Apologies for the second email, I found the relevant setting in the docs: proxy.config.http.chunking.size for anyone else interested
On Oct 31, 2016, at 2:38 PM, Eric Friedrich (efriedri) <[email protected]<mailto:[email protected]>> wrote: Thanks Bryan- If I set proxy.config.http.chunking_enabled to: 3Generate a chunked response if the client request is HTTP/1.1 and the origin server has previously returned HTTP/1.1. ATS will store the object as not chunked, what defines the size of the chunk? —Eric On Oct 31, 2016, at 1:51 PM, Bryan Call <[email protected]<mailto:[email protected]>> wrote: inline reply -Bryan On Oct 31, 2016, at 10:32 AM, Eric Friedrich <[email protected]<mailto:[email protected]>> wrote: How does Traffic Server handle Chunked-Transfer encoded content? I see in the docs the setting for proxy.config.http.chunking_enabled but have some additional questions. 1) Is the chunk size cached as part of the content? Will ATS always use the exact same chunk size in the client response or can this be modified? ATS caches the object as not chunked. 2) How long will Traffic Server keep a connection to the origin server for pending additional chunks? You can configure this to what you would like by setting proxy.config.http.transaction_no_activity_timeout_out. # this is our current setting on a host [bcall@l10 trafficserver]$ traffic_ctl config get proxy.config.http.transaction_no_activity_timeout_out proxy.config.http.transaction_no_activity_timeout_out: 60 3) In the docs, what does pipelining mean? Does it mean that multiple connections are made from a cache to the origin server to avoid blocking other requests? Pipelining is used for HTTP/1.1 connections and most browsers disable this by default. As I recall we open the requests serially. https://en.wikipedia.org/wiki/HTTP_pipelining Thanks, Eric
