[zeromq-dev] Detecting sending failure due to over-full queue

Iñaki Baz Castillo ibc at aliax.net
Tue Oct 5 16:35:03 CEST 2010


2010/10/5 Pieter Hintjens <ph at imatix.com>:
> Iñaki,
>
> Could you explain your architecture?  Are you using a queue device or not?

It's just a theoric question for now :)

- A client open a XREQ socket and connects to a server (REP/XREP).
- It set the water mark to 5.
- It sends 5 requests using non-block. All of them fail (so they are
queued) in the XREQ outgoing queue.
- When it tries to send a new request zqm_send should reply EAGAIN
(the message couldn't be sent neither queued, am I right?).

My question is: how can the client determine that the request couldn't
be sent due to over-full queue in its XREQ socket?

I hope it's clear now :)

Thanks a lot.


-- 
Iñaki Baz Castillo
<ibc at aliax.net>



More information about the zeromq-dev mailing list