Re: [squid-users] 206 Partial Content

From: Amos Jeffries <squid3_at_treenet.co.nz>
Date: Fri, 25 Nov 2011 13:17:41 +1300

On 25/11/2011 12:34 p.m., Ghassan Gharabli wrote:
> Hello Amos,
>
> Squid was not able to cache partial content 206 Responses such as this
> link below:
>
> http://dc122.4shared.com/img/97459254/dcf5c10e/dlink__2Fdownload_2Fe1D3g7qW_3Ftsid_3D20111124-220446-e01441a0/preview.mp3?sId=qnwkz3H9Cqm2NezD&t=1322172503748&s=fb361cc0390859a718b67a4646f7c16c
>
> AS I can see , the response doesnt have Content_range which make it
> harder to understand ..
>
> What are the tricky things to cache this response using Squid ?

 From redbot.org report it appears the server sends 206 responses even
if the request did not specify support for ranges or a range to receive.
This will break the workaround Squid uses to fetch full content (by
removing the range specifiers).

The tricky bit is that Squid stores the object along with its response
headers and sometimes meta information about the clients fetching it.
This is stored together in one cache object by most of the cache types.
The caching storage code does not yet have any logics to handle storing
only part of an object, or for handling when multiple pieces are known,
removing gaps etc.

Squid is only capable today of converting the range request into a
full-object request and responding with the needed ranges from a full
object. Or of relaying the request and not storing the partial response.

Sorry there is no easy fix.

Amos
Received on Fri Nov 25 2011 - 00:17:52 MST

This archive was generated by hypermail 2.2.0 : Fri Nov 25 2011 - 12:00:02 MST