On Mon, Apr 11, 2011 at 8:51 AM, Sean Grove <[email protected]> wrote: > I'm using ZeroMQ 2.0.11, and the ruby ffi-rzmq (0.7.2) bindings, and I seem > to run into this. The code in question is a REQ/REP pair running in separate > processes, and everything works fine the first time through. The second > time, I get the above error. > I'd love to better understand what state the socket is in that it's dying > here - how can I go about fixing this?
If the second time through your REQ socket complains, it's because it didn't get a reply to its request, and you're trying to send a second request. Usually you can make this work by sticking to a synchronous request-reply dialog. If you really want an asynchronous dialog, switch to a dealer (xreq) socket. If you want to use a synchronous request-reply dialog but be able to resend a request, use the Lazy Pirate pattern from Ch4 of the Guide, i.e. close the socket, reopen, send again. -Pieter _______________________________________________ zeromq-dev mailing list [email protected] http://lists.zeromq.org/mailman/listinfo/zeromq-dev
