Greg's diary--July 2001

Sunday, 1 July 2001	Boston –> Austin
		Top of page
		next day
		last day

I had meant to catch up on my sleep this morning, but somehow it didn't work, and I was up before my alarm would have gone off. Spent the morning catching up on mail and other documentation, then decided it would be a good idea to get sound working on sydney, my Dell Inspiron 7500, since I had bought a couple of CDs on Wednesday. That proved to be more difficult than I thought: even on -CURRENT it didn't work, though everything looked OK. Finally it dawned on me: there's a real, hardware volume control wheel on the left of the machine, just below the PCMCIA slots, and it was turned off. How times have changed.

Off to the airport, and armed with my new knowledge of how to authenticate with MobileStar, was connected within a minute. I wonder if anybody else of their customers has used their service over here. Monitored the air waves for other signs of life; it reminds me of “On the beach”. As in that film, there was little evidence. I'd guess that the chance of anybody being “logged in” at any one moment in any specific airport lounge is about 20%.

The plane out of Boston was delayed 90 minutes due to a storm. While waiting around, I noticed a bloke who looked surprisingly familiar, and I was reminded that I had been supposed to talk to Peter Salus during the conference, in order to invite him to the AUUG2001 conference. I confirmed that it was, indeed, Peter Salus, and that he was prepared to come to Australia. He changed seats, and we had a more interesting time on the way to Austin as a result.

Monday, 2 July 2001	Austin
		Top of page
		previous day
		next day
		last day

Off to IBM after breakfast, and was able to locate both Mark Peloquin and Steve Best, the latter of whom had time for me. They're all located in the same area. The most interesting thing about Steve's JFS project is the reason they chose the OS/2 version as the starting point. We had all assumed that it was a bit of a copout because they didn't want to release the really high performance AIX version, but the real reason was that the old AIX JFS was being phased out, and the new one, also based on the OS/2 version, wasn't ready yet.

Shopping in the evening. I've decided I need a new combination laptop and overnight bag, and that took a bit of looking for. Ate at the Taj Palace, which seems to have changed little over the years.

Tuesday, 3 July 2001	Austin
		Top of page
		previous day
		next day
		last day

Spent most of the day talking to Mark Peloquin and friends about EVMS. Also had time to describe and demonstrate Vinum, which on closer examination competes surprisingly well with the other open source volume managers out there. It's the only one with RAID-5, for example. EVMS looks like it will be a very strong contender when it's finished, however.

Ate at a Korean place on 1325 almost opposite IBM. Food wasn't too bad for America.

Wednesday, 4 July 2001	Austin
		Top of page
		previous day
		next day
		last day

Up late this morning, and ended up wondering whether to have breakfast or find somewhere for brunch. I chose the latter, which proved to be a bad idea: most restaurants, including the Fonda San Miguel, were closed, and I ended up at Tien Hong on Burnet, which is good for Dim Sum, but today it was so empty that it wasn't worth the trouble, and I ended up with a normal boring lunch.

Then out to Jerry Dunham's place, where we spent the afternoon getting his new machine into reasonable shape. In the process discovered the secret of running multiple X servers, somewhat marred by the fact that the machine kept crashing, probably because of thermal issues. Ended up dropping core and I/O voltage to 1.6 and 3.3V respectively, which gave it enough stability to do a make world (in only 52 minutes; not bad).

Ate at a pseudo-Italian restaurant next to the Outback Steakhouse, apparently run by the same mob. Not too bad, though I missed the Fonda San Miguel.

Thursday, 5 July 2001	Austin –> Los Angeles –>
		Top of page
		previous day
		next day
		last day

Another funny day. I didn't have much to do until the afternoon, so spent the morning catching up on copious outstanding mail and doing some last-minute shopping. Austin traffic has become really congested. It's difficult to think that it's much smaller than Adelaide, and that it has a much more extensive road network. Where do the people all come from?

In the afternoon to meet Dwayne McConnell, who had come back early from holiday specifically to meet me. That was well worth while; it gives me a much better perspective of what is going on, and I also got some useful docco from him.

Off relatively early to the airport so I could use the network connectivity there. Again I had no problems authenticating (Skynet Global have still not told me how to do it, and I suspect nobody knows). Again there was nobody else active in the lounge, though just before I left I saw somebody with a laptop and a suspicious looking card. MobileStar have one annoying habit, though: before authentication, their firewall sends incorrect replies to the SMTP server, instead of just sending a “destination unreachable” ICMP. As a result, the mail fails and I have to bounce it again when the link is up. Sent a second message to them asking them to do something about it.

Friday, 6 July 2001	–>
		Top of page
		previous day
		next day
		last day

Another day lost over the Pacific.

Saturday, 7 July 2001	–> Sydney –> Echunga
		Top of page
		previous day
		next day
		last day

An uneventful flight. Got to Sydney and the Customs people found that the pills I had bought for my father (in England), available over the counter in the USA, are prohibited in Australia, and confiscated them. Grrr. It makes me look like a criminal, when in fact there's no indication that there's anything wrong with the things.

To the Qantas club to get back on the net and send my mail. SkyNet Global have messed up their authentication yet again, and I wasn't able to log in. Called their support people, who said they would do something, but of course they didn't by the time I took off. I have been able to use this “service” about one time out of three. This is really annoying me. Got details of the people at Qantas to contact about it. Once again confirmed that nobody was using the service. I suspect this time it's a network problem: a traceroute shows:

$ traceroute -n www.lemis.com
traceroute to www.lemis.com (206.239.198.97), 30 hops max, 40 byte packets
 1  10.2.21.1  4.374 ms  4.188 ms  4.277 ms
 2  10.2.253.250  31.691 ms  32.858 ms  31.611 ms
 3  * * *
 4  * * *
 5  * * *

Previous traceroutes before logging in went a lot further:

$ traceroute -n 192.109.197.137
traceroute to 192.109.197.137 (192.109.197.137), 30 hops max, 40 byte packets
 1  10.2.21.1  4.805 ms  4.224 ms  4.105 ms
 2  10.2.253.250  20.034 ms  18.837 ms  18.660 ms
 3  10.250.0.9  19.803 ms  20.328 ms  19.665 ms
 4  10.2.253.252  21.050 ms  21.548 ms  21.079 ms
 5  *^C

Back in Adelaide, on a hunch, fired up my laptop while waiting for the baggage. The baggage claim area is just downstairs from the Qantas club, and sure enough, it worked fine. And I was not able to authenticate.

In catchup mode in the afternoon. It was cool and wet, temperatures around 10°, but after being in the hot summer in the USA, I felt overly warm inside.

Sunday, 8 July 2001	Echunga
		Top of page
		previous day
		next day
		last day

More catching up to do. Somehow all the flame wars start when I'm not in a position to do anything about it.

Yana is going to Tellurian for two weeks for work experience. Spent a lot of time preparing for that.

Monday, 9 July 2001	Echunga
		Top of page
		previous day
		next day
		last day

Into Adelaide to take Yana to her work experience. Saw Martin Leadbeater first. He has a PowerBook with MacOS X, and we spent a bit of time looking at that, the first time I've looked at it in any detail. It looks quite good.

Dan in and got Yana started. We seem to have ssh compatibility problems. Grr.

Then to IBM and met Hugh Blemings and Alan Modra, also got my dog badge and some AIX CD-ROMs. While we were there, discovered that our new RS/6000 workstations had arrived and had both been sent on to Alan's place, way out in Golden Grove. To lunch with some of the local IBM people, and then out with Alan to pick up my workstation. Back home via Cudlee Creek, and spent some time setting up the machine.

Tuesday, 10 July 2001	Echunga
		Top of page
		previous day
		next day
		last day

Another day spent chasing my mail. Also spent more time with the RS/6000, which I have called sphinx. It seems that it doesn't come with a compiler, which is rather sad.

Booted up with Greg Rodgers' Linux disk, which went fine, modulo the fact that my machine didn't have the same PCI configuration, so I had to change the /etc/fstab, which mercifully worked without any problems.

Wednesday, 11 July 2001	Echunga	Images for 11 July 2001
		Top of page
		previous day
		next day
		last day

Another day of thrashing. I really need to get my paper for FAST finished, but so much happened that I didn't get around to it. At the last minute we got a couple of high-profile applicants for AUUG2001, and spent a bit of time deciding whether we could accept them or not.

In the afternoon to Tellurian and spoke with Dan Shearer and David Newall, then with Yana to the SAGE-AU quiz night, which was a little disappointing. I had to leave early because of a teleconference tomorrow morning, but was rather disappointed by the food, which wasn't what was advertised. I suspect the SAGE people were a bit surprised too.

17 August 2012: No mention of the people I was sitting with. Two of the people there were speakers. As I recall, one was Elizabeth Zwicky, who was quite interesting. The other was a rather quiet person with little to say, but I had expected him to be interesting. Unless I'm very much mistaken, this was Monty Widenius, who later told me that he had had great difficulties because at the time he didn't speak English very well.

9 May 2013: I found I had taken photos of the dinner. They're not very good, but they clearly show that the person next to Elizabeth was not Monty:

Image title: Dinner 1          Dimensions:          1536 x 2048, 370 kB

Make a single page with this image Hide this image

Make this image a thumbnail Make thumbnails of all images on this page

Make this image small again Display small version of all images on this page

All images taken on Wednesday, 11 July 2001, thumbnails          All images taken on Wednesday, 11 July 2001, small

Diary entry for Wednesday, 11 July 2001 Complete exposure details

That's Elizabeth second from the right. I don't know who is to her right. But the next photo, unfortunately out of focus, shows somebody who almost certainly is Monty, at the extreme right. You can even imagine the MySQL on his T-shirt:

Image title: Dinner 2          Dimensions:          1536 x 2048, 277 kB

Make a single page with this image Hide this image

Make this image a thumbnail Make thumbnails of all images on this page

Make this image small again Display small version of all images on this page

All images taken on Wednesday, 11 July 2001, thumbnails          All images taken on Wednesday, 11 July 2001, small

Diary entry for Wednesday, 11 July 2001 Complete exposure details

Image title: Dinner 2 detail

Complete exposure details

Dimensions: 287 x 712, 18 kB

Dimensions of original: 287 x 712, 18 kB

Display this image:
thumbnail    hidden   alone on page

Display all images on this page as:
thumbnails    this size

Show for Wednesday, 11 July 2001:
thumbnails    small images    diary entry

Thursday, 12 July 2001	Echunga
		Top of page
		previous day
		next day
		last day

Finally got some work done on my paper, but somehow other activities are taking up too much of my time. I think I need to drop some of them.

Friday, 13 July 2001	Echunga
		Top of page
		previous day
		next day
		last day

Finally I'm making some headway with my paper. Good thing too, it's due today (really tomorrow, since it's US time).

Also managed to get some laptops updated. Things are looking easier now.

Saturday, 14 July 2001	Echunga	Images for 14 July 2001
		Top of page
		previous day
		next day
		last day

BSD barbecue today. Off to Hahndorf with Yana to pick up some bread, and when we got back Luke Mewburn and wife were already there. Hacked around with various stuff for a while until the rest arrived, then had lunch, after which we decided to hack on my Tivo: Geoffrey Bennett has one, and he had all the software on his laptop.

We had difficulty getting the thing into serial console mode, so I decided to get a TV to attach, went out to the shed, and in the process discovered that our Davey water pump, last date of failure 1 September 2000, had failed again (it had sheared off the bolts holding the housing together) and was in the process of draining the water tank:

https://lemis.nyc3.digitaloceanspaces.com/grog/Photos/20010714/big/Burst-pump-1.jpeg

Image title:	Burst pump 1	Dimensions:	2048 x 1536, 688 kB
Make a single page with this image		Hide this image
Make this image a thumbnail		Make thumbnails of all images on this page
Make this image small again		Display small version of all images on this page
All images taken on Saturday, 14 July 2001, thumbnails		All images taken on Saturday, 14 July 2001, small
Diary entry for Saturday, 14 July 2001		Complete exposure details

In a foul mood replaced it with the one from the sprinkler system. Davey is going to have to give me some good answers on this one. There are more photos here.

Apart from that, quite a pleasant day. I gave up on the Tivo, but it seems that Geoffrey is planning to talk about it at the LinuxSA meeting on Tuesday, so I'll bring mine along to try out.

Late night talking to Luke. We agreed that we need to get more cooperation between the NetBSD and FreeBSD core teams.

Sunday, 15 July 2001	Echunga
		Top of page
		previous day
		next day
		last day

Finally the pressure is letting off. I had intended to do nothing at all today, but unfortunately my spate of laptop updates over the last two days left Yana's machine unable to start X or run her Ethernet card, so I ended up having to give her mine. Then I discovered that the two I was left with both couldn't mount my camera flash card. sigh: the joys of running -CURRENT.

In the afternoon out riding, even more briefly than last time. Yvonne was riding La Tigre, who is still recovering from the kick she got a month ago, so we couldn't go very far. It was enough to tire me out, though, and by 8:30 pm I was in bed.

Monday, 16 July 2001	Echunga
		Top of page
		previous day
		next day
		last day

Four years ago today we moved into Wantadilla. Things have changed a bit since then.

I later decided that the date was 17 July 1997.

Finally got a bit of infrastructure work done for Linux PPC. We're using Bitkeeper, which takes a bit of getting used to.

Tuesday, 17 July 2001	Echunga
		Top of page
		previous day
		next day
		last day

Spent most of the day getting a PPC kernel compiled on the Sphinx. In the process discovered that current RS/6000s running Open Firmware might be in a better position to run NetBSD or FreeBSD than I thought; certainly it's an encouraging sign.

In the evening to the LinuxSA meeting, in which Geoffrey Bennett demonstrated a Tivo, and got the entire talk on tape, something I should have done with Tridge's talks before. I had brought mine along, but we didn't have time to do anything with it.

Wednesday, 18 July 2001	Echunga
		Top of page
		previous day
		next day
		last day

Long conference call this morning. After that spend most of the time working on netbooting the RS/6000. In the process ran into a strange bug where dhcp on echunga was replying with an address associated with a different interface. Put in a PR about that, and ended up starting dhcpd on only the interface I wanted. After sorting that out and getting the right kernel, I was able to boot. Great. Now to jump in and rewrite the Linux kernel.

Thursday, 19 July 2001	Echunga
		Top of page
		previous day
		next day
		last day

Another long conference call today, and after that to the ADUUG meeting, where I found all the BSD people who didn't show up at the barbecue on Saturday, including Phil Kernick, whom I haven't seen for a long time. Interesting discussion; it seems things are happening in Adelaide.

Friday, 20 July 2001	Echunga
		Top of page
		previous day
		next day
		last day

And another week goes by. Got my dhcp problems sorted out, and started modifying the PPC Linux kernel for testing purposes, also pulled in the JFS sources. It's still strange to have to do so much manually, but I did get some help using Bitkeeper from Paulus, and I'm now less unhappy than I was.

Saturday, 21 July 2001	Echunga
		Top of page
		previous day
		next day
		last day

Long sleep-in this morning; I needed that.

After that did some catching up on mail; this is really keeping me busy a significant proportion of the time. After that Geoffrey Bennett came along and effectively installed the Tivo software for me. It's becoming clear that we've only scratched the surface of the possibilities so far.

Sunday, 22 July 2001	Echunga
		Top of page
		previous day
		next day
		last day

Spring is in the air. The daffodils are blooming, the horses are beginning to lose their winter coat, and the sun was shining most of the day. A perfect day to go out riding, so we did. Unfortunately Yvonne was riding Pebbles, who may also have sensed the coming spring; in any case, she was very hot, and Yvonne was completely unable to keep her in a walk. Not as short a ride as last year, but still not as long as we had planned.

Monday, 23 July 2001	Echunga
		Top of page
		previous day
		next day
		last day

Finally got some work done on the sphinx. Built a Linux kernel with JFS and booted it. Worked first time, modulo the fact that I didn't have any JFS file systems. I still need to get that differential SCSI adapter, but that's not going to happen this week, so put an AHA 2940 in the machine and watched Linux's idea of drive names change. Ah well, at least we're making progress.

Did my first recording with the Tivo. It didn't want to work, because the time was set wrong (too late). How do you set the date on a Tivo? I don't know, but it wasn't obvious. I took the easy way out: connected my laptop to the serial port and used date(1) to set it. The Tivo doesn't seem to have the concept of time zones, which will need to be fixed. Ah, isn't it nice to have a familiar user interface?

Tuesday, 24 July 2001	Echunga
		Top of page
		previous day
		next day
		last day

More playing around with Sphinx hardware today, during which I learnt more about Open Firmware than I really wanted to. In particular, I discovered that I can't just put any old PCI card in it and expect it to be recognized, at least not without some magic that I don't understand yet. As a result, I was not able to use the Adaptec host adaptor I put in yesterday, and ended up kludging together a second drive box to connect to the external SCSI connector. That worked without incident, and I was able to bring up JFS and run it on the second disk. This is almost too easy.

Wednesday, 25 July 2001	Echunga
		Top of page
		previous day
		next day
		last day

The electricians came in today to do a number of wiring tasks, mainly network and telephone. Finally we'll have a 100 Mb/s network throughout the house, and I discovered to my surprise that we'll be able to max out an 8 port switch in the HiFi cupboard: uplink to the office, three connections to the bedrooms, one to Yana's office, one to the Tivo, one to the wireless gateway, and a spare for a laptop in the lounge room.

Also did some work on getting EVMS working on sphinx. It went smoothly, and I was able to boot a kernel with EVMS support within hours. Unfortunately I didn't have the documentation, and I couldn't find it. Played around a little, but it didn't really do very much.

Thursday, 26 July 2001	Echunga	Images for 26 July 2001
		Top of page
		previous day
		next day
		last day

After all the work of the last couple of days, I found that I was left waiting for responses from other people, so I had a chance to catch up on background tasks. I probably need to spend more days like this.

More work on getting laptops working in the afternoon. It seems that there have been some undocumented changes in the way pcccardd works in -CURRENT: specifically, you need to start it with

# pccardd -I -i interrupt

where interrupt is the IRQ of the pcic0 (from dmesg). Warner's planning to fix that Real Soon Now, so if you read this in more than a week's time, check first.

Friday, 27 July 2001	Echunga
		Top of page
		previous day
		next day
		last day

More admin work today, also spent some time working on PCCARD stuff. Keeping a number of machines up to date can be a full-time occupation.

The electricians finally got the network connect to the old part of the house finished. Moved the wireless gateway machine (formerly fhss-gw.lemis.com, now air-gw.lemis.com) into the HiFi cupboard, and reconfigured it. With this change, the 10 Mb/s network is going away, which will involve a number of changes: in particular, flame.lemis.com has been up for nearly 300 days, and it was connected to the 10 Mb/s network. For some reason I can't umount the NFS file systems, so I have to leave it connected to the old IP address. I'll probably have to leave echunga with dual addresses until I next reboot flame.

Also connected the TiVo to the network. That was more difficult than I had expected; I obviously didn't watch Geoffrey as well as I should have last Saturday, and discovered that the machine didn't even have vi, only ed. Also the promise of NFS mounts didn't materialize. They failed silently, but didn't mount. Tried to copy some files across and discovered that rsync didn't work because there was no rsh on the machine; I need to find what I'm missing there.

Saturday, 28 July 2001	Echunga
		Top of page
		previous day
		next day
		last day

Another quiet day. I'm getting back to something closer to a 5 day week, which may be a good thing.

In the afternoon out riding again. We need to work more on floating horses, and they (Darah and La Tigre) were pretty hot when we finally got to Kuitpo Forest. We need to ride more.

Sunday, 29 July 2001	Echunga	Images for 29 July 2001
		Top of page
		previous day
		next day
		last day

Another quiet day today, but I did spend some time trying to work out what's wrong with the PCMCIA code at the moment. In the process, I brought zaphod and canberra up to date with today's -CURRENT. The build times surprised me. Here's the summary for zaphod:

--------------------------------------------------------------
>>> elf make world completed on Sun Jul 29 13:53:42 CST 2001
                        (started Sun Jul 29 11:26:31 CST 2001)
--------------------------------------------------------------
     8831.44 real      5553.48 user      3388.25 sys
     18500  maximum resident set size
       866  average shared memory size
       901  average unshared data size
       126  average unshared stack size
  12939275  page reclaims
      4250  page faults
         0  swaps
      9700  block input operations
    637142  block output operations
   4358011  messages sent
   4327105  messages received
         6  signals received
  10080943  voluntary context switches
    851078  involuntary context switches

And here's canberra:

>>> elf make world completed on Sun Jul 29 13:35:33 CST 2001
                        (started Sun Jul 29 12:21:08 CST 2001)
--------------------------------------------------------------
     4464.61 real      1937.89 user       730.10 sys
     15848  maximum resident set size
       991  average shared memory size
      1039  average unshared data size
       129  average unshared stack size
  12739181  page reclaims
     17072  page faults
         0  swaps
     67562  block input operations
      4624  block output operations
       135  messages sent
       135  messages received
         6  signals received
    795720  voluntary context switches
    588939  involuntary context switches

The interesting thing here is that zaphod is a two-processor Celeron, 466 MHz, while canberra is a laptop with a single 1000 MHz Pentium III processor. Looking at the frequencies, you'd expect them to be roughly similar in performance; obviously the difference in bus speed is really significant. And no, it's not because only one CPU was working: the sum of user and system time for zaphod is (barely) more than the elapsed time.

Out riding again. I don't know when we last went riding twice on the same weekend. Darah finally got enough movement to satisfy her, though I'm sure she could have gone a lot longer. La Tigre is still too young for long, tough rides, so I ended up going up and down with Darah while La Tigre followed in a walk.

Monday, 30 July 2001	Echunga
		Top of page
		previous day
		next day
		last day

More catching up today. Something is wrong with -CURRENT PCMCIA/CardBus code, and it's getting on my nerves. Spent a lot of time looking for that, but I didn't really know what I was looking for, so I didn't get very far.

Tuesday, 31 July 2001	Echunga
		Top of page
		previous day
		next day
		last day

Another quiet day. The electricians are still here. We thought it would take 2½ days.

PCMCIA is still in bad shape. Warner's doing what he can, but it's taking a long time. mojave is now in an unrecoverable state: I can't get any Ethernet card recognized, and I need that to build a new kernel. Grrr.