|
| Greg's diary |
| November 2000 |
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Wednesday, 1 November 2000 | Echunga | |
| Top of page | ||
| next day | ||
| last day |
From time to time it looks as if I'm getting nothing done. More mail:
On 31 October 2000 you received 1327 mail messages.
Finally I'm getting the Malaysian slides working. For future reference, the description of how to import PostScript files into groff is described in grops(1), and it's complicated enough that it's probably better to use the PSPIC macro. Unfortunately, I didn't find a daemon image with adequate resolution, so it looks as if we're not going to make it.
| Thursday, 2 November 2000 | Echunga –> Melbourne | Images for 2 November 2000 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
More slide writing, somewhat disturbed by Yvonne arriving with the news that Darah had fallen on Duffield road and scraped both front knees:
Looked a nasty sight (the photo shows it after treatment), but the vet came and said he didn't think she'd have any lasting problems for it, but at least I now have a reason to ride Miss Teak.
sigh The spammers and the “virus” makers have teamed up. This evening I got about 20 messages from a Peter Wagner in Germany with a .vbs file obviously containing some malicious program. I called up the company and confirmed that Peter Wagner does, indeed, exist. “OK, I'll tell him” said the sysadmin. Sheesh.
| Friday, 3 November 2000 | Melbourne | Images for 3 November 2000 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Worse than yesterday's “virus” were the automatic answers. By the time I got up, I had had about 300 of them:
On 2 November 2000 you received 1643 mail messages.
At least they were easy enough to remove. I pity users of Microsoft MUAs who have to delete every single one of them manually.
In the afternoon to Melbourne and to the Rydges hotel in Swanston St., Carlton, on the same block where I used to live as a kid. Wandered around Lincoln Park, Grattan St. and the South-East corner of the University, and took a couple of photos of our old house:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Ah, the nostalgia.
After that, called Luke Mewburn, who was carrying on after the security symposium somewhere down south of the Yarra, and Benno Rice, who was just down the road in Flemington Road. Benno joined me at the hotel, and we then went down to the pub, where, true to form, just about the entire AUUG executive committee were also present. Had dinner at a forgettable Indian restaurant and decided (with Luke, who's on the NetBSD Core Group) that we needed to forge better ties between the BSDs. Michael Paddon (OpenBSD) was less involved, but in favour.
| Saturday, 4 November 2000 | Melbourne | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Up bright and early this morning and to Lygon St. for breakfast at the inappropriately named Cafe [sic] Notturno. Then to the appointed place for the exec meeting, arriving too early as usual, and found myself hacking on my laptop outside on the street before a locked door until Con turned up on the dot of 10 am and let us in.
I'm getting used to these exec meetings now. This time we had even fewer Davids present than last time, and Malcolm ran the meeting. David Purdue called in on the phone, but David Newall is in KL, and I'll probably be the first to see him in a couple of days time. As yesterday evening and at other times along the way, we noticed the hidden agenda of the 08 area code: the FreeBSD core team now has 3 members of 9 from the 08 area code (and also the security officer), and the AUUG Exec committee has 5 of 9.
With Peter Gray's help, Malcolm got through the agenda on time for the first time in living memory, and we went off for a drink at the Oxford, on Swanston St. opposite the City Baths. At 6 pm sharp we got thrown out of the pub; I thought we got rid of that law decades ago. The East coasters left us, and Luigi, Malcolm, Sarah and myself wandered up to Lygon St. to have another forgettable dinner, while Luigi let it be known that, like all Italians, he comes from a family of restaurateurs. On further for another drink, and found ourselves back at the hotel round midnight trying (in vain) to debug Microsoft Outlook (undebuggable) and Wireless Ethernet (no two cards with the same standard).
| Sunday, 5 November 2000 | Melbourne –> Singapore | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Despite the late night, up round 9 am to have breakfast with Sarah, who didn't respond to her phone. Down to a Greek restaurant on the corner of Latrobe and Russell Streets, and while having breakfast, Sarah walked past. It eventuated that her phone wasn't loud enough when hidden in her handbag...
After breakfast up to the old museum to see if Phar Lap was still there. He wasn't. Neither was the museum: it had turned into an art gallery. Looked round for a while, then left Sarah and off to the airport and spent the day catching up with my mail in a very cramped Qantas club.
Off to Singapore, where I had to wait far too long for “priority” baggage: they had forgotten a container on the plane, allegedly the fault of the people in Sydney. To the hotel (Méridien Changi), where I shall not go again.
| Monday, 6 November 2000 | Singapore –> Kuala Lumpur | Images for 6 November 2000 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
On to KL this morning. A total of 400 km, but I had to leave the hotel at 9 am, and didn't get to my hotel in KL until 1:30 pm, thanks at least in part to the baggage handling at KL airport.
Before leaving Singapore, had a chance to try out their 802.11b network in the airport departure lounges. Good news: it works. Less good news: nobody can tell you how. First, there are two separate ways to get wireless Internet connection. One is infrared, which seems to be preferred way, and it seems to be Palm Pilot oriented. The other is real 802.11b, but most people don't have the cards. They're available for free loan (or for purchase), but not many people seem to use them. There are signs around advertising the feature, but no technical details, and the people at the information stands will tell you that you have to use the infrared method.
For further reference: it's standard 802.11b in managed mode, no encryption (so make sure you're using encrypted protocols!). I had some trouble because my initialization routines kept turning on encryption, and by the time I discovered that I had already called the service people, who told me to set an SSID (I think; they didn't know the term) of ANY. I'm not sure whether this is a Microsoft thing meaning “empty”, or whether the SSID needs to be ANY. Anyway, it got me connected:
=== grog@sydney (/dev/ttyp6) ~ 1 -> wicontrol -i wi0NIC serial number: [ 00UT28336821 ] Station name: [ FreeBSD WaveLAN/IEEE node ] SSID for IBSS creation: [ ANY ] Current netname (SSID): [ SIA Lounge Wireless ] Desired netname (SSID): [ ] Current BSSID: [ 00:60:1d:f2:2a:17 ] Channel list: [ 2047 ] IBSS channel: [ 3 ] Current channel: [ 10 ] Comms quality/signal/noise: [ 6 60 54 ] Promiscuous mode: [ Off ] Port type (1=BSS, 3=ad-hoc): [ 1 ] MAC address: [ 00:02:2d:04:09:3a ] TX rate (selection): [ 3 ] TX rate (actual speed): [ 2 ] RTS/CTS handshake threshold: [ 2347 ] Create IBSS: [ Off ] Access point density: [ 1 ] Power Mgmt (1=on, 0=off): [ 0 ] Max sleep time: [ 100 ] WEP encryption: [ Off ] TX encryption key: [ 1 ] Encryption keys: [ ][ ][ ][ ]=== <font color=red>grog@sydney</font> (/dev/ttyp6) ~ 2 -> ifconfig wi0wi0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500 inet6 fe80::202:2dff:fe04:93a%wi0 prefixlen 64 scopeid 0x9 inet 196.168.1.68 netmask 0xffffff00 broadcast 255.255.255.255 ether 00:02:2d:04:09:3a
They appear to have blocked traceroute:
=== grog@sydney (/dev/ttyp6) ~ 3 -> traceroute freebie.lemis.com
traceroute to freebie.lemis.com (192.109.197.137), 30 hops max, 40 byte packets
1 196.168.1.1 (196.168.1.1) 2.781 ms 6.651 ms 8.375 ms
2 *^C
but name services and ping work, and so does ssh. As you might expect, you set up connections using dhcp.
In the evening wandered out trying to work out where I was. Finally found myself on Ipoh Road, and walked up into Sentul, and then through Jalan Kolam Ayer, now respelt Jalan Kolam Air, thinking of when my friends all lived there. I barely recognized some of the newest houses built when I was in KL, now just about falling apart. Most of the Tamils are gone, too, replaced by Malays. Melbourne on Friday was easy to recognize; here it was almost impossible. Returned to the hotel via a new walkway down the side of the Batu River (we used to have to wade through it to get anywhere) and past Maxwell College back to the hotel. I've now decided it must be where there used to be a padang along the side of Jalan Maxwell when I was a kid, just past where the Batu river (left) flows into the Gombak river:
|
|||||||||||||||||||||||
| Tuesday, 7 November 2000 | Kuala Lumpur | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Nothing much to do today; the conference and my work don't start until tomorrow. Listened to Richard Sharpe's talk on Ethereal, very interesting, but probably over the head of most people. I certainly think he was optimistic in his attempts to get people to build dissectors for it. I suppose I should probably take a look at the functionality, though.
After the event did a bit of testing my laptop with the projector (worked fine, sort of) and the SGI box I am to use for installation on Thursday. Good thing I did: for the first time in a long time I found a keyboard which didn't work with FreeBSD. We had to replace it with a Compaq keyboard.
In the evening into town with David Newall, who is very taken with the place. I'm less enthusiastic. Took the LRT to a station whose name I forgot, but which is slap bang above the middle of the Mountbatten Road (oops, what's that called nowadays) bridge. There's even an old left-over beggar still left on the bridge, to David's concern, but there's as good as nothing left of what used to be KL's main shopping centre. Wandered down past the real centre of town, where the Gombak river (behind, and muddier) flows into the Klang river, before going back up to Ampang Road and Bilals for a forgettable meal. That place has gone badly downhill.
| Wednesday, 8 November 2000 | Kuala Lumpur | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
And into another conference. This one was interesting for a couple of reasons: there was a very good turnout (I'd guess about 300 people), and during the keynote speech (“first keynote” in the terminology of the organizer), Gavin Sherry asked how many people could program in C. Three people raised their hands: Richard Sharpe, myself, and one other person. As they say themselves, it's very commercially oriented.
In the afternoon, did my talk on the BSDs. At least people were interested enough to ask questions, which is generally not the case.
In the evening, a “speakers' reception”. I didn't meet many people I didn't know, but Robert Hart of RedHat was there. It appears he's back in Australia again.
| Thursday, 9 November 2000 | Kuala Lumpur –> Singapore | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Today was my installation presentation (“how to get FreeBSD up and running in 45 minutes while maintaining a constant patter”). It's a good thing I tried out the installation on Tuesday: it worked fine then, and I ran into troubles with X today, probably because I didn't specify any explicit resolutions, and the whole thing came up in something that the projector didn't understand. Still, managed to get it up and running KDE within the 45 minutes, and we had enough audience to swallow up the 200 CD-ROM sets that BSDi donated. It looks like there is potential for more BSD advocacy in Malaysia.
In the evening back to Singapore.
| Friday, 10 November 2000 | Singapore | Images for 10 November 2000 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
This morning to visit eLinux. Spent most of the morning overloading myself with another 3000 downloaded mail messages, some of which required semi-urgent attention, and then spent the afternoon talking to the techies, who had a number of questions.
In the evening to the LUGS meeting, where Michael Boman, one of the eLinux people, presented a paper on security. I took a couple of photos, mainly to demonstrate the technique of mounting the camera flash memory as a disk and using standard UNIX commands to manipulate the images:
=== root@sydney (/dev/ttyp0) ~ 23 -> mount -t msdos /dev/ad6s1 /camera=== root@sydney (/dev/ttyp0) ~ 24 -> dfFilesystem 1048576-blocks Used Avail Capacity Mounted on /dev/ad0s2a 987 848 59 93% / /dev/ad0s4e 11372 9394 1067 90% /home /dev/ad0s1c 2992 1805 1187 60% /C: procfs 0 0 0 100% /proc /dev/ad6s1 45 33 11 74% /camera=== root@sydney (/dev/ttyp0) ~ 25 -> ls /camera/dcim/100nikon/dscn0067.jpg dscn0076.jpg dscn0085.jpg dscn0094.jpg dscn0103.jpg dscn0112.jpg dscn0068.jpg dscn0077.jpg dscn0086.jpg dscn0095.jpg dscn0104.jpg dscn0113.jpg dscn0069.jpg dscn0078.jpg dscn0087.jpg dscn0096.jpg dscn0105.jpg dscn0114.jpg dscn0070.jpg dscn0079.jpg dscn0088.jpg dscn0097.jpg dscn0106.jpg dscn0115.jpg dscn0071.jpg dscn0080.jpg dscn0089.jpg dscn0098.jpg dscn0107.jpg info.txt dscn0072.jpg dscn0081.jpg dscn0090.jpg dscn0099.jpg dscn0108.jpg dscn0073.jpg dscn0082.jpg dscn0091.jpg dscn0100.jpg dscn0109.jpg dscn0074.jpg dscn0083.jpg dscn0092.jpg dscn0101.jpg dscn0110.jpg dscn0075.jpg dscn0084.jpg dscn0093.jpg dscn0102.jpg dscn0111.jpg=== root@sydney (/dev/ttyp0) ~ 26 -> mv /camera/dcim/100nikon/d* /var/tmp=== root@sydney (/dev/ttyp0) ~ 27 -> ls /camera/dcim/100nikon/info.txt=== root@sydney (/dev/ttyp0) ~ 28 -> dfFilesystem 1048576-blocks Used Avail Capacity Mounted on /dev/ad0s2a 987 882 25 97% / /dev/ad0s4e 11372 9394 1067 90% /home /dev/ad0s1c 2992 1805 1187 60% /C: procfs 0 0 0 100% /proc /dev/ad6s1 45 0 45 0% /camera=== root@sydney (/dev/ttyp0) ~ 29 -> umount /camera
We also discovered that it's possible to thoroughly confuse Linux systems by opening many (about 40) telnet connections to localhost. I tried that on my FreeBSD laptop, with an infinite number of connections, and noted that, apart from resource problems, nothing went wrong. We hit the process limit with about 350 telnets, which I then had to kill. The people present found that impressive.
After that to dinner, and finally found some good satay in Singapore. I'm quite impressed with the technical people here in Singapore; they seem pretty clever and very motivated.
| Saturday, 11 November 2000 | Singapore –> | Images for 11 November 2000 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Up late this morning. I must really catch up on my sleep, but today wasn't the day. After yesterday's exorbitant breakfast ($34), decided not to have breakfast in the hotel, and went to the food court in Raffles City instead, only to discover that, at 10:30 am, it still hadn't opened. Down North Bridge Road to the Funan centre, where I had a forgettable Penang Laksa ($3.10) and then went to look for the third edition of Evi Nemeth and Co.'s System Administration, but they were sold out.
Back to the hotel and checked out, then to Sim Lim Square looking for accessories for my camera. After a lot of searching, became convinced that there are no rechargeable batteries for it (pseudo 6V lithium) available in Singapore. During the search took my camera out of my bag to take a photo and discovered that the thing had somehow turned itself on. Since the lens cap was on, it just displayed an error message—and didn't time out! The battery compartment was almost too hot to touch, and of course the battery was dead. Grrr. Another firmware bug: if the lens cap had been off, it would have timed out after 30 seconds.
Bought a 48 MB flash card at a price which was good here and bad in the USA, then on to Mohamed Mustafa Samsuddin and bought some watches for Yvonne and Yana. I'm coming to the conclusion that:
Then to the airport, where the 802.11b support was down, so no further testing. Back home.
| Sunday, 12 November 2000 | –> Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Say what you will, I prefer long overnight flights to short ones. Flying from LA to Sydney (14 hours) gives you time to get 8 hours sleep. This short 7 hour flight gives you a maximum of 4 hours sleep, and in this case I had not one but two screaming babies in the immediate vicinity, so I didn't get much sleep.
Back home and took a look at our new arrival, La Tigre before going to bed until lunch time.
In the afternoon, rode Miss Teak for the first time in months. She was surprisingly cooperative. With a bit of work, I'm sure she'll make a good horse.
Took it easy in the afternoon. Round evening, Greg Edmonds came around with the most worn out floppy I have seen in a long time. I can't see much hope.
| Monday, 13 November 2000 | Echunga | Images for 13 November 2000 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
And back into the daily grind, this time with quite a bit to do. Real Work on the one hand, preparing for the Australian Open Source Symposium 2 on the other. As if that weren't enough, while I was fighting my remaining 2,500 mail messages the modem connection dropped and couldn't reconnect. The phone line was dead. Fine, for once they didn't all die together, so I could use the spare line the way I intended.
In the afternoon, the spare line died too. So did all the rest. Funny, it wasn't raining. Rang up Telstra, who didn't know about any problems, and then drove up the road to the old familiar area. I don't know what it is about that place, but this time there were three linesmen sitting there splicing a 2 metre length of cable into the line. Somebody had been there with a backhoe. Well, at least the lines weren't down long. This is the only case I know where Telstra has honoured the legislated Customer Service Guarantee.
| Tuesday, 14 November 2000 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
More Real Work today. One of these days I must stop biting off more than I can chew.
In the afternoon, another power failure. They're really overdoing this stuff lately.
| Wednesday, 15 November 2000 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Spent most of the day playing with the Yoyodyne box and learning things about Linux installations that I long knew about for FreeBSD. It's funny being a beginner again.
Arranging the AOSS turns out to be more of a problem than I thought. We should really get our act together better next time.
| Thursday, 16 November 2000 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
During the morning, my main gateway machine echunga.lemis.com paniced again. This has been a regular business (about once every 3 or 4 days) since I installed a buggy -CURRENT kernel a couple of months ago. I don't want to upgrade to the new -CURRENT until I've written the lightweight interrupt threads (which at least gives me an incentive :-), so decided to install 4-STABLE instead. I had a new disk to connect to it, so I left the old one as secondary master so that I didn't need backups, and I could copy files across as I needed them. Wouldn't take long for an expert like me, right?
Well, wrong. First I found I didn't have a CD-ROM drive on echunga, and that NFS and ftp installs didn't work too well, at least partially because the other machines on the network were hung on NFS requests to echunga. Finally connected a CD-ROM to the machine and installed like that. But that shouldn't be necessary.
Things weren't fixed there. A number of things had changed on the way, including the location of sendmail.cw, so the first mail messages which arrived were rejected. Dozens of other minor things happened, but at least this time I was able to keep notes, which should come in handy for my planned tutorial “FreeBSD: The instant server”.
| Friday, 17 November 2000 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Off to the monthly ADUUG (Adelaide UNIX User Group) lunch today. We had 10 people this time, one of the larger gatherings, but the place (Manna, in Wakefield St.) was very noisy, and we could hardly hear ourselves talk. Got to talking about flash memory, and with a little work we were able to take a photo in my digital camera and read it in Alastair Dick's Psion notebook. He didn't have appropriate software to read 2048x1536 images, but it worked fine at 640x480, modulo the monochrome display.
Later in the afternoon, talking about compact flash on IRC. For the fun of it, I tried to create a ufs file system on CF in sydney.lemis.com. Well, it worked, but not before I mistyped the disk unit number and discovered that I could newfs a mounted file system if I used an overlapping partition with a different name. sigh another new install... At least I had backups.
| Saturday, 18 November 2000 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
All this reinstallation of systems has somehow got on my nerves, so for a change of something to do I finally got round to installing NetBSD 1.4.2 for my diskless Sun 3/60s. That turned out to be a little more work than I had planned, though not too much. I obviously need to learn more about Sun 3s, but for now at least one of them works. The other appears to have the Ethernet port set to AUI, and though it can send, it doesn't receive. One day I'll find out how to change that.
In the evening, Magia had her foal, a colt, over a week early. He doesn't look like there's anything wrong with him, though.
| Sunday, 19 November 2000 | Echunga | Images for 19 November 2000 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
More playing around with the Sun 3s today, including hardware chop and changes. The big discovery was how slow they are. Back 12 years ago when they were new, they were considered to be fast machines, and in fact they have the same spec (and more memory) than the Tandem LXN I still have. I once benchmarked that LXN as having the fastest CPU of any current Tandem machine; that must have been in late 1989 or early 1990. Now I find that it took 3 hours to build bash, while wantadilla does it in exactly one minute. Some of that is probably because of the low memory—only 8 MB in the machine I used—but it's still a factor of 180 to 1.
| Monday, 20 November 2000 | Echunga | Images for 20 November 2000 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Not much to report today. More testing of the Yoyodyne box, interesting but confidential.
EFnet seems to have died a death. After three days, I still can't find a server which stays up. Telstra have shut theirs down, and the script kiddies seem to be doing their best to kill the rest. I wonder when people will take effective steps to either improve the protocol or bring the script kiddies to justice.
| Tuesday, 21 November 2000 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
The big news today was that somebody damaged an international phone cable 100 km off the East Coast of Singapore. It was on the news in the morning, along with the information that services had been restored. Nonsense: I had up to 80% packet loss across the Pacific, www.auug.org.au disappeared off the net (it's in Canberra, but the packets were routed to Perth before being dropped, presumably into the Indian Ocean), and my connection to front.linuxcare.com.au, also in Canberra, went via San Francisco and Los Angeles. The new, highly vaunted Southern Cross Network doesn't seem to have helped much. It could be up to a week before the cable is repaired: the damage is in Indonesian waters, and so far they don't even have authority to send a boat there.
In the evening to the Linuxsa monthly meeting, and delivered my speech about the AUUG. We gave away two memberships; we'll see how many people turn up on Saturday.
| Wednesday, 22 November 2000 | Echunga | Images for 22 November 2000 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
More work on the Yoyodyne box. Also, the pace seems to be stepping up to the AOSS symposium. It looks like we'll be paying tribute to the original etymology of “symposium”: from Liddell & Scott's Greek—English Lexicon, 7th Edition:
Sympósion, tó: a drinking party, symposium Theogn., Hdt., etc. II the room in which such parties were given, drinking-room, Luc.
This is really the entire entry. The Oxford English Dictionary adds that the topic of discussion might be intellectual. It's sad how such customs have fallen into disuse.
| Thursday, 23 November 2000 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Spent most of the day catching up on documentation and arranging the AOSS. It's amazing how you can spend the whole day doing nothing of interest.
| Friday, 24 November 2000 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Spent most of today documenting things and preparing for the AOSS tomorrow. We didn't get the final programme finished until the evening.
Luigi Cantoni arrived in the afternoon, and I went down to Glenelg to pick him up. Yvonne put on some good food in his honour.
| Saturday, 25 November 2000 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Early into town this morning to the AOSS. We need to do something about signing and directions; we had to call Sarah to even be able to get on the campus, and she had enough difficulty describing it to us that she had to come out and find us.
The symposium got off to a good start. The facilities weren't too bad, with the exception of Yet Another 800x600 projector, with which a number of people had problems. I was very pleased with the quality of the talks, especially considering that 6 of the 9 papers were delivered by people personally involved in the projects, and 7 papers were by local people.
After the symposium, without even enough time for a beer (what kind of symposium is that?), off to the Festival Theatre to meet Yvonne and Yana for a concert (Appalachian Spring, Korngold Violin Concerto and the New World Symphony). Possibly not quite a world-class performance, but nevertheless very good. We should go there more often.
| Sunday, 26 November 2000 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Up early to prepare for the barbecue that we had planned for the AOSS speakers. We got off to a slow start, and Sarah had difficulty finding her way. Went out to the street to look for her, but all I found was Monty, dead, having been hit by a car. Sadness. Luigi, bless him, braved the heat (34°) and buried him for us. Now we only have one cat left, and I think we'll leave it like that. This is obviously just too dangerous a place to keep cats.
The barbecue was about as enjoyable as it could be under the circumstances. It's nice to have out-of-town geeks around.
| Monday, 27 November 2000 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Finally I have time to catch up on the stuff I haven't had time for in the last few months. Spent a lot of time catching up on the more urgent mail (over 2000 messages), and also working on the Linux port of rawio. It's interesting how much there is to learn about differences in as UNIX-like an operating system as Linux.
In the evening Yet Another power failure, about 90 minutes. I'm getting a bit fed up with this. During the night we had a further one, but it didn't wake me, and it wasn't long enough for the UPSs to discharge.
| Tuesday, 28 November 2000 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Teleconference with San Francisco this morning. While looking for the number, I came across another message with “conference” in the subject: yesterday was the deadline for summaries of papers for next year's Freenix track at USENIX. It's a good thing that it was still yesterday in California, so I made the deadline.
Finally decided that the installation I had on zaphod was not worth saving, and decided to create a snapshot CD of FreeBSD 5-CURRENT instead. The buildworld died in Vinum; turns out it wasn't my fault directly, somebody had committed some changes without doing a buildworld first. At least I was in a position to fix the breakage.
Got some negative feedback about the AOSS. I think we need to do something about setting expectations.
| Wednesday, 29 November 2000 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Got in this morning and discovered that my make release had failed. I really needed zaphod as a test machine, so off to completely reinstall stuff there. In the meantime tested the Yoyodyne box with Bonnie, which gave less than spectacular results.
Installing FreeBSD is getting easier all the time. I suppose practice doesn't make it any more difficult, but the installation itself is also getting better. But copying data around still takes a long time.
| Thursday, 30 November 2000 | Echunga | Images for 30 November 2000 |
| Top of page | ||
| previous day |
Somehow I'm doing too many things at once, and end up thrashing between one and the other. Tried out bonnie under FreeBSD on the yoyodyne box and got surprisingly much better results than with Linux: in sequential I/O, it was more than twice as fast on the same hardware, and even on random seeks it was 20% faster. Linux gave:
-------Sequential Output-------- ---Sequential Input-- --Random--
-Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
Machine MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU
ping 2000 11123 98.4 21214 14.3 11055 15.0 12049 97.3 28741 9.5 108.5 0.7
ping 2000 11127 98.3 21114 14.2 11012 15.1 12009 96.9 28782 9.4 108.7 0.6
ping 2000 11130 98.3 21252 14.2 10916 14.9 12046 97.2 28806 9.0 106.9 0.5
By contrast, FreeBSD gave:
ping 2000 20165 58.5 20758 17.9 10721 12.9 21716 93.7 29355 21.4 121.0 0.9 ping 2000 20637 60.0 20770 17.8 10758 13.0 22431 96.9 28567 21.5 121.2 1.3
I'm puzzled by these figures. While I'm in no doubt that they're “correct” (in other words, done in ostensibly the same circumstances), there may be reasons other than the operating system to explain them.
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 |