- From: Mark Nottingham <mnot@mnot.net>
- Date: Thu, 2 Feb 2017 12:28:11 +1100
- To: Tom Bergan <tombergan@chromium.org>
- Cc: HTTP Working Group <ietf-http-wg@w3.org>
> On 2 Feb 2017, at 12:26 pm, Tom Bergan <tombergan@chromium.org> wrote: > > On Wed, Feb 1, 2017 at 5:20 PM, Mark Nottingham <mnot@mnot.net> wrote: > > > On 2 Feb 2017, at 7:41 am, Tom Bergan <tombergan@chromium.org> wrote: > > > > > Applications sometimes want requests to be retried by > > > infrastructure, but can't easily express them in a non-idempotent > > > request (such as GET). > > > > nit: did you mean "in an idempotent request (such as GET)"? > > Thanks, fixed in source. > > > > > > A client SHOULD NOT automatically retry a failed automatic retry. > > > > Why does RFC 7230 say this? I am aware of HTTP clients that completely ignore this suggestion, and I can't offhand think of a reason why this is a good rule-of-thumb to follow. > > Good question. The immediate answer is that RFC2616 said it, and RFC2068 said it before that (and apparently introduced the requirement). > > If we end up revising the text regarding retries, that's something we should consider updating too. > > Maybe it there was a concern about accidental DoS? Infinite retries are probably a bad idea without exponential backoff. I suspect so. This algorithm doesn't seem to match reality, as you point out, and so we'd need to either specify a different one, or just loosen it to a more general requirement. Cheers, -- Mark Nottingham https://www.mnot.net/
Received on Thursday, 2 February 2017 01:28:43 UTC