrived, is a complete mystery to me!

However, the second of his victims is almost back up and running again, bar some weirdness with the network adaptor drivers - doubtless nothing that a fresh brain won't fix tomorrow. I'm really glad that it's a short week, though... I've seen enough dead NT servers recently to last me a year... Oh, yes, and Insight Manager is now reporting a SMART disk warning on the current primary of this server team, so that makes all four of our ProLiant 3000s failing in a week! I guess maybe they are getting old...

26th March

Success with one of the dead servers, today - I installed a second copy of NT along-side the first, and after some false starts I managed (with the able assistance of NT-guru Mike, who happened to be on-site today) to get booted and onto the network so that I could siphon the data off onto another server. The hardware will need some careful soak-testing to identify any problems, but the crisis has passed for the moment. The "spare" server, also dead, can be an excellent test of the new Backup Exec system - by chance I'd performed an extremely thorough backup of that server right before my week off, so it shouldn't be too hard to bring it back from the grave.

My only real concern, now, is that I don't know what actually happened to the original server... All the disk problems seemed to go away after booting the second NT and running a CHKDSK, so it now seems to be down to Act Of God or high-energy cosmic rays... I guess they're the same thing...

However, the mutterings of the Compaq engineer about obsolete hardware (they're only three years old, dammit!) may have persuaded my management to upgrade some of our core servers - that will be the usual mix of glee and panic for me, I'm sure, and is yet another thing to "look forward to" this year...

25th March

Definitely a bad server day at work, today... the primary of the pair of external customer-facing servers died with symptoms of a RAID controller failure while I was off sick last week, and the Compaq engineer who attended in my place not only failed to repair the dead server, but managed to break the hot-spare whilst trying. In the end the DBA switched over to the secondary server, more of a last-resort than it should be because the V7 Oracle app that runs the the system apparently can't be configured to replicate in real-time - so there were around three minutes of transactions trapped on the dead primary. NT would load around half-way, and then fail with a blue screen ("Inaccessible Boot Device" - a patently false statement!), but I was fairly sure that the separate RAID array that holds the databases was still intact.

I spent some time on both servers, using the spare to test some ideas, but didn't really achieve anything with the standard repairs - but although I could only lay my fingers on the read-only demo versions of NTFSDOS or ERD Commander (it really is time we bought that - the latest version looks great!), they were enough to confirm that the mirrored boot partition definitely had some data corruption but that the data volume appeared intact.

I shall try again tomorrow with a fresh brain.

23rd March

Finally left the house to go phone shopping, today, and it was so nice to see the sky again... I've settled on a set of three matching DECT handsets based around a central transceiver and answer-phone from BT, the Synergy 2150. Somewhat to my surprise, BT's online shop was around the cheapest, and the hardware should be with us at the start of the week. We're hoping that they don't prove to be too small and fiddly, though - it seems impossible to buy a digital wireless phone significantly less minute than an average cellphone, these days, and I'm yet to be convinced that it's a good thing. In general though, DECT seems to both an extremely mature technology, and also an unusually extensible one - there are all sorts of funny little bits of combined voice/data hardware coming onto the market to fill peculiar little niches, and I suspect that it could win the low-end wireless home networking market away from Bluetooth and 802.11 if anyone was seriously pitching it that way.

I spent a while tonight tidying the Space Models pages, and also adding another image to the Digital Art page... which also required trying to trace an old friend to a new email address to check for his permission, it being a heavily-reworked photo of him from our salad days...

22nd March

Almost back in the land of the living, and fancied doing a little techy-something to get the blood flowing again... When the phone line died last weekend, the BT engineer routed around the problem by connecting a new wire from the street into the wallbox of one of the old data lines, unused since we went to broadband. This minimised the amount of fuss (for both of them and us!) and had us back up and running quickly, but meant that the extensions hung from the bypassed line were also now dead! I'd need to run completely new lines to bedroom and basement, as I have no idea where the current wires are routed and little inclination to start pulling up floorboards, but for the phones in the main room I planned to re-arrange the existing extension cables.

Easier said than done, however, as the phone extensions were probably the first wire laid when we moved into the house, and are now behind, beneath and buried by everything else that the last few years has accumulated - lots of CAT5 network cabling and lots of S-Video between my PC and the AV hardware, as well as the more conventional domestic string. I was trying not to jerk or pull too hard while I was unthreading it, but evidently it took exception to being disturbed and when I hooked it all back up again it didn't work... The connections at the ends look Ok, so it's probably a broken wire - not uncommon in cheap cable that's been laid for a long time.

I could just go out and by more bell wire, but if the funds will stretch (and I think they will) re-working the whole thing as a cordless DECT solution with multiple handsets seems to be infinitely preferable... It's shopping time again!

20th March

<sneeze> <sniffle> <cough>

In my occasional lucid moments I've been fiddling with the rather bewildering variety of drivers and applications for the LCD panel. There is certainly no shortage, but as yet it's proved impossible to find one that really stands out from the pack... So far I've looked at: LCDriver, LCDcenter, LCD Smartie, LCDC, LCDStats and finally, for a welcome change in naming convention, Crystalline. Oops! Another search has just tapped into a whole new list, and one of them seems to echo the standard Win2K PerfMon counters out to the LCD - now, that's more like it! There's more exploring to be done, obviously, when I have a brain again...

19th March

Looks like a bug sneaked through my own anti-virus defences... coughs, sneezes, aches, pains and total lethargy - but at least I'm not sending email to everyone in the address book behind my own back...

17th March

Today's hardware failure - one of the phone lines. I think we must have gremlins breeding in the basement...

The LCD is in place, though, and the basics are working. Machining the hole was a fairly long job, but I did achieve as professional a finish as I was hoping - at least from the outside, as the LCD module is currently attached with neat cubes of Blu-Tack on the inside! I was hoping for a tight push-fit, but had to remove another fraction of a millimetre to keep the top edge straight and lost the snugness - but the Blu-Tack works well, and I'll probably leave it that way.

It looks better in the flesh - but was hard to photograph!

I think the LCD's onboard micro-controller may be crashing intermittently, though - I seem to be able to talk to it quite happily for a while, with various applications, but then it suddenly refuses to display anything until I cycle the module's power. It needs investigating, but Matrix Orbital have a well-attended tech support board and from my previous browsing it's highly likely that someone there will be able to help.

In other news - my favourite tech site, Dan's Data, has a comparative review of three of the market-leading heatsink compounds going head-to-head against toothpaste and Vegemite. The results are surprising, to say the least...



Click here for 1st Mar to 16th Mar



Weblog Archive