[zeromq-dev] zeromq-dev Digest, Vol 46, Issue 5

[Pieter Hintjens]

On Mon, Oct 3, 2011 at 9:41 PM, Johan Ström <johan at stromnet.se> wrote:

> Good, then I wasn't totally confused! But the theory behind ping
> silence as it was written in the comment is correct, right? As soon as
> a server have sent any kind of reply, we know it's identity has been
> seen, and thus any further pings (to a dead server) will be queued.
> But until we've heard anything, we can ping as much as we want without
> any risk of filled queues.

Yup, I guess so. Sorry, it's been a while since I wrote that code.
There's a page with some heartbeating strategies here:
http://www.zeromq.org/topics:heartbeating

-Pieter