Thursday, January 27th 2005

Photos (from left to right): Rootes pump, Carl von Linde’s Air Liquiefaction equipment (taken at the Deutsches Museum, in Munich).

Finally uploaded more photos from last weeks’ trip to Munich. I enjoyed that visit, but I learned that Adidas gloves aren’t necessarily all they’re supposed to be since the ones I bought in Munich have already broken at the seams. Bah.

Today I’m trying to get some stuff finished off at home in preparation for being in Canada from next Tuesday. I’m flying out on a Boeing 767 300ER (the same family I was on in the summer – last time it was plane equipment number 647 if anyone cares) at 16:35 on Tuesday afternoon and shall return on the 21st, arriving in Heathrow at around 11:15. I’m staying in a Victorian guesthouse in Ottawa for most of the first 11 nights, but am still planning a day trip to Toronto which has to fit in somewhere. I’ll be in Kingston for a few days, on the road to Boston, MA (via Maine, New England) and then in Boston during the show itself. I’ve arranged to meet up with Jeremy Allison (I think I can now mention he’s doing a column in the magazine) – I’m hoping he can put me in touch with the guys working on various embedded Linux media stuff at HP too – and hopefully one or two friends from other companies over there. I also have recently corresponded with some GNU/Linux companies in Ottawa (anyone else want coffee?) and today spoke with the folks over at Novell about meeting up for a few hours.

After recently moving some servers around, it became necessary to abort an installation of Ubuntu and recover the previous Mandrake installation from a bunch of soft RAID level 5 disks on a particular box. Luckily, only the partition tables had been nuked at the point that the install was cancelled. This meant that a few hours with xxd, less, fdisk, dd, disktype, and friends – coupled with the fact that RAID superblocks live not at the end of a blockdev, but at the last 64k aligned block (very important to realise, not well documented however) – allowed me to manually locate and read off the superblock contents, performing necessary endian conversions, and reconstruct the appropriate partition table entries which had existed previously. Unfortunately however, the geometry of the disks had been forced to be different previously (4866 cylinders rather than the real 4865, for various random reasons) which meant that I also had to figure this bit out. Anyway, at this point it’s almost ready to force mdadm to use it – anyone who likes non-Free Software: how would you do this with your favourite OS?

I added a bunch more entries to recently, am doing some reading, and keeping up the daily exercise within certain limits – I’ve also upped the number of situps to 40 each morning and evening whenever possible, and 20 pressups each time too. I have not yet gotten around to that serious python study I was going to do – but I hope to spend some time in the evenings while I’m away with my safari.

Tomorrow I’m in Bracknell to have some more coffee with the folks at Monta and talk about an article we had planned. I need to mail several other Embedded Linux guys I know about participating in a feature on the realtime work going in to support the various new uses we’re seeing for Linux through various devices.


Leave a Reply