*scottstuff*

My wife’s Mac mini (G4, 1.42 GHz) has been crashing several times per day lately, and I’m almost out of options. I think I’ve found an issue with Lacie Firewire drives and PPC Mac minis, but I’ll let you be the judge of that.

The system in question is running OS X 10.4.8 with 1 GB of RAM. It boots off a Lacie mini Hard Drive & Hub. It’s been crashing with a kernel panic that looks about like this:

Mon Jan  8 22:06:24 2007
panic(cpu 0 caller 0x000E3DDC): vnode_writedone: numoutput < 0
Latest stack backtrace for cpu 0:
  Backtrace:
     0x00095138 0x00095650 0x00026898 0x000E3DDC 0x000D8A24 0x2D8F2230 0x2D8E4FF4 0x2DFA4C80 
     0x2DFA7A38 0x2DFA7B38 0x2D7F1FFC 0x2D7F222C 0x2DA7727C 0x2DA776B8 0x2D9D2C6C 0x2D9D26EC 
     0x2D594A44 0x2D576040 0x2D575940 0x2D5E1AD4 0x2D5D1450 0x002D0B94 0x002CFA5C 0x000A9314 
  Kernel loadable modules in backtrace (with dependencies):
     com.apple.iokit.IOSCSIReducedBlockCommandsDevice(1.4.9)@0x2dfa3000
        dependency: com.apple.iokit.IOStorageFamily(1.5)@0x2d8df000
        dependency: com.apple.iokit.IOSCSIArchitectureModelFamily(1.4.9)@0x2d7ea000
     com.apple.iokit.IOFireWireSerialBusProtocolTransport(1.4.4)@0x2da75000
        dependency: com.apple.iokit.IOSCSIArchitectureModelFamily(1.4.9)@0x2d7ea000
        dependency: com.apple.iokit.IOFireWireSBP2(1.7.5)@0x2d9cf000
        dependency: com.apple.iokit.IOFireWireFamily(2.2.5)@0x2d56c000
     com.apple.iokit.IOFireWireSBP2(1.7.5)@0x2d9cf000
        dependency: com.apple.iokit.IOFireWireFamily(2.2.5)@0x2d56c000
     com.apple.iokit.IOStorageFamily(1.5)@0x2d8df000
     com.apple.iokit.IOSCSIArchitectureModelFamily(1.4.9)@0x2d7ea000
     com.apple.driver.AppleFWOHCI(2.5.3)@0x2d5cb000
        dependency: com.apple.iokit.IOPCIFamily(1.7)@0x275f7000
        dependency: com.apple.iokit.IOFireWireFamily(2.2.5)@0x2d56c000
     com.apple.iokit.IOFireWireFamily(2.2.5)@0x2d56c000
Proceeding back via exception chain:
   Exception state (sv=0x2762FA00)
      PC=0x00000000; MSR=0x0000D030; DAR=0x00000000; DSISR=0x00000000; LR=0x00000000; R1=0x00000000; XCP=0x00000000 (Unknown)

Kernel version:
Darwin Kernel Version 8.8.0: Fri Sep  8 17:18:57 PDT 2006; root:xnu-792.12.6.obj~1/RELEASE_PPC

The second line is the key; it’s crashing due to vnode_writedone: numoutput < 0. Searching for that error gives me 2 other people who have had the same problem, both within the last month. Both other cases involve PPC Mac minis and Lacie FW hard drives. This can’t be that new of a problem, because I’ve been suffering since August when I was running 10.4.7.

I’ve tried a number of fixes with no success. I’ve dusted out the mini, reseated the RAM, performed disk and permission repairs, re-installed the 10.4.8 combo updater, monkeyed with power saving settings, and removed all non-essential hardware, all without any success.

I just opened a ticket with Lacie. If anyone else is seeing similar problems, let me know and I’ll add it to the Lacie ticket.

Like most of the people I know, I spent at least 60 hours per week staring at text on a monitor. These days, I’m mostly using 24” Dell LCDs–3 at work and 1 at home. They’re nice, but I keep finding myself wishing I had something bigger, with more pixels. Unfortunately, today’s batch of 30” LCDs doesn’t quite cut it–they all require a dual-link DVI connector, and of the 5 devices that I’ve plugged into my LCDs lately, only 1 can drive a 30” screen at full resolution. Half of them won’t even plug in at all. I want something that I can use at a high resolution for computer work, and then fall back to HD or even lower to play games or watch movies.

It looks like Westinghouse is going to announce something cool at CES: a 52” 3840x2160 LCD TV at CES. My math says that the screen will be roughly 45.25”x25.5” at about 84 DPI. That’d make an awesome full-desk monitor, although the DPI’s a bit low and the corners would be difficult to use for anything that you needed to focus clearly on. Since it’s a TV, I can only assume that it’ll have The Mother of All Video Scaling Chips in it, so my Xbox 360 will still work with it, even if it’s only at 720p. Assuming that the computer side is about like the IBM/Viewsonic/Iiyama 3840x2400 displays that were on the market for $7,000 a couple years ago, it’d do around 25 Hz with a single dual-link DVI and 50-60 Hz with 2 dual-link connectors. I could probably live with that. Sure, there are a few shortcomings, but I’m sure I’d find a way to cope if someone felt the need to mail me one.

While cleaning my office, I found a 6-year-old CD with scanned 35mm pictures from my 10-year high school class reunion. Amazingly enough, the CD was still perfectly readable, so I uploaded the pictures to Flickr. Hopefully someone will find these and enjoy them eventually.

When my wife and I moved into our house almost 7 years ago, I claimed one of the bedrooms as a home office. I crammed a desk, an 8’ folding table, and 2 big bookshelves into the room, filling it up with Linux systems, big CRT monitors, 2 or 3 printers, a couple scanners, and other big, heavy bits of hardware. I wanted to maximize my desk space, so I put the desk and table up against the walls and worked with my back to the door most of the time, never quite understanding why I didn’t enjoy working in there as much as I’d expected.

In 2002, I bought a PowerBook and discovered that I could work from any room in the house; I wasn’t tied down to big chunks of glass and 30 lb computers anymore. So I mostly stopped using my office for anything but storage. There are a couple servers in the corner, but most of the space is full of hardware that I don’t really need.

I finally realized a month or two ago that I actually want a quiet place to work, and that most of the problem with my home office is that the layout is designed to maximize the number of big CRTs that it can hold. That’s irrelevant now, and has been for years. What I really want is a quiet, comfortable room with one or two LCDs, a place to read, and some place to plug in my laptop.

So, I’ve started on a quest to fix my office. My long-term goal is to repaint, replace the nasty old carpet with something more livable, and maybe even pick up a couch to fit along one wall, but for now I’m starting by moving the desk around so my back is away from the door and getting rid of the big folding table. Then I’ll add a bit of extra lighting so it doesn’t feel like a cave anymore, and replace the two big bookshelves with someplace to cleanly store cables, random camera gear, CDs, and a couple dozen books.

I moved the desk around last night, then moved my 24” LCD onto it. Then I dug the speakers that I bought in 2001 and have never really been able to use out of the mess, plugged them into a nice little surround sound decoder, and wired my Xbox 360 and Wii into the LCD and decoder.

I’m only half-done with the “clean the room up” phase of the plan, but it’s already drastically better. Getting the desk away from the door made a huge difference. I don’t know why I didn’t do it years ago.

Well, the power’s back on, and it looks like we survived The Big Storm. It wasn’t really all that bad here; our power was down for a bit under 48 hours, but we still had water and gas, and there was a supermarket with power under a mile away, so it hasn’t been too hard to cope. Our house lost a screen door, and there were a handful of small branches in the yard, but that’s pretty minimal. I’ve heard from a few co-workers with similar stories, but half of my team is still missing, hopefully due to continuing power outages.

The scariest part for me was the complete lack of cell coverage that we had for about 12 hours–I wasn’t able to make or receive calls from home until late in the afternoon yesterday. No network access, no cell phone, no power–it was as disconnected as I’ve been in years.

What is it with Seattle weather this year? Until a month ago, my worst commute ever was 1:30, tied a couple times in the 8 years I’ve been living in the Seattle area. Now I’ve broken that twice in a month. First there was last month’s freak snowstorm that gave me a 3:30 commute, and tonight it took me 1:40 to get home, just because of rain. I mean, I can understand Seattle being unable to handle snow, but rain? What the heck? Admittedly, it was a lot of rain–our office’s underground parking garage was filling up, and the parking lot was about to overflow into the lobby when I left at 5:15, but still–we’re supposed to be able to cope with water around here.

Sheesh. I’m tempted to call in sick for the rest of the winter.

I was going to post something here saying that I’ve now been at Google for one year, but I received some mail today that kind of preempts that. I just learned that Ryan VanderWall, one of our young helpdesk guys at work, was diagnosed with advanced, metastasized liver cancer a couple weeks ago. He’s spent the last couple weeks in the hospital undergoing radiation therapy.

Go check out Ryan’s blog, and consider donating if you can.

Update: December 23, 2006.

One of the things that I like about working for Google is the commute–in perfect traffic, I have an 11 minute drive. Unfortunately, tonight’s commute was far from perfect–instead of taking 11 minutes, it took three and a half hours. That’s an average of under 2.5 MPH.

Seattle just can’t cope with snow.

Yesterday, Slashdot pointed out that “The Web” just turned 16. The first web page at CERN went up in November of 1990, about a month after I first used the Internet.

I heard rumblings about the web off and on for the next few years, but it wasn’t really interesting until NCSA Mosaic was first released in 1993. I remember reading the release announcement for Mosaic 1.0 and being amazed–wow, a graphical viewer that could handle fetching and displaying multiple file formats over multiple network protocols. How cool. I uploaded one of the first (the first?) Mosaic binaries for Linux a month or so later.

I still remember days where I started browsing at the NCSA’s “what’s new” page and finished when I’d read every single page that I could discover on the Internet.

I upgraded my home Asterisk server from 1.2.x to 1.4.0-beta3 today. Unfortunately, I’d ignored a bunch of deprecation warnings when I upgraded from Asterisk 1.0 to 1.2, and it looks like 1.4 removed almost everything that was deprecated in 1.2. So I spent most of an hour replacing DBget(var=some/key) with Set(var=DB(some/key)). I guess that’s an improvement, although neither form is exactly elegant.

The single biggest improvement that I’ve seen so far is Jabber support. Now, every time the phone rings, my Gmail account gets an instant message with the caller ID information.

My Asterisk config is now around 2 years old and getting kind of crufty. Sooner or later I’m going to need to rewrite it in a better language; either AEL or one of the various Ruby plugins for Asterisk. AEL looks like a big improvement over Asterisk’s traditional config language, but it’s still not a real programming language.

I just finished upgrading the hardware that runs this blog. Everything tests okay, and the logs look clean, but major upgrades always make me nervous.

As I’ve mentioned before, this site was running on a 700 MHz Athlon for years. That worked fine when there wasn’t a whole lot of traffic, but I’ve kept adding new services and sites over the years, and it all adds up eventually. The poor little Athlon has been running for 42 days since it’s last reboot, and it’s averaged nearly 50% CPU utilization the entire time.

Unfortunately, the very thing that makes it slow also makes it hard to upgrade–the poor little system runs a dozen websites, acts as my home router, runs Asterisk, and handles email for at least three domains. It’s been up and running since May of 2000, so there are probably minor services that I’ve completely forgotten about over the years. Frankly, if someone handed me this as a project at work and said “fix it,” I’d probably run screaming.

I have a three-phase plan for fixing things:

1. Outsource as much as possible. For instance, I’ve stopped using my local IMAP mail server and switched to Google Apps for Your Domain. I still have SMTP and mailing lists running locally, but they’re a lot easier to maintain.
2. Move each service on the old system onto its own Xen virtual machine on a new Athlon X2 3800+ system. This should be a bit easier to maintain then just lumping everything onto one single Linux system, and it has 6x the CPU power and 4x the RAM of the old system.
3. Once everything is migrated, all that will be left on the old system is routing and my firewall. I’ll migrate that onto something less power-hungry; I’m not sure what yet.

At of about a half hour ago, this blog is now running on the new system. It should be substantially faster then before, especially since it currently has the entire X2 to itself. I’m planning on upgrading it to the Typo trunk in a few days, but I don’t like making too many upgrades all at once.

Update: Well that was fun. I’ve had the new box running quietly under a desk for over a month without problems. A couple hours after I move traffic onto it, one of the hard drives failed, killing the system. I’m not quite sure how that happened–it’s running RAID 1. Even better, Xen’s network configuration script only works correctly for me on every second boot. Some days I just love computers.

It seems to be kind of in vogue to point people to Steve Yegge’s blog these days–I’ve seen 2 or 3 people that I read link to him already this week. And I try to avoid me-to-isms. But you probably ought to go read his take on Agile programming. I don’t completely agree with everything that he says, but it’s a fascinating read. Even if he does overstate things slightly–he said that if you ask “[Google’s office] movers will show up the next day to put you in your new office with your new team.” It’s taken Google’s movers almost a week to throw me out of my office and give it to Steve’s new team. Inaccuracy!

I bit the bullet this morning and upgraded my E61 to the new firmware release that’s been floating around. I was planning on putting it off until I got back home from California, but there was kind of an E61 upgrade party going on in the office (E61s are kind of popular at Google), and the guy before me didn’t have any problems, so I went ahead and bit the bullet. It took 20 minutes or so, but everything seems to work fine.

The process was simple:

1. Run Tools/Memory and select ‘Backup phone mem.’ to back up everything to your flash card.
2. Run the Windows-based Nokia firmware upgrader.
3. Run Tools/Memory and restore your backup.

The only real change that I’ve seen is that my phone’s browser doesn’t lock up with a specific HTTPS site that I use anymore. Other then that, everything seems the same, including the inane behavior with IMAP IDLE servers. Unlike firmware 1.x, though, it doesn’t seem to hang the messaging app when it loses track of the IMAP session, so SMS still seems to work.

I’m in Mountain View this week for a few meetings, and my social calendar is pretty empty. Is anything happening this week?

A number of people have pointed out that Typo’s Flickr API key has expired. The situation is actually slightly more complex, but the fix is dead simple. See changeset 1256 for the fix. If this works for everyone, then I’ll roll 4.0.4 over the weekend.