|
| Greg's diary |
| January 2006 |
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Sunday, 1 January 2006 | Echunga | Images for 1 January 2006 |
| Top of page | ||
| next day | ||
| last day |
The New Year is upon us, and with it a complete lack of information about ADSL connectivity:
After some examination, discovered that the problem was in the calculation of the time zone offset. I do it with the following code:
# Calculate time zone offset from UTC # First get current time. We need this mainly because of DST. NOW=`date +%m%d%H%M.%S` # Convert to seconds at UTC UTC=`TZ=GMT date -j $NOW +%s` # And in local time zone LOCAL=`date -j $NOW +%s` # The difference is the time zone offset. TZOFFSET=`expr $UTC - $LOCAL`
This worked fine most of the time, but the clue was in the date format, %m%d%H%M.%S. This gave, for example, 01010918.30 for the local time. That's fine, but since I didn't specify a year, converting the time back again defaulted to the current year in the time zone: 2005 for UTC and 2006 for Australia. This resulted in a time zone offset of slightly over 12 months, since before I started keeping the records, so nothing was shown. Fixing the problem was straightforward enough:
--- getdslstats 2005/11/03 00:49:19 1.8 +++ getdslstats 2005/12/31 23:22:04 @@ -12,7 +12,7 @@ Y2K=946684800 # Calculate time zone offset from UTC # First get current time. We need this mainly because of DST. -NOW=`date +%m%d%H%M.%S` +NOW=`date +%G%m%d%H%M.%S` # Convert to seconds at UTC UTC=`TZ=GMT date -j $NOW +%s` # And in local time zoneNote the checkin time for the new version: still in 2005 UTC.
What a change in the weather! The last couple of days of last year were really hot, with temperatures in the high 30s. Today it rained all day with a persistence that I have seldom seen in Australia, and the temperature barely made 20°.
As if to protest, our cat Lilac has disappeared. We've had a couple of cats run over over the years, the last in November 2000, but this time seems different: first, she went away and came back again, and though we didn't see her today, the food we left for her last night was gone. Spent some time both looking for her (if she got run over, she did it somewhere where we didn't expect it) and wondering what might have happened.
Didn't do much all day; played around with linphonec and got it to understand ^D, and also tidied up the code, something I really shouldn't do. But why is it that so much recent “open source” software looks so terrible?
| Monday, 2 January 2006 | Echunga | Images for 2 January 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Lilac has still not returned home, but we haven't found any evidence that she's been run over, either. Having searched all places where she might have been locked in, considered the possibility that she might have been bitten by a snake, and set off on horseback to search the (20 hectare) property for us.
That didn't quite work out the way we expected. Somewhere near the south end of the property, a herd of black bulls came running up from the neighbouring property and scared Lady, whom Yvonne was riding. She reared up, lost her balance and fell over backwards, right on top of Yvonne. Yvonne wasn't knocked out, but obviously it hurt a lot. Called the emergency services—how stupid these people are! They don't even have the information about what state I'm calling from, and they're obviously ignorant of the names of the places. Had to spell “Echunga” three times before I was finally connected to the ambulance services. In a life threatening emergency, that could make the difference between life and death.
The ambulance was, on the other hand, very quick; they were with Yvonne in the paddock within 25 minutes of the accident, a very good response considering that we're in the middle of nowhere. She was in some pain, not surprisingly, and after some consideration, they took her off to the Royal Adelaide Hospital for observation:
|
|||||||||||||||||||||||
On the way to the hospital, between Echunga and Hahndorf, ran into the next emergency: a sheep on the road. Sheep are stupid animals at the best of times, but you can rely on them when they're separated from the herd—to do the most stupid thing they could. Called the emergency services again, and again had trouble:
She didn't ask me for any further details; I suspect that she just dropped the call, but I was left wondering whether it's even worth calling the emergency services from a mobile phone. With the help of a passing motorcyclist, managed to get the sheep off the road, and a little later found the owner (it turned out to be a pet lamb) and pointed him in the direction.
One of my main news sources is ABC Radio. Or it used to be. Since about Christmas they've turned it in to an electronic equivalent of a tabloid; instead of international events, we now hear about shootings and accidents that happen in Adelaide. I don't know why ABC should think that anybody is interested in this stultification. Today, however, we were touched by one of the items: they reported that a woman was shot in the groin in Marion and taken to the Royal Adelaide Hospital, presumably the one who was received in an emergency bay right next to Yvonne. She was accompanied by five policemen and three security guards, one of whom took the chair I had been sitting on at the end of Yvonne's stretcher. She (the woman who had been shot) appeared to be high on drugs, and spent a lot of time cursing everybody in sight. It was quite a relief to get the information that Yvonne had nothing beyond some bruising, and got her home by about 7 pm.
While waiting at the hospital, took another opportunity to read the Asterisk book, this time the sections on POTS technology and VoIP protocols. By the time I went to bed, I had finished the two chapters and confirmed my impression that the book is worse than useless. The two chapters supply no new information, refer to things that they neither define nor describe (what's a central office? It's an obfuscatory American word for what the rest of the English-speaking world calls a telephone exchange, as I discovered only 15 years ago when I became involved with Telco work), and it doesn't describe the VoIP protocols at all. What's the difference between SIP and IAX? No idea. How does SIP work? You won't find it here. How do I set up Asterisk to connect to multiple services via SIP? I haven't been able to find any useful information in the book. This is the second book about VoIP by O'Reilly that I have read, and they've both been terrible.
| Tuesday, 3 January 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Slow day. As the doctor predicted, Yvonne was in considerable pain and spent most of the day in bed, so apart from having to look after her, I also needed to do her work. Did that as superficially as possible, but it still kept me busy.
Back to looking at bugs. I have this one bug, both serious and simple to fix, that has been hanging now because I can't reproducibly build the software. Very frustrating.
| Wednesday, 4 January 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Another day with little to show. The gardeners turned up and, amongst other things, found the leak in the underground pipes that had been worrying us for some time, but weren't in a position to fix it.
Yvonne was still in pain, so I had to look after her again, in addition to other things.
In the evening, Lilac returned. It's been over four days since we last saw her. No idea where she's been, but she was in good condition and looked relatively well fed, so it's possible that she's been at one of the neighbours.
| Thursday, 5 January 2006 | Echunga | Images for 5 January 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Busy day today, without much in the way of outcome. Yvonne is gradually recovering from her accident on Monday, but it'll be a while before she's back to her usual self.
| Friday, 6 January 2006 | Echunga | Images for 6 January 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
More planning for travel today. It now looks as if Christine Aker will join Brian, Arjen and myself on the journey to Dunedin in two weeks.
More work on debugging, and finally worked out a few things that aren't as well documented as they should be:
my.cnf is frequently referred to as the defaults file in the documentation.
To specify the defaults file for the server to read, use --defaults-file foo.cnf.
Arjen and I are not sure that the man pages in the bk tree are up to date. Certainly the standard UNIX practice of leaving the stated date way in the past doesn't help.
Arjen and I are agreed that we really need both a man page and a fully-populated sample file. I need to put in a request.
With the aid of all that, managed to start the server the way I want and start working my way through the life of a request. Interesting stuff, and some that gives me food for thought about threading and performance.
| Saturday, 7 January 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Brew day today, which as usual kept me going all day. Today I changed things slightly: apart from using oats (3.5%) in the grist for the first time, I also made a total of 40 litres of wort and fermented them with two different yeasts. This is pushing the limits of my vessels, and I had problems with the wort boiling over; even so, I needed to make a concentrated wort and dilute it with pre-boiled and cooled water before fermenting.
After that, still in the kitchen, tried a recipe from Bonniers kokbok, Sjömansbif, a baked beef and potato dish that is particularly useful for using up stocks of less tasty beer. Great success. We must make more dishes from this book.
| Sunday, 8 January 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
As a result of Yvonne's accident, I'm behind with my projects. Spent some time today on my documentation, but didn't continue it on Wiki: Wiki is just too slow and painful. I'm doing it in good old groff instead, and when I'm done I'll convert it back to Wiki.
This leaves me wondering what good Wiki is. I see it as a reaction against XML. XML is just too complicated for Mere Humans, and Wiki grew up as a more usable alternative. It's interesting to note that this is exactly the way the ad-hoc UNIX markup languages developed; but groff is several orders of magnitude more mature and powerful. It's a pity that it had to be reinvented.
Also more debugging work, walking through the server. Got lost inside the parser; I suppose it's time to use the standard debugging tools. Mañana.
| Monday, 9 January 2006 | Echunga | Images for 9 January 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
As planned, on with my bug analysis today; instead of tracing the server with gdb, planned to use the builtin trace facility. That was a slight problem: the documentation assumes that you start the server with mysqld_safe, and documentation for other methods is less helpful. Didn't get very far before I had to mover on to other work.
That other work was my documentation, which is due on Wednesday. At least that's looking in reasonable shape.
| Tuesday, 10 January 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
More work on the parser bug. Finally found a way to start the server: reading mysql-test/mysql-test-run shows that there's a --debug option (and a lot of other stuff), and with that I got things started. But it's still very frustrating not to know why it doesn't work the way I tried. Found my way up into the parser, after which things became less clear. Got some help, but I'm not there yet.
More work documenting I/O issues. Somehow documentation seems to be an open-ended issue. Still not finished.
| Wednesday, 11 January 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
AVIS have still not responded to my strongly worded letter, despite a promise to do so by 9 January. I wonder how they can stay in business with such practices.
2½ hours on the phone today, all intercontinental! It's nice that VoIP is cheap, but the time doesn't come any cheaper because of it.
More work on the documentation, and finally got a first draft finished. It's brought home to me how little I understand about Microsoft. What I've heard suggests that the most efficient ways to run under Microsoft are not at all closely related to UNIX. We really need somebody with a better understanding of Microsoft to complete this work.
| Thursday, 12 January 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
More work on my I/O efficiency document today; finally took the plunge and investigated what Microsoft does, and came up with some useful information (including that, apart from their horrible syntax, their I/O structure isn't too bad). Also got a chance to find out some of the details of how InnoDB works under Microsoft. That took most of the day, however.
AVIS finally replied to my demands—three days over the limit and without answering any of the issues I raised. They still appear to be assuming that I damaged the car—one of the key issues is that I did not, and they haven't provided anything to substantiate their claims. At least they did it without reminder.
| Friday, 13 January 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Finally the pressure is off. Well, inevitably there are more requests for other information, some of which will potentially cause significant more work, but today I was able to spend some time thinking about online backups. There are a number of questions still to be answered; probably the biggest one is “what is a consistent backup of a non-transactional engine?” I haven't come up with a good idea about that one; suggestions are welcome.
More mail from AVIS. They finally included the invoice —issued in the name of AVIS Australia. They admit that they can't justify the sums, but after over a month, their own very generous (to them) limit time, they're not prepared to refund the money. To make things easier, the bloke who was doing the work has gone on holiday for the rest of the month.
| Saturday, 14 January 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Quiet day. Bottled some beer in the morning, pottered around for the rest of the day, and cooked in the evening.
Lesson for Groggy: When chopping deep-frozen chicken livers with a Chinese cleaver, cover them with plastic foil first.
Lesson for Yvonne: When defrosting deep-frozen chicken livers in a microwave oven, cover them first.
| Sunday, 15 January 2006 | Echunga | Images for 15 January 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Decided to do something completely different today, but discovered that there were so many different things to do that I only got some of them done.
Arjen Lentz is interested in one of my old SPARC boxes, so downloaded the latest NetBSD ISO image (why are they so small? They don't even have packages on them; they're in a separate place), burnt a CD, and installed (on a spare SPARCStation 5). The funny thing was that it Just Worked. Now to see whether I can just drop the disk into an IPC or IPX (both without CD-ROMs).
One of the things that I learnt while travelling Europe in the last few months was that I take too much stuff with me. For my trip to Linux.conf.au I really don't need two laptops. Well, maybe not. My new Inspiron 6000 has a 1920x1200 display, and I need to reduce it to 1024x768 for typical projectors. Set to setting up a corresponding configuration file, only to find that it was completely ignored. I put this in the Screen section:
Section "Screen"
Identifier "Screen0"
Device "Card0"
Monitor "Monitor0"
SubSection "Display"
Viewport 0 0
Modes "1024x768"
Depth 24
EndSubSection
EndSection
And still it started in 1920x1200 mode—only. I couldn't even switch to 1024x768. It took a while to realize that it was still using a default of 16 bits for the display, even though I hadn't told it: I also hadn't told it not to use 16 bits. Managed to fix it like this:
Section "Screen"
Identifier "Screen0"
Device "Card0"
Monitor "Monitor0"
DefaultDepth 24
SubSection "Display"
Viewport 0 0
Modes "1024x768"
Depth 24
EndSubSection
EndSection
This is taking the concept of defaults just a little too far.
My Apple died two days ago, and I didn't notice. This time, though, I couldn't revive it: the “yo yo” power supply is dead. Spent some time looking for a new one, wondering if the cost were justified, and sent out some messages asking for advice. Why are power supplies so expensive? You can buy a brand new 350W ATX power supply for $20, but I couldn't find any replacement for Apple's toy for under $70, and that was a special offer.
| Monday, 16 January 2006 | Echunga | Images for 16 January 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Things are getting saner. Had a number of questions to answer about my I/O document, adding to it in the process. As usual, the difficult questions are more about what to present and how to present it than the technical information. For example, is it relevant to talk about the limited size of the kernel file table (which limits the number of files that can be concurrently open)? I haven't decided yet.
Apart from that, and the inevitable Monday mail, finally got down to writing a requirements document for online backup. We already have a road map, but for some reason I've been having difficulty getting my head around it, and writing a new, differently structured document makes it a lot easier for me to understand. At least I have the feeling that I'm getting somewhere.
With a bit of mail discussion, it's become pretty clear that my PowerBook power supply is dead. Tried turning it on again and... it worked! For about 20 minutes. Then it failed again. There's a lot of goo in the case, probably a dead electrolytic capacitor:
|
||||||||||||||||||||||
Bought a couple of power supplies on eBay, adjustable 13 to 25 V (the Apple needs 24 V) and hopefully powerful enough: they deliver 2.1 A; the Apple power supply is rated at 75 W, which corresponds to 3 A less loss. We'll see whether it works. For $24.64 for the pair, including postage, I'll find other things to do with them, like run my garden sprinkler system, which I still haven't repaired since last summer.
Finally did something about the sprinklers, running them manually until the power supplies arrive. It's been a moist summer, and so far we haven't needed them (normally we'd have had to turn them on in early December). What a lot of mess collects in these things!
| Tuesday, 17 January 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Frustrating day. I had an appointment in town in the afternoon, and somehow that messes up the whole day. Spent the morning picking at my backup design document, without making any tangible progress.
In town, nothing of use happened. Ramana tried to access a web site, an important part of the meeting, but the site (run by Macquarie Bank) kept crashing. I'm pleased to note that it appears to be Microsoft-based, but it's still frustrating to spend most of an afternoon doing nothing when there's enough work waiting at home.
Back in November I renewed my main domain, lemis.com, with GANDI. The update still hasn't shown, though I was able to confirm that the money has been debited. Sent them an “email” (i.e. filled out a toy web form). Hopefully they'll do something about it; I only have 2 weeks.
By the time I got back from town, it was 16:30. Didn't feel like doing any more work on my design document, so grabbed a SPARCstation IPX as a possible alternative to give to Arjen. To my surprise, the thing had 224 MB of memory in it, more than any other SPARC I know about here. Put in a SCSI-2 disk, but it didn't locate it. I don't know if I can be bothered; let him have the SPARCstation 5 instead.
| Wednesday, 18 January 2006 | Echunga | Images for 18 January 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Another day from hell! Slept badly, and the carpenters arrived at 7:40 am to replace the eaves which had had to be damaged when replacing the roller shutter belts over a month ago—they had promised to come immediately after that. Shortly after that, the electricians arrived to examine why the generator wasn't connected to the power outlets in the Mike Smith Memorial Room any more. Before finding obvious wiring faults in the switchboard—how can professional electricians make such basic errors in such simple circuits?—they managed to dispose of a screwdriver in a spectacular fashion:
|
||||||||||||||||||||||
The other end of that was the switch (also destroyed) that connected my office power (via UPS) to either generator or mains, so it had the side effect of taking power from the UPS. They managed to bridge that with about 3 minutes of UPS time to spare. More work to come, obviously.
In parallel with that were the gardeners, who attended to the sprinkler system. We've only just started using it again this year, and in the meantime many things need adjustment. Discovered later in the day why the lawn is still so green: we have another leak in the underground water system (Wayne Graetz only fixed the last one last week). This one seems to be my own doing: it's the one I “fixed” in November 1998. The appearance has changed a lot since then.
In the afternoon, a painter arrived to take a look at the work that needs to be done; fortunately that didn't take too long.
I'm off to New Zealand for Linux.conf.au next week, and there's lots of mail going round the lists. One useful bit of information: New Zealand only has one mobile phone provider (Vodaphone), who are thus able to charge what they want. It would cost me about AUD 4.00 per minute to call back to Australia, and it even costs AUD 0.65 to receive a call—more than I pay here to make one. So spent some time trying again to get linphonec to work on eucla, my new laptop. I had set it up previously, but I couldn't get the microphone input to work. With some help from Daniel O'Connor, got that working: to use it, you need the following commands:
=== grog@eucla (/dev/ttyp6) ~ 5 -> mixer rec 80Setting the mixer rec from 0:0 to 80:80.=== grog@eucla (/dev/ttyp6) ~ 6 -> mixer =rec micRecording source: mic=== grog@eucla (/dev/ttyp6) ~ 7 -> mixer mic 0Setting the mixer mic from 70:70 to 0:0.=== grog@eucla (/dev/ttyp6) ~ 8 -> mixerMixer vol is currently set to 72:72 Mixer pcm is currently set to 80:80 Mixer speaker is currently set to 0:0 Mixer line is currently set to 100:100 Mixer mic is currently set to 0:0 Mixer cd is currently set to 75:75 Mixer rec is currently set to 80:80 Mixer ogain is currently set to 100:100 Mixer line1 is currently set to 75:75 Mixer phin is currently set to 70:70 Mixer phout is currently set to 0:0 Mixer video is currently set to 75:75 Recording source: mic=== grog@eucla (/dev/ttyp6) ~ 9 ->
This is the first that I have seen of the =rec mic syntax. It appears to have occurred since I last looked, quite some time ago.
Despite all that, still found some time to work on my requirements document, where some of my problems were with the technology: my new Brother laser printer has one difference from the other: it complains if you ask for a different paper size (typically US “letter” instead of ISO A4)—on every page. So far I've been unable to turn it off. There's a configuration setting Page Protection, which the typically superficial documentation implies might have something to do with it. I've set it to off, but it still doesn't work.
Spent some time looking at groff trying to work out why it always told the printer that the paper was “letter”; it obviously had nothing to do with the specifications I gave when running groff, and there was nothing obvious in the man pages. Finally found it: the file (typically) /usr/share/groff_font/devps/DESC contains a paper size specification. This patch fixes it, at least for me:
--- /usr/share/groff_font/devps/DESC 2004/05/08 03:13:27 1.1 +++ /usr/share/groff_font/devps/DESC 2006/01/18 05:37:14 @@ -10,5 +10,5 @@ tcommand postpro grops broken 7 -papersize letter +papersize a4 print lpr
AVIS are thoroughly getting on my nerves. Called up today to speak to Denise Fogasavaii. who is supposed to be an the case. She knew nothing about the incident, and claimed not to have received the email message which I copied her on last Friday, and in which I asked for immediate resolution. She was quiet for a while, then claimed to have “Just received mail from our counterparts”. She promised to forward it to me.
In the meantime, called up the bank, who will send me another dispute form, and the Office of Consumer and Business Affairs (phone number (08) 8204 9777), where I was advised to let the bank take its course first. They also seem to be the best people to handle potential issues of fraud on the part of AVIS.
90 minutes later, by which time she was no longer in the office, I received a stupid message from Denise which made it pretty clear that she hadn't read the message I had sent her: they continue to talk about excesses, while I'm asking them to explain what damage they are claiming. What incompetence! Sent her a reply asking her to finally read and reply to my previous message. But how much is that going to help if she doesn't read it in the first place?
I find it baffling that such incompetence can exist in a high-profile, world-renowned company (never mind the slogan “We try harder”, which really seems out of place now). If any reader has a suggestion about how to proceed by civil means, I'd be grateful. I'm keeping a possibly interesting online record of this whole sorry mess.
Also got a call from Dell to organize the repairs to my laptop. He was in Malaysia, so he couldn't do very much about it except to “organize a technician”, something that I had already had promised two days ago. He also didn't know about the damage to the PC card ejector. Basically, he contributed nothing.
| Thursday, 19 January 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Finally getting back to normal. Hopefully I can now get down to work. Spent more time going through the requirements document than I expected, but the end is now in sight.
Another phone call from Dell this afternoon to organize the repairs to my laptop, this time from Sathyan. He said effectively the same thing as the bloke yesterday, and that somebody would contact me tomorrow. That's now been three days since Teh said the same thing. He also confirmed that the technician would have the material to repair the ejector for my PC-Card slot, in a manner that makes me suspect he didn't understand the question.
| Friday, 20 January 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
More work on the requirements document today, including a massive two-hour phone call to discuss it. Ended up with a lot of scribbled notes; I wonder when I'll find time to incorporate them into the document, seeing as though I'm off to Linux.conf.au tomorrow.
Finally got a phone call from Charles, the local Dell technician, at 15:31. He called on my mobile phone despite my request not to, and left a message. From what he said, it appears that he didn't know anything about the ejector. Clearly he wasn't able to do it this week. I'll see if he calls back again next week.
| Saturday, 21 January 2006 | Echunga –> Brisbane | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
The flight to Brisbane was only a couple of hours, but somehow it took all day, with preparations. One of those was setting up the SPARCstation 5 for Arjen, not helped by the fact that I couldn't find my 3R13 to VGA adaptor. Spent another half an hour setting up the thing enough that it would come up on the network without a head, a time comparable with setting it up in the first place.
Off to Brisbane and to Arjen's, having to help the taxi driver find the place. Set up the SPARCstation without too much difficulty, once we got the disk connected correctly. Brian and Christine arrived, and spent the rest of the evening eating and talking. Far too late to bed.
| Sunday, 22 January 2006 | Brisbane –> Dunedin | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Up in the middle of the night today, after little more than 4 hours' sleep for a 5 am taxi to the airport: Freedomair insists on people being at the airport 2½ hours before takeoff. Had breakfast at the airport and then waited in the plane 30 minutes after scheduled takeoff waiting for latecomers and wondering how that compared with the requirement to be at the airport so long in advance.
Slow flight to Dunedin: instead of the accustomed tail winds, there were head winds, with the net result that we arrived 65 minutes late. Into Dunedin—pretty town—and off to look for something to eat. What with the lack of food on the flight and the time difference (3 hours), it was 7 pm before we had “lunch“, in a restaurant with Moroccan and Greek pretensions in the Octagon. “We” were the MySQL delegation: apart from Arjen, Brian, Christine and myself, also Stewart Smith and Morgan Tocker. Had a “lamb tagine”; I should have been warned by the fact that the spelt it “tangine“. Not good, and not enough to eat.
Off after that with Stewart to a pub where we met up with some other people from the conference, including Andrew McMillan of Catalyst, who are doing some high-profile database stuff using open source; they're running the New Zealand electoral roll on PostgreSQL. Interesting stuff.
| Monday, 23 January 2006 | Dunedin | Images for 23 January 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
First day of the Linux.conf.au miniconferences today, and started with a VoIP tutorial by Ewen McNeill, located in Wellington. There seem to be a number of clever people there. The tutorial was exactly what I had been looking for (and complaining vociferously) in the past: an overview. Of course it couldn't go into enough detail in the hour allotted, but it was full of links, some of them presumably good.
Went to a couple of other talks, and then into town with Arjen and Stewart for lunch at the Curry Box, not a place I'm liable to return too. The rice was well cooked, but the food was very westernized and rather lacking in enthusiasm. Still, there are a lot of places round there that promise good food.
In the afternoon worked on my presentation; I'm planning a surprise or two. Things finished off early, and spent a while waiting for people to decide to do something interesting. Finally down to the local pub to try still more of the local beer (Emersons). They have a surprising variety, and they're planning a brewery visit tomorrow.
| Tuesday, 24 January 2006 | Dunedin | Images for 24 January 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Into a talk this morning by Edward Murphy (my ex-contact at iHug Satellites) and Simon Lyall entitled A few useful Linux tools you might not know about. The name is appropriate; they had tested a whole lot of packages that I might have tried had I had time. Useful to have somebody else doing the work.
Unfortunately, in the middle, while duplicating stuff back home, the network connection back home failed and stayed down. Further investigation showed that only the link home had failed. Called Yvonne, who confirmed that the new UPS had failed (“Err 13”). Damn. Spent some time getting things up, somewhat confused by an incomplete firewall initialization.
After that, had lunch at a Korean restaurant that Chris Yeoh had recommended; not bad, but slower service than we could have done with. As a result, Morgan had to leave early to deliver his paper. I arrived 15 minutes after the scheduled start time to discover that there had been a fire alarm, and everything had been postponed by 30 minutes. The paper was a good overview; it's interesting to note how much interest MySQL papers arouse.
After that off to Emersons brewery to take a look around. It's a relatively small operation:
|
|||||||||||||||||||||||
They've just moved to larger premises, and still haven't come up to speed. They have a capacity of 140 hl per week, but are still currently only delivering about 300 hl per month.
In the evening the speaker's dinner, the first time in a (high-class) pizza place.
|
|||||||||||||||||||||||
Quite good food, but the noise level was high enough to make me hoarse.
| Wednesday, 25 January 2006 | Dunedin | Images for 25 January 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
First day of the main conference today, and started off with an Introduction to Asterisk, which got off to a slow start: first they had to wake the presenter. Then they couldn't get his laptop to talk to the projector, so the first half was without video. Apart from that, it wasn't the best talk I've heard on the topic, and he said a number of things that didn't correspond with my experience. For example, he said that VoIP soft phones and 802.11 were both individually unsuited to VoIP, and that there was no way to solve the problem. He also said that you needed really good headsets, and that the average Sennheiser headset wouldn't work. I happen to have a Sennheiser, and I have had no problem with it (nor with the el cheapo ones that he says are completely useless), and in the course of the day I used linphonec over 802.11 and got surprisingly good communications quality. The real problems are in the user interface, not the quality.
After that, off to lunch at a Malaysian place that I will remember so that I can ensure that I'll never go there again. Not all restaurants here are good.
After lunch spent some time working on my talk; some of the things I'm doing with graphics require updating my macros. Then to the Linux Australia AGM, after which spoke with Matt Mackall, the originator of Mercurial. A number of people, including Rusty Russell and Ted T'so, are using it and are happy with it. Brian Aker is less convinced, but what I've heard all sounds good.
Then off to dinner with Stephen Rothwell, Ben Elliston and Arndt Baumeister, to another Korean restaurant. It's not as good as the one we went to yesterday. Back to Unicol, intent on an early night, but first sat down in the lobby (where there's wireless networking), and ended up talking with people as they came in; ended up getting to bed later than the previous nights.
| Thursday, 26 January 2006 | Dunedin | Images for 26 January 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Any intentions of sleeping in this morning were shattered at 8 am sharp by a lot of jack hammers being started outside my bedroom. During breakfast discovered that—of course—we had had a power failure back home, and again battunga hadn't come up. That was a pointer to the firewall problems: I only allow ssh through to two specific systems, and I had specified them by name. And battunga was the name server. Fixed that, and hopefully we won't have any further problems.
After that did some more work on my presentation, mainly image manipulation. Then to Paul McKenney's talk on “Steamroller testing”, and then missed the next paper while manipulating images. The issue was the MySQL logo, which was in an EPS format that broke both groff and xv. Installed gimp, with the surprising discoveries that the installation worked and that gimp was able to correctly transform the EPS into a PostScript file.
In the afternoon to listen to Arjen talk about Concurrency control in RDBMS. He had shown me the slides earlier, and I had been amazed how little material he wanted to cover in his allotted 2 hours. I was wrong: he didn't get through it. Clearly the concepts are of interest, but difficult to understand.
Then the VoIP BoF, which didn't really have enough time to talk about much before leaving by bus for the professional delegates' “networking” session at Larnach castle, very picturesque, with a piper to meet us:
|
|||||||||||||||||||||||
Late to bed without having had any real food to eat.
| Friday, 27 January 2006 | Dunedin | Images for 27 January 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Started the day trying to set up my laptop to display on the projector for my talk this afternoon, not successfully. There's something I don't understand about the dual headedness of this chip set, and x.org seems to ignore parameters when it suits it. Nothing I did could change the horizontal frequency, which was out of range for the projector. Ended up copying the talk to a Microsoft box. Not happy.
The first paper was Brian Aker talking about hacking MySQL internals. There's not much you can say on the topic in an hour, but got some useful information.
Then on with MySQL with Stewart Smith talking about the NDB storage engine. What I didn't know was that he wanted to do some practical work, and I didn't have an appropriate installation on my laptop—shades of my recent startup problems again. Went to one of the internal machines, but the build I had there wasn't appropriate for the workshop.
Then lunch, far too short again—there should be at least a 90 minute lunch break at these conferences—and then back to listen to Maddog talking about “open sores”, in fact very similar to an earlier talk I had heard. After that came Van Jacobson with a modest proposal to help speed up and scale up the linux networking stack—only an fivefold speed increase! Finished with a standing ovation, after which I had to run off to my own presentation, which went well enough given the video problems.
Conference dinner in the evening, sitting with Van Jacobson, Paul McKenney, Hollis Blanchard, Ted T'so, Jim Gettys and two others whose name I forget. By chance talked about ADSL connections, and discovered that neither Van, Ted nor Jim had one: they're not available where they live (Van on Skyline behind Palo Alto near where Kevin Coughlin used to live, Ted near Boston and Jim in a place I didn't recognize, but presumably also in the Boston area). They're all much more built-up than where I live. Van and Jim are using cable instead (something not available here) and getting higher data rates than Telstra is currently supplying, round 5 Mb/s. Ted is stuck on 512 kb/s SDSL.
Maddog started auctioning off a copy of John Lions' Commentary on the Sixth Edition UNIX Operating System, signed by just about everybody, including notably Ken, Dennis, Kirk, Eric, Peter Salus and Linus. The proceeds are intended for the new John Lions chair of computer science at UNSW. Rusty ended up pushing him aside and proceeding to raise a record $10,000, in the process getting a lot of collateral promises, including shaving two beards (David Miller's and mine), a moustache (Rusty Russell) and hair (Jeff Waugh) if the final sum passed various sums <= $10,000. That'll be fun.
| Saturday, 28 January 2006 | Dunedin | Images for 28 January 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
The last day of the conference started with a talk by Mark Shuttleworth about cooperation between projects. Good stuff, but I was expecting a bit more fireworks. After that went to listen to Matthew Garrett talking about ACPI, which left me wanting to hear more.
After that, a barbecue lunch and a depilation ceremony. David Miller got shaved first, followed by myself:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Unfortunately the battery of the camera failed before any shaving got done, but there are
many on the web. I'm aware of photos taken by Marc
Merlin,
Horms, Dave Miller,
Gustavo Sverzut
Barbieri and
Davyd
Madeley.
Looking back five years later, it's sad to see how many of these links have died or dried up. I don't have any alternatives. If you know, please contact me.
After me, Rusty shaved his own moustache (and promptly became unrecognizable), and then Maddog shaved Jeff Waugh's head. They auctioned the right to shave. My beard got $50 (much more than I got for it on eBay two years ago), and Maddog paid $150 for Jeff's scalp. I forget how much was paid for Dave Miller, and of course Rusty spoilt the fun by doing it himself.
After conference close, down to the Captain Cook pub where we (MySQL) had invited people for a drink. That went well, and afterwards a lot of us went into town to Ananda, an Indian thali restaurant that I had been looking at for some time. I wasn't disappointed.
| Sunday, 29 January 2006 | Dunedin –> Brisbane –> Adelaide –> Echunga | Images for 29 January 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Up particularly early this morning for the speakers' event, since Brian wanted to go on a train journey afterwards. Who didn't show up? Brian. Arjen sent him an SMS and got the distressing news that he had “food poisoning”, not the only one: it seems that a lot of people had it.
At breakfast this morning, talking with Stewart Smith (Linux Australia Committee member), Jonathon Coombes (AUUG Secretary) and Maddog Hall about Linux Australia and AUUG. I told Maddog that I was an ordinary committee member of AUUG. Maddog: “Greg, there'll never be anything ordinary about you”. That went down like butter.
Off on the helicopter tour round the estuary and back. The helicopter flight would have been fun at any time, but the landscape here is particularly pretty, and the weather was perfect, not a cloud in the sky:
|
|||||||||||||||||||||||
After that, discussed with Arjen what to do. We had heard from Brian that he was not in serious danger, but he didn't want to do anything, so decided to leave today and get back to do some Real Work. Changing flights was mercifully easier than booking them. While at Unicol was forcibly faced with the unappetizing sight of one of the people affected by the “food posioning”, sitting on the floor with a sick bag. Waited until he had finished and confirmed that he had seen a doctor, and then off with the others first to lunch and then to the airport.
At the airport, bumped into Pia Waugh and some others who were travelling back with Mark Shuttleworth in his private jet. She too had been taken down with the bug early in the morning, but was now feeling better, which is a good sign. Then off to Brisbane, passing over the Southern Alps:
|
|||||||||||||||||||||||
The same unusual weather has been prevailing all week. When we came down, it delayed us by 45 minutes. Today it sped us up by the same time, just in time for Morgan to leave the plane before the stomach bug hit him. Went off looking for a first aid person and discovered that the officials at Brisbane airport were woefully prepared for the event; it took about 5 minutes for them to even work out how to handle the issue. Arjen and I were left wondering what would have happened in a real emergency.
Apart from that, Brisbane airport left me less than pleased. Followed the sign to “domestic transfer” and discovered that it was only for specific airlines; there was little help for people who weren't booked through. Finally found my way to the train, where I had to pay the comparatively princely sum of $4 to get to the domestic terminal. This is the only airport I know that charges anything at all, and $4 is more than the return bus fare from Echunga to Adelaide. Not a good advertisement. At the domestic terminal, we were, of course, obliged to remove laptops from their bags (which isn't required in New Zealand). Virgin tried to make it more acceptable with this sign:
|
|||||||||||||||||||||||
As usual, they get things wrong. That also upset the security staff, who saw a security breach (it's obviously in the area of the security checks), and who didn't understand that the sign invited people to take photos of it. As I said, it depends on your imagination.
The security guard handed me a well-worn laminated A4 sheet telling me generic stuff about why I had been “randomly” selected. I later discovered that I took it with me, which was clearly not the intention.
After that, I was “randomly” selected for an inspection of my laptop with one of these sniffers which I suspect don't do anything useful. At least he was businesslike about it.
Back home without much further ado, apart from the stewardess being unable to close the plane door, a potentially serious issue which doesn't reflect well on Virgin Blue.
| Monday, 30 January 2006 | Echunga | Images for 30 January 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Back to work today, first attacking the 3000 odd mail messages that I have accumulated over the past week. One thread was the question of “food poisoning”, which now seems likely to be a viral infection (norovirus), probably brought in from Christchurch by a select few, and nothing to do with any food served at the conference. A surprising number of people have been infected.
Also surprising is the amount of technical stuff that has failed over the last week: the new UPS that failed last Tuesday and the pressure pump for the garden sprinklers with a molten power connector:
|
||||||||||||||||||||||
In addition we have another wet patch in the garden; still more underground pipes to fix. And two phone lines failed completely again; this time it proved to be an issue with the pair gain system which runs them. Maybe Telstra will get that one sorted out before their Customer Service Guarantee expires.
| Tuesday, 31 January 2006 | Echunga | Images for 31 January 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Another day full of interruptions. I've been waiting impatiently for the release of the slides that Van Jacobson presented last week, and finally sent out a message to the LCA list asking where they were. They weren't on the web, it seems: Van replied with slides and the request to put the up there, so put them up in my documentation pages. That generated a lot of interest.
Charles from Dell came along. He didn't have the replacement parts for the PC Card ejector: he hadn't been informed that there was a problem. He did have a new lid for the laptop, though, nicely covered in plastic film:
|
||||||||||||||||||||||
Two more bad points for Dell: supplying a used machine as “new”, and not sending the parts despite multiple reminders. Now I'm going to have to call them up all over again.
Another Charles showed up shortly after that, this time from Telstra. He was able to fix the pair gain system relatively quickly; for once they've upheld their Customer Service Guarantee.
In the afternoon, on IRC, discovered that my attempts to run ndb during Stewart's tutorial on Friday hadn't passed without trace:
<lars> groggy|rep: Hi. Hope it is ok if I kill your ndb and mysql processes on rpl-a and rpl-b, they seem old... <groggy> lars: Sure :-) <lars> groggy|rep: thx <groggy> lars: I had some trouble last Thursday. Sorry I left junk behind.
I've been given authorization to hire somebody, maybe even two people, for the online backup team. We have a job advertisement for Senior Software Developers on the web site, but that doesn't seem to be attracting enough attention. Spent some time pondering how to improve on that. There are so many people out there, many of them looking for jobs, and yet I haven't been able to find anybody who would fit.
Decided to look at a couple of the myriad bugs allocated to me in the last few days, and ran into trouble with bitkeeper: despite correct keys, I kept getting the message
pull bk-internal.mysql.com:/home/bk/mysql-5.0 -> file://src/mysql/mysql-5.0 ================================= error ================================= your license for bitkeeper has expired. you are beyond the grace period and you may not create any changesets until you have contacted bitmover to purchase a license. ============================================================================
The problem was, these were the same keys that other employees of MySQL use. Spent some time discussing the problem with Brian and Elliot, without resolution, so sent in a support request through their tiny support form.
Do you have a comment about something I have written? This is a diary, not a “blog”, and there is deliberately no provision for directly adding comments. It's also not a vehicle for third-party content. But I welcome feedback and try to reply to all messages I receive. See the diary overview for more details. If you do send me a message relating to something I have written, please indicate whether you'd prefer me not to mention your name. Otherwise I'll assume that it's OK to do so.
| Top of page | Previous month | Greg's home page | Today's diary entry | Next month | Greg's photos | Copyright information |
