[pca] PCA failling with some inetnum resolutions of a248.e.akamai.net

[Glenn Satchell]

On 04/07/10 18:22, Martin Paul wrote:
> Michael Jackson wrote:
>> Has anyone else noticed any change in the behaviour of the Akamai
>> download
>> servers in the last couple of days -- particularly if you're in the UK?
>
> Seems as if you were (temporarily) redirected to a different set of
> local servers, and at least one of them had problems. Judging from my
> logfiles, the name a248.e.akamai.net is constant (it's probably Sun's
> handle with Akamai).
>
> Here in Austria, I've been redirected to Akamai servers in the
> 193.170.140.* network almost exclusively for the last months.
>
>> Martin: As far as PCA is concerned, would it be possible to add
>> something in
>> the debugging output that showed a) the date/timestamp of the last
>> successful
>> Akamai download, and b) the a248.e ip address successfully downloaded
>> from.
>
> I'm not sure whether I understand what you really want here, but all
> that output is coming directly from wget anyway, so there's not much I
> can change here. If you want to collect debug data, I recommend setting
> up a local caching proxy with debug enabled. I have that, and so I could
> e.g. grep for "Resolving a248.e.akamai.net" in /tmp/pca-proxy-debug.txt
> easily to see whether something has changed.
>
> Martin.
>
That's the magic of akamai.net. Today this resolves to the following IP 
addresses, which just happen to be at my ISP, so I get nearly full speed 
downloads for patches :)

% host a248.e.akamai.net
a248.e.akamai.net has address 150.101.98.85
a248.e.akamai.net has address 150.101.98.72

I'm pretty sure these particular akamai servers have been in place for a 
couple of years at least, they are used by windows update and 
sun.com/downloads (solaris ISOs for example) as well.

-- 
regards,
-glenn