|
| News | Profile | Code | Photography | Looking Glass | Projects | System Statistics | Uncategorized |
Blog |
I was downloading VMware Server 2.0 (RC1) today, and noticed something odd when running wget:
Resolving download3.vmware.com... 24.25.26.50, 24.25.26.83
So, 24.24/14 is Road Runner, and 24.25.26.0/24 resides in the Raleigh RDC (regional data center). Further investigation reveals that Road Runner is indeed hosting Akamai caches in certain regions, to (evidently) reduce transit costs. Here's a traceroute from Charlotte:
% tracepath download3.vmware.com.
1: destiny.prolixium.com (10.3.5.107) 0.112ms pmtu 1500
1: et3.starfire.prolixium.net (10.3.5.254) 0.629ms
1: et3.starfire.prolixium.net (10.3.5.254) 0.482ms
2: et0-0.e.prolixium.net (10.3.253.2) 1.884ms
3: 10.218.32.1 (10.218.32.1) 14.022ms
4: gig2-2.chrlncdnb-rtr1.carolina.rr.com (24.93.65.161) 15.616ms asymm 5
5: srp0-0.chrlncsa-rtr2.carolina.rr.com (24.93.67.175) 14.864ms
6: ge-7-3-0.chrlncpop-rtr1.southeast.rr.com (24.93.64.23) 23.056ms asymm 7
7: ge-0-3-0.rlghncpop-rtr1.southeast.rr.com (24.93.64.172) 27.707ms
8: te-0-25.akamai-swt.akamai.net (24.25.24.158) 21.955ms asymm 7
9: rdc-024-025-026-083.southeast.rr.com (24.25.26.83) 25.052ms reached
Resume: pmtu 1500 hops 9 back 57
And here's one from Menands, NY:
% tracepath download3.vmware.com.
1: cpe-74-76-220-132.nycap.res.rr.com (74.76.220.132) 0.242ms pmtu 1500
1: cable1-0.albynybhi-ar401.nyroc.rr.com (74.76.216.1) 12.582ms
1: cable1-0.albynybhi-ar401.nyroc.rr.com (74.76.216.1) 14.133ms
2: gig2-1-8.albynywav-rtr01.nyroc.rr.com (24.29.39.33) 13.471ms
3: ge-1-0-0.albynywav-rtr03.nyroc.rr.com (24.24.7.21) 12.537ms asymm 4
4: ge-0-0-0.syrcnycsr-rtr03.nyroc.rr.com (24.24.7.173) 32.309ms asymm 5
5: syrny-aka-cache-01.akamai.com (24.24.29.102) 21.402ms
6: syrny-aka-cache-node99.akamai.com (24.24.52.99) 19.799ms reached
Resume: pmtu 1500 hops 6 back 59
Now, the funniness comes into play when one realizes that Comcast is Akamai-powered:
% tracepath www.comcast.net.
1: destiny.prolixium.com (10.3.5.107) 0.114ms pmtu 1500
1: et3.starfire.prolixium.net (10.3.5.254) 0.728ms
1: et3.starfire.prolixium.net (10.3.5.254) 0.535ms
2: et0-0.e.prolixium.net (10.3.253.2) 2.034ms
3: 10.218.32.1 (10.218.32.1) 17.462ms
4: gig2-2.chrlncdnb-rtr1.carolina.rr.com (24.93.65.161) 15.766ms asymm 5
5: srp0-0.chrlncsa-rtr1.carolina.rr.com (24.93.67.161) 12.675ms
6: gig0-2.chrlnccrm-ubr2.carolina.rr.com (24.93.66.197) 17.814ms asymm 5
7: ge-7-3-0.chrlncpop-rtr1.southeast.rr.com (24.93.64.23) 30.691ms
8: ge-0-3-0.rlghncpop-rtr1.southeast.rr.com (24.93.64.172) 31.193ms asymm 7
9: te-0-25.akamai-swt.akamai.net (24.25.24.158) 27.696ms asymm 7
10: rdc-024-025-026-073.southeast.rr.com (24.25.26.73) 28.550ms reached
Resume: pmtu 1500 hops 10 back 57
Yeah, Road Runner is (indirectly) hosting Comcast's website.
I received a music CD and Roomba battery from Amazon, today. I'm a little confused about the extra jewel case, considering its condition:
It's not even the right kind of jewel case.
Another impulse buy of mine.
The good:
The suck:
Yeah, so I'm stuck with EDGE for now.
Update: HSDPA/3G doesn't work because I bought the wrong phone. The Nokia E71-1 (RM-346) doesn't support the frequencies that are used on AT&T's network in the United States. Blast.
Nope, not the movie, destiny uptime!
[destiny:17:50]% uptime 22:49:32 up 300 days, 2:57, 17 users, load average: 0.11, 0.09, 0.09 [destiny:22:48]% uname -a Linux destiny 2.6.22-gentoo-r5 #4 SMP Mon Sep 10 19:47:50 EDT 2007 \ x86_64 Intel(R) Core(TM)2 Quad CPU @ 2.66GHz GenuineIntel GNU/Linux
Yes, it's my desktop. Next checkpoint is at 365 days!
Day late and a dollar short, oh well.
Happy belated 4th of July! It took us over 1 hour and 45 minutes to exit the Charlotte Knights stadium after the fireworks were finished. Bleh.
One of our Linux boxes at work (that had 668 days of uptime!) blew up yesterday, and needed a reboot. Unfortunately, the box's sole purpose is to provide OOB connectivity via modem for some gear we have at remote locations, and the internal modem wasn't being picked up by the new 2.6.24-1-686 (Debian) kernel I installed after the box came back up.
So I picked up the only USB modem I saw at Best Buy, a Zoom V.92 USB unit (model 3095F). Although it's USB-powered and nice & small, it needs a small workaround in cdc-adm.c on Linux:
@@ -1242,6 +1242,9 @@ static struct usb_device_id acm_ids[] =
{ USB_DEVICE(0x22b8, 0x7000), /* Motorola Q Phone */
.driver_info = NO_UNION_NORMAL, /* has no union descriptor */
},
+ { USB_DEVICE(0x0803, 0x3095), /* Zoom Telephonics Model 3095F USB MODEM */
+ .driver_info = NO_UNION_NORMAL, /* has no union descriptor */
+ },
/* control interfaces with various AT-command sets */
{ USB_INTERFACE_INFO(USB_CLASS_COMM, USB_CDC_SUBCLASS_ACM,
See the original post to LKML here. It looks like it made its way into 2.6.25.6.
You know me, I'm fairly old-school, so I use MRTG for all network monitoring. Although I converted to RRDtool (rateup sucks) awhile back, I've still been sticking with the classic MRTG look via the mrtg-rrd CGI script.
FreeBSD ports upgraded RRDtool to 1.3.0 recently, and ended up breaking mrtg-rrd. The mrtg-rrd page looked like this:
It turned out the width and height (classic ol' HTML) parameters to the image tag were wrong. width was being set to 8 or 0, and height was being set to an outrageously large number. The image itself was intact, though. mrtg-rrd was populating these values from the 2nd and 3rd array indices of the return value from RRDs::graph, $rv[0] and $rv[1]:
my @rv = RRDs::graph($file, '-s', "-$back", @local_args,
@{$target->{args}}, "VRULE:$oldsec#ff0000",
"VRULE:$seconds#ff0000", @local_args_end);
Bug in RRDtool? Maybe. Could be something that RRDtool linked against. Anyway, the workaround is easy since the width and height are available from the config:
@@ -230,8 +230,8 @@
return unless defined $tgt->{$ext};
my @values = @{$params->[0]};
- my $x = $params->[1];
- my $y = $params->[2];
+ my $x = $tgt->{config}->{xsize};
+ my $y = $tgt->{config}->{ysize};
$x *= $tgt->{xzoom} if defined $tgt->{xzoom};
$y *= $tgt->{yzoom} if defined $tgt->{yzoom};
Now graphs are looking better, again.
Looks like ATDN caved and peered with Cogent. This elevates Cogent's status to a tier 1 network.
No, tier 1 networks don't necessarily provide better connectivity.
Displaying page 38 of 121 of 965 results
     |
This HTML for this page was generated in 0.007 seconds. |