Category: computers

Best Buy Woes

February 24, 2005

So, for the last while the MENU button on my iPod has been intermittantly non-responsive. I can’t help but think that this is a systemic problem with the third generation iPods, partially leading Apple to switch to mechanical buttons (via a rocker wheel) on the latest iPods.

Regardless, I have a four-year warranty for my iPod through Best Buy. I took the device in today to see if they’d swap it, and while the “tech” had the problem happen to him, he noted that it “wasn’t happening anymore” after the first problem with it. Based on that, he determined the iPod wasn’t having problems anymore. He offered to send it out to the repair center for 3-4 weeks to see what they could do… Well… That’d be nice, not having an iPod for 3-4 weeks.

I had been hoping that the “tech”, after having it demonstrated to him that the button occasionally fails to work, would realize that the device really does have issues and get it replaced with a new one. Best Buy’s support isn’t going to be able to repair a capacitive switch that occasionally fails after the device has been sitting for a few minutes.

…

February 20, 2005

Hmm, this is kind of disappointing. I acquired the Broadcom BC5805 SSL accelerator and I’ve got it installed in a test machine. While it seems to accelerate rsa and dsa signing under OpenSSL, it doesn’t actually accelerate des and md5 stuff like I had expected.

So I did some digging, and it seems that OpenSSL is actually looking for /dev/ubscrypt and /dev/ubskey so that it can hook into the card itself and accelerate things.

I find this disappointing, because there is not a driver available for providing such things under FreeBSD, only under Linux.

computers electronics moved from livejournal

!!!

February 9, 2005

Hey all you Windows-using people… Be sure you run Windows Update and get all your updates and such. Some of the recently released bulletins from Microsoft (specifically MS05-011 and MS05-012) point to some pretty nasty things that could be lingering on the horizon. I’m sure you don’t want your machine randomly infected just because it’s sitting there connected to the interweb.

(Gee, can you guess what I’m working on today?)

computers moved from livejournal

Weird Referrers…

February 8, 2005

Any of you out there who run webservers… Have you ever seen anything like this? It appears to be grabbing lots and lots and lots of images. In tailing the nuxx.net log I haven’t seen it grabbing anything else yet.

207.155.199.163 – – [08/Feb/2005:23:28:38 -0500] “GET /albums/wallpaper/collectiveextension_1024.sized.jpg HTTP/1.0” 200 32768 “-” “Mozilla/4.0 (compatible; MSIE 4.0; Windows NT; ……/1.0 )”

It’s a really weird referrer, and I’m getting it from a bunch of IPs all at once. I’d normally say it’s someone indexing my site, but some of the IPs it’s coming from are:

207.155.199.163 (UUNET)
12.17.130.27 (AT&T)
65.164.129.91 (Microsoft Sprintlink)
208.252.91.3 (UUNET)

None of those reverse, and the addresses aren’t allocated to any customers, they are just held by the big ISPs.

Seems kinda weird to me…

UPDATE: Upon more digging, I’ve found this:

207.155.199.163 – Concentric Dialup, run by XO.
12.17.130.27 – Traceroute seems to just stop, like it’s part of some absolutely massive netblock.
65.164.129.91 – Similarly weird traceroute result. Weirdness after Seattle.
208.252.91.3 – Another oddly terminating traceroute. Weird stuff starts happening after Seattle.

Sounds like someone is trolling through my site looking for addresses. Possibly from stolen netblocks? I haven’t bothered to look up weird routes because it’s bed time.

UPDATE2: It almost seems to be coordinated… Check this out:

65.164.129.91 – – [08/Feb/2005:23:46:34 -0500] “GET /gallery/livingroom_speakers HTTP/1.0” 200 15551 “-” “Mozilla/4.0 (compatible; MSIE 4.0; Windows NT; ……/1.0 )”
207.155.199.163 – – [08/Feb/2005:23:46:42 -0500] “GET /albums/livingroom_speakers/speakers_hung.highlight.jpg HTTP/1.0” 200 6193 “-” “Mozilla/4.0 (compatible; MSIE 4.0; Windows NT; ……/1.0 )”
12.17.130.27 – – [08/Feb/2005:23:46:50 -0500] “GET /gallery/livingroom_moulding HTTP/1.0” 200 15701 “-” “Mozilla/4.0 (compatible; MSIE 4.0; Windows NT; ……/1.0 )”
65.164.129.91 – – [08/Feb/2005:23:46:56 -0500] “GET /albums/livingroom_moulding/DCP_0920.highlight.jpg HTTP/1.0” 200 6225 “-” “Mozilla/4.0 (compatible; MSIE 4.0; Windows NT; ……/1.0 )”
65.164.129.91 – – [08/Feb/2005:23:47:05 -0500] “GET /gallery/livingroom_painting HTTP/1.0” 200 52236 “-” “Mozilla/4.0 (compatible; MSIE 4.0; Windows NT; ……/1.0 )”

It’s like a distinct set of addresses are rummaging through my site, one request per second.

computers moved from livejournal nuxx.net

Photo People…

February 7, 2005

Hmm, here’s a question. It seems that my new camera does a resolution that when resized to 25% of it’s original size, it’s 648 x 486. My gallery currently resizes photos to 640×480 for their normal, standard-viewing size.

I’m wondering if I should change it so that the photos are done at 648×486, because resizing to 25% original size could possibly produce a better resized image, as it’s exactly 4:1 instead of 4.05:1. I’d think that sampling exactly four pixels down into one would just be better and less likely to induce moire.

Thoughts?

computers moved from livejournal nuxx.net

Broadcom BCM95805

February 4, 2005

Okay, so I just won an auction for a Broadcom-based hardware crypto card. Since this is supported under FreeBSD 5.2-RELEASE for accelerating crypto(4), ipsec(4), and random(4) operations, I’d like to do some benchmarking of it to see just how much it does speed things up. I’m thinking of the following for doing a before/after test of random(4):

bornslippy:~% dd if=/dev/random of=/dev/null bs=1048576 count=100
100+0 records in
100+0 records out
104857600 bytes transferred in 16.046981 secs (6534413 bytes/sec)
bornslippy:~%

I’m also looking at maybe using John the Ripper as john –test if I can determine if it actually uses crypto(9). I think that OpenSSL’s benchmark openssl test would be a good idea, because it will provide the ability to look at the effects of the card on OpenSSL, as I’m fairly certain that it does use crypto(9).

Does anyone have any other ideas as to what might be a good benchmark? I don’t think I can really test ipsec(4), but since I won’t be using it, I don’t think it’ll matter too much.

Hmm, I hope my understanding of how the driver registers itself and what it utilizes it is in line with reality…

computers electronics moved from livejournal

:::[—]:::

February 4, 2005

So after eating today’s lunch over at Somerset Mall (sorry, The Somerset Collection) we took a walk through the Apple store. Turns out they had a demo Airport Express for $99, so I picked it up for the living room. Not a bad deal, I don’t think, seeing as student price is $119. It’s a little scratched up, but it’ll be tucked behind the stereo, so that doesn’t matter.

Also, all my email has been moved in to nuxx.net and is being accessed via IMAP. I’m still trying to figure out a good mail client to be used from home. Squirrelmail isn’t bad for web-based stuff, but it has two problems. One is absolutely massive CPU utilization when opening a large mailbox, which may be tweakable by either adding an SSL accelerator to the box and/or a Squirrelmail setting to have it cache stuff. The second is that (as far as I’m aware) I can only send email through it using the account which I logged in with.

I’m also wondering if moving the mail to a database instead of the filesystem would make sense. This will be a more drastic change, though, so I’m not so certain I want to do it.

The last idea is to upgrade the mail server, but there’s no reason why a P2-500 (which is what bornslippy.nuxx.net is) can’t push big quantities of IMAP. I think the clients are a problem.

Oh, and yes, I’m looking into using mutt. That’s one of the things on my plate to get working today.

computers moved from livejournal nuxx.net

Yay SOCKS!

February 3, 2005

Well, I got the SOCKS thing figured out. It seems that current builds of Mozilla, Firefox, Thunderbird and IE (?!?!) don’t support doing name resolution on the remote side of a SOCKS proxy. They all try to resolve the name locally, then use SOCKS for the actual connection itself. That just does’t make sense to me. heh.

Anyway, after digging around online for a few hours I found a two-year Bugzilla thread which discusses this. It finally looks like it’s coming around. So, I grabbed the most recent Mozilla nightly and… What do you know… There’s a checkbox for the option and it works great. Whee!

computers moved from livejournal

SOCKS v5

February 3, 2005

Has anyone here had any luck making Firefox and/or Thunderbird work properly with a SOCKS proxy? I’ve got one running via PuTTY and it works great for AIM and SOCKScap. I can’t get the damn thing to work with IE, Firefox, or Thunderbird. It appears that the apps are attempting to do their name resolution locally, so they don’t know where to connect to…

Any ideas?

computers moved from livejournal

Protected: !!!

February 3, 2005

computers moved from livejournal work