W3C home > Mailing lists > Public > ietf-http-wg@w3.org > January to March 2016

RE: GOAWAY after GOAWAY

From: Mike Bishop <Michael.Bishop@microsoft.com>
Date: Thu, 7 Jan 2016 19:01:40 +0000
To: Stefan Eissing <stefan.eissing@greenbytes.de>, HTTP Working Group <ietf-http-wg@w3.org>
Message-ID: <CY1PR03MB13741BA156465BBE4F3D9E7A87F50@CY1PR03MB1374.namprd03.prod.outlook.com>
Yes, it SHOULD send the GOAWAY, which implies that it can also reasonably close the connection without sending it.  It's not a MUST, because sending it is always optional and sometimes serves no real purpose.  You've identified one of the latter.

-----Original Message-----
From: Stefan Eissing [mailto:stefan.eissing@greenbytes.de] 
Sent: Thursday, January 7, 2016 10:39 AM
To: HTTP Working Group <ietf-http-wg@w3.org>
Subject: GOAWAY after GOAWAY

Question arising during implementation: 

A hypothetical http server receives a GOAWAY from a client with no streams open. Should it attempt to send a GOAWAY back or can it safely close the connection?

rfc 7540. 6.8 seems to say that endpoints SHOULD always send a GOAWAY. But it seems not useful in such a case. Probably I am overlooking something?

-Stefan
Received on Thursday, 7 January 2016 19:02:11 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 22 March 2016 12:47:10 UTC