[Outages-discussion] [outages] ipv6.netflix.com and www.ipv6.cisco.com down

[Ryan Rawdon]

Netflix took giant leaps backwards on the IPv6 front a few months ago when they started pushing more of their infrastructure to Amazon and elsewhere.  

After that, ipv6.netflix.com went to their IPv6-only site still, but just about any clickpath would get you onto their v4-only sites very quickly (generally by redirecting to movies.netflix.com instead of staying on ipv6.netflix.com like it used to).  Streaming could still be done over IPv6 if you manually changed the URL.

About a week after they moved more of their content to the cloud, their IPv6 streaming couldn't be accessed even by manually changing the URL to ipv6.netflix.com, it would redirect back to a v4-only hostname.

Their ipv6 site has about 1% of the IPv6-capable functionality it had 6 months ago.




On May 2, 2011, at 4:51 PM, Owen DeLong wrote:

> Shame on you Netflix...
> 
> Apparently they resolved the infinite recursion loop by simply redirecting
> to their IPv4-only name.
> 
> This is a giant step backwards for Netflix.
> 
> Owen
> 
> On May 1, 2011, at 1:58 PM, Frank Bulk wrote:
> 
>> Interesting.  If I hit it from a web browser it works, but from NAGIOS it
>> fails.
>> 
>> nagios:/usr/lib/nagios/plugins# ./check_http -6 -H ipv6.netflix.com -f
>> follow
>> Name or service not known
>> HTTP CRITICAL - Unable to open TCP socket
>> nagios:/usr/lib/nagios/plugins#
>> nagios:/usr/lib/nagios/plugins# ./check_http -6 -H ipv6.cnn.com -f follow
>> HTTP OK: HTTP/1.1 200 OK - 102475 bytes in 1.707 second response time
>> |time=1.707327s;;;0.000000 size=102475B;;;0
>> nagios:/usr/lib/nagios/plugins#
>> 
>> I added in the "-f follow" last week.  I know that ipv6.netflix.com
>> redirects to a v4-only site and this must cause that v6-specified check to
>> fail.  I will need to make a separate http check for such sites.
>> 
>> Thanks,
>> 
>> Frank
>> 
>> -----Original Message-----
>> From: Majdi S. Abbas [mailto:msa at latt.net] 
>> Sent: Sunday, May 01, 2011 3:13 PM
>> To: Frank Bulk
>> Cc: outages-discussion at outages.org
>> Subject: Re: [outages] ipv6.netflix.com and www.ipv6.cisco.com down
>> 
>> On Sun, May 01, 2011 at 03:01:54PM -0500, Frank Bulk wrote:
>>> Minor on the grand scale of things, but ipv6.netflix.com has been down
>> since
>>> Tuesday, 6:25 pm (Central) and www.ipv6.cisco.com since Saturday 10:50 am
>>> (Central).
>> 
>>   Funny, I can reach ipv6.netflix.com just fine from 2914:
>> 
>> -snip-
>> 3  ae-1.r06.snjsca04.us.bb.gin.ntt.net  37.971 ms  34.733 ms  34.014 ms
>> 4  2001:590:2:2b::1  53.566 ms  67.476 ms  60.543 ms
>> 5  ae0-70g.cr1.sjc1.us.nlayer.net  36.525 ms  81.402 ms  36.028 ms
>> 6  edge01.sv1.netflix.com  38.038 ms  34.834 ms  35.979 ms
>> 7  ipv6.netflix.com  45.034 ms  36.304 ms  36.940 ms
>> -snip-
>> 
>>   And yes, the page loads.  Say, http://ipv6.netflix.com/WiHome
>> for example.
>> 
>>> Of the roughly 15 IPv6 websites I monitor, ipv6.cogentco.com has had the
>>> greatest uptime over the last 31 days (100%), followed by ipv6.google.com.
>> 
>>   Where are you monitoring from?  
>> 
>>> Please direct any comments to outages-discussion.
>> 
>>   Can you expand on your methodology?
>> 
>>   Thanks,
>> 
>>   --msa
>> 
>> _______________________________________________
>> Outages-discussion mailing list
>> Outages-discussion at outages.org
>> https://puck.nether.net/mailman/listinfo/outages-discussion
> 
> _______________________________________________
> Outages-discussion mailing list
> Outages-discussion at outages.org
> https://puck.nether.net/mailman/listinfo/outages-discussion