Monday, February 3, 2014

Comcast issues update

2/3/14 Update. 

The Comcast tech came out, as expected found nothing - all signals good, everything good.  He was nice enough to call around a few people and they confirmed they are seeing the packet loss I am seeing (at the same time when I see it).  One interesting finding - I can reliably observe packet loss and high latency when running the upstream portion of the Comcast speed test.  I believe this points to a capacity constraint on the upstream channels.

The tech also mentioned that in the last two weeks he's seen several people complain about the same issue, and he's always been unable to find anything wrong, and they have not been happy about it.  He thanked me for being nice and not taking it out on him.

I exchanged voicemails with someone in the Executive Customer Support, whatever that is.  We'll see if that yields any results.  

Sunday, February 2, 2014

Comcast issues

For the benefit of Comcast, here's a summary of the problem with my internet connection.

At times, I am observing a high level of latency trying to go out to internet sites.  This latency at times leads to complete packet loss, rendering my internet connection not only slow, but unusable.
During these times of latency/loss, I can continue to ping my side of the modem with zero packet loss and no change in latency, so the problem is not on my side.

So far comcast has:
  • replaced the line from the pole to the house
  • verified signal levels are perfectly acceptable
  • verified that when the tech is here, I get the speed I should be getting.  I've had about 3 visits in the last month, and the last one will be tomorrow (because I am cancelling if you don't fix this).
I have on my own:
  • confirmed that the problem is not in my wireless, home firewall, or modem
  • despite that, replaced all the above equipment just to be sure it's really not me
The problem happens sporadically, and happens even when I connect my laptop directly to the (brand new) modem, so it's not my wireless, my router, my modem or any other piece of equipment.

This morning I (yet again) observed the problem and called Comcast support (well, I first called retention to cancel, but they don't work on Sunday apparently).

As I type this I am on the phone with Comcast.  I am observing packet loss when pinging the very first hop past the modem:


$ traceroute speedtest.comcast.net
traceroute to speedtest.g.comcast.net (69.241.64.6), 64 hops max, 52 byte packets
 1  192.168.1.1 (192.168.1.1)  1.282 ms  1.266 ms  1.370 ms
 2  50.174.106.1 (50.174.106.1)  20.315 ms  18.839 ms  26.483 ms
 3  te-7-4-ur03.walnutcreek.ca.sfba.comcast.net (68.85.191.121)  17.233 ms  12.697 ms  15.860 ms
 4  te-1-5-0-3-ar01.sfsutro.ca.sfba.comcast.net (68.85.154.194)  16.972 ms  13.595 ms  15.883 ms
 etc..



$ ping 50.174.106.1
PING 50.174.106.1 (50.174.106.1): 56 data bytes
(normal pings deleted here for space)
64 bytes from 50.174.106.1: icmp_seq=566 ttl=255 time=18.855 ms
64 bytes from 50.174.106.1: icmp_seq=567 ttl=255 time=10.305 ms
64 bytes from 50.174.106.1: icmp_seq=568 ttl=255 time=9.342 ms
64 bytes from 50.174.106.1: icmp_seq=569 ttl=255 time=13.433 ms
64 bytes from 50.174.106.1: icmp_seq=570 ttl=255 time=10.279 ms
64 bytes from 50.174.106.1: icmp_seq=571 ttl=255 time=53.718 ms
64 bytes from 50.174.106.1: icmp_seq=572 ttl=255 time=72.584 ms
64 bytes from 50.174.106.1: icmp_seq=573 ttl=255 time=90.848 ms
64 bytes from 50.174.106.1: icmp_seq=574 ttl=255 time=114.551 ms
64 bytes from 50.174.106.1: icmp_seq=575 ttl=255 time=128.611 ms
64 bytes from 50.174.106.1: icmp_seq=576 ttl=255 time=143.017 ms
64 bytes from 50.174.106.1: icmp_seq=577 ttl=255 time=154.146 ms
64 bytes from 50.174.106.1: icmp_seq=578 ttl=255 time=170.938 ms
64 bytes from 50.174.106.1: icmp_seq=579 ttl=255 time=180.391 ms
64 bytes from 50.174.106.1: icmp_seq=580 ttl=255 time=191.445 ms
64 bytes from 50.174.106.1: icmp_seq=581 ttl=255 time=184.840 ms
64 bytes from 50.174.106.1: icmp_seq=582 ttl=255 time=100.994 ms
64 bytes from 50.174.106.1: icmp_seq=583 ttl=255 time=123.883 ms
64 bytes from 50.174.106.1: icmp_seq=584 ttl=255 time=78.710 ms
64 bytes from 50.174.106.1: icmp_seq=585 ttl=255 time=97.007 ms
64 bytes from 50.174.106.1: icmp_seq=586 ttl=255 time=120.788 ms
64 bytes from 50.174.106.1: icmp_seq=587 ttl=255 time=137.364 ms
64 bytes from 50.174.106.1: icmp_seq=588 ttl=255 time=146.512 ms
64 bytes from 50.174.106.1: icmp_seq=589 ttl=255 time=19.653 ms
Request timeout for icmp_seq 597
Request timeout for icmp_seq 598
Request timeout for icmp_seq 599
Request timeout for icmp_seq 600
Request timeout for icmp_seq 601
Request timeout for icmp_seq 602
Request timeout for icmp_seq 603
Request timeout for icmp_seq 604
Request timeout for icmp_seq 605
Request timeout for icmp_seq 606
Request timeout for icmp_seq 607
Request timeout for icmp_seq 608
Request timeout for icmp_seq 609
64 bytes from 50.174.106.1: icmp_seq=608 ttl=255 time=2862.261 ms
64 bytes from 50.174.106.1: icmp_seq=609 ttl=255 time=1861.965 ms
64 bytes from 50.174.106.1: icmp_seq=610 ttl=255 time=862.106 ms
64 bytes from 50.174.106.1: icmp_seq=611 ttl=255 time=10.016 ms
64 bytes from 50.174.106.1: icmp_seq=612 ttl=255 time=22.393 ms
Request timeout for icmp_seq 615
Request timeout for icmp_seq 616
Request timeout for icmp_seq 617
Request timeout for icmp_seq 618

The Comcast support rep was unable to find anything wrong, despite understanding all the information above.  He said he could not receive email so I could not supply this info and could only suggest that yet another tech should come out.  

10 minutes after I agree to send another tech and hang up, the connection is again working fine:
$ ping 50.174.106.1
PING 50.174.106.1 (50.174.106.1): 56 data bytes
64 bytes from 50.174.106.1: icmp_seq=0 ttl=254 time=38.244 ms
64 bytes from 50.174.106.1: icmp_seq=1 ttl=254 time=13.285 ms
64 bytes from 50.174.106.1: icmp_seq=2 ttl=254 time=12.524 ms
64 bytes from 50.174.106.1: icmp_seq=3 ttl=254 time=11.720 ms
64 bytes from 50.174.106.1: icmp_seq=4 ttl=254 time=10.986 ms
64 bytes from 50.174.106.1: icmp_seq=5 ttl=254 time=14.048 ms
64 bytes from 50.174.106.1: icmp_seq=6 ttl=254 time=12.569 ms
^C
--- 50.174.106.1 ping statistics ---
7 packets transmitted, 7 packets received, 0.0% packet loss
round-trip min/avg/max/stddev = 10.986/16.197/38.244/9.048 ms

I monitor my home connection with Pingdom.  The 30 day results are quite telling:



Yes, my connection has been down 102 times in the last 30 days.  If you look at the last 7 days, the week I've had to work from home recovering from knee surgery, when I really need my connection to be reliable:



The Comcast tech support guy agreed with my assessment that there's got to be a problem somewhere upstream, and that yes, there's some advanced networking group within Comcast whose job it is to fix this issue. He did not however know how to contact that group, nor do I.  

In a previous attempt to cancel, the retention specialist vowed to find that group and get them engaged.  But I never heard back and continue to have this issue.  Then the next retention specialist who begged for one last chance said she would do a "thorough review" of the account notes and call me back.  She never did.

Why am I still with Comcast?  Because when it works, it works well and is somewhat reasonably priced (as long as you threaten to cancel if you don't get the promo pricing).  ATT Uverse is not a great option, and there's only one other cable provider in my area (Astound) with decent bandwidth but their cable offerings are subpar.

So Comcast, here's the challenge: fix my internet, this week, or I'm switching to Astound for internet and I'll happily go back to Dish for my TV needs, or just cut the cord.


Friday, June 17, 2011

Velocity 2011


I wasn't able to attend the sessions at Velocity 2011, but here is what I was able to glean from the Twitter feed.  If you have more links, feel free to tweet or comment!

Tuesday:










Wednesday:
Yslow bookmarklet:

mobile html5: http://t.co/5AA47Ft







Thursday:

Google Pagespeed: http://t.co/SkDDDPN







Wednesday, September 2, 2009

Cisco ASDM is broken in Snow Leopard

See link: http://discussions.apple.com/thread.jspa?messageID=10091457&tstart=0
No known solution yet...

Wednesday, June 24, 2009

building a triple boot Dell Mini 10v

The goal: to put OSX on a Dell Mini 10v. But why stop there? I decided to make it a triple boot.
After much installing and reinstalling, here are some tips that helped me along the way.

First off, if all you want is OSX, follow this guide:
http://dellefi.mechdrew.com/guide/method1.shtml
It's pretty straightforward, but follow everything to the letter. For example, if you decide to call your disk "Fred" instead of "OSXMINI9" the install will fail.

Now, if you want to do two or more different OSs, you need to know:
a) Don't partition the drive as GUID, partition as MBR. You can make a gparted boot disk to do this, but if you've got the modified OSX installer, that will be just fine.
b) To install OSX on a MBR partitioned disk, you need to modify the installer. See this:
http://dellefi.mechdrew.com/guide/advanced3.shtml

Quirks I had to work around:
a) Installing Linux and Windows7 were trivial. The main problem I had was with OSX.
At various times, it would boot and then panic, sometimes in plain text mode, sometimes with the equivalent of the BSOD. To fix that, I found I had to boot from the OSX Installer USB key, open up a terminal, and issue this command:
rm -rf /System/Library/Extensions/AppleIntelCPUPowerManagement.kext/

b) After Windows7 was installed, the system would only boot into Windows. I booted up using the USB key, booted my Ubuntu partition, and used gparted to look at the drive. The only partition marked "boot" was the Win7 partition, so I switched boot to the Ubuntu partition. After that, everything worked fine.

So, I now boot up into the Chameleon bootloader. If I want to switch from the default (OSX), I hit the - key, and it shows me a list of my OS choices.

c) You can use an external CD/DVD. I found that the Ubuntu installer kept failing, so I made a USB installer which worked fine. Could be bad media, bad drive, not sure. Stick to USB is my suggestion.

Thursday, April 30, 2009

turvy over topsy

Just blocked a new bot/crawler called Topsy. On their webpage, they pretend to be cool and tell you just exclude them with robots.txt. But this is how they crawl you:
"Mozilla/5.0 (Macintosh; U; Intel Mac OS X; en-US; rv:1.8.1.1) Gecko/20061204 Firefox/2.0.0.1"

Does that look like a bot to you? Nah, me neither. Uncool. I blocked their IP (208.74.66.43) and will add the subnet if necessary.

Wednesday, April 29, 2009

libtool problems on ubuntu

I was trying to build sysbench on ubuntu, and hit some libtool issues. I saw these on several versions of Ubuntu (8.04.2 and 9.04) so clearly it couldn't be a strict ubuntu issue.

The symptom:
../libtool: line 838: X--tag=CC: command not found
../libtool: line 871: libtool: ignoring unknown tag : command not found
../libtool: line 838: X--mode=link: command not found
../libtool: line 1004: *** Warning: inferring the mode of operation is deprecated.: command not found
../libtool: line 1005: *** Future versions of Libtool will require --mode=MODE be specified.: command not found
../libtool: line 2231: X-g: command not found
../libtool: line 2231: X-O2: command not found
../libtool: line 2400: Xsysbench: command not found
X: user not authorized to run the X server, aborting.
../libtool: line 2412: Xsysbench: command not found
../libtool: line 2420: mkdir /.libs: No such file or directory
mkdir: cannot create directory `/.libs': Permission denied
make[2]: *** [sysbench] Error 1
make[2]: Leaving directory `/home/dan/sysbench-0.4.12/sysbench'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/dan/sysbench-0.4.12/sysbench'
make: *** [all-recursive] Error 1

The solution:
libtoolize --force --copy; aclocal; autoconf; automake; make