[Discuss] Annoying network problem

[dsr at randomstring.org (Dan Ritter)]

Laura Conrad wrote: 
> 
> I have an intermittent but quite frequent problem with my network.
> Symptoms include:
> 
>          Backups via deja-dup to google drive which are supposed to take
>          place nightly, a couple of times a week don't finish.
> 
>          When I come back to my computer after an absense of hours, the
>          tabs in my browsers need to be reloaded (e.g. amazon.com).
> 
>          The most annoying is that when I tell my mail reader (gnus) to get my
>          mail, it hangs.
> 
> I'm quite sure this is comcast's fault, because when I visit my friends
> in Arlington, I had this problem when they were on comcast but when they
> switched to RCN it stopped happening.  Switching to RCN is not an option
> either in Cambridge, where I live, or in Fall River, where I'm currently
> a homeless refugee because of a fire.
> 
> I believe this output demonstrates the problem:
> 
> ------------------------------------------------------mtr output---------------------
> mtr -rwzbc100 gwene.org
> Start: 2022-03-24T13:09:30-0400
> HOST: fortepiano                                                         Loss%   Snt   Last   Avg  Best  Wrst StDev
>   1. AS???    _gateway (10.0.0.1)                                         0.0%   100   50.6  39.6   2.5  56.8  14.8
>      AS7922   96.120.70.177
>   2. AS7922   96.120.70.177                                               0.0%   100   50.4  39.3  10.7  59.1  14.1
>   3. AS7922   96.108.56.173                                               0.0%   100   50.3  39.7  10.5  57.9  13.7
>   4. AS7922   96.108.68.178                                               0.0%   100   52.3  38.4   9.8  57.2  14.0
>   5. AS7922   be-341-ar01.needham.ma.boston.comcast.net (96.108.69.189)   0.0%   100   53.1  42.4  13.9  59.0  13.7
>   6. AS???    ???                                                        100.0   100    0.0   0.0   0.0   0.0   0.0
>   7. AS3356   4.69.162.181                                                0.0%   100  116.1 122.8  92.3 164.5  15.1
>   8. AS3356   AS33891-NET.edge7.Amsterdam1.Level3.net (212.72.41.234)     0.0%   100  125.9 120.8  93.4 137.7  13.6
>   9. AS201011 ae6-2011.nbg40.core-backbone.com (80.255.14.246)            0.0%   100  139.1 140.0 111.4 156.2  13.8
>  10. AS201011 core-backbone.hetzner.com (81.95.15.6)                      0.0%   100  131.7 133.3 104.3 150.5  13.6
>  11. AS24940  core11.nbg1.hetzner.com (213.239.229.161)                   0.0%   100  137.7 133.9 103.9 160.5  14.5
>  12. AS24940  ex9k1.dc5.nbg1.hetzner.com (213.239.245.2)                  0.0%   100  134.1 134.0 103.9 157.3  14.3
>  13. AS24940  ciao.gmane.io (116.202.254.214)                             0.0%   100  130.7 132.1 102.6 148.8  14.0
> 
> ------------------------------------------------------end mtr output---------------------
> 
> I would guess that that node that's losing 100% of the packets isn't
> supposed to do that.  Does anyone know a way to convince comcast to look
> at the problem and fix or work around it?

Sorry, that's not evidence of the problem that you are actually
having.

That's evidence that there is a router owned by Comcast or L3,
probably Comcast, that is not responsive to ICMP requests. This
is almost certainly by policy on their side, and the fact that
there is no packet loss to any point past that is strong
evidence that this router is not the problem.

It's also the case that gwene.org/gmane.io are not the
destinations of your Google Drive backups, so you're measuring
to the wrong endpoint.

I suspect you have a real and persistent issue, but we can't
diagnose what it is from this.

-dsr-