[zeromq-dev] ZeroMQ 3.2.2 on Linux [ My broker is not brokering ]

Pieter Hintjens ph at imatix.com
Thu Feb 14 19:44:13 CET 2013


Hi Alan,

These lbbroker examples have client, broker, and worker tasks as
threads in one process. You can run them; they do a minimal to and
fro, and exit.

If you want a real load-balancing broker, use mdbroker.c, mdclient.c,
mdworker.c and the respective client/worker APIs if you want to try
your own applications.

It's explained in the Guide in chapter 4, if you read the whole thing.

-Pieter



On Thu, Feb 14, 2013 at 6:40 PM, Alan Vogt <alanovogt at gmail.com> wrote:
>
> Hi All:
>
> I'm trying to find a load balancing router, whether done in C or C++.
>
> I've compiled and linked (and launched) several of the sample applications
> including:
>> lbbroker
>> lbbroker2
>> lbbroker3
>> lruqueue3
>
> Doing this with
> * Ubuntu 12.10 (GNU/Linux 3.5.0-22-generic i686) 32 bit
> * ZMQ 3.2.2
> * Some of these samples are based on CZMQ
>        which I built fresh just the other day.
>
> My problem is that the broker never seems to broker anything.
> Take lruqueue3 for instance: It compiles and links (and launches)
> successfully.
> But then it sits there apparently stalled, when I know that workers and
> clients have been actively generating messages.
>
> I added in some printf() statements and a request for VERBOSE, which gets me
> this far:
>
> MAIN-Checkpoint 01: Startup: Memory has been allocated.
> MAIN-Checkpoint 02: Startup: ROUTER sockets have been created.
> MAIN-Checkpoint 03: Startup: Endpoints have been bound.
> MAIN-Checkpoint 04: Client threads have been spawned.
> MAIN-Checkpoint 05: Worker threads have been spawned.
> 13-02-12 15:26:55 I: zloop: register ROUTER poller (0x95edbf0, 0)
> MAIN-Checkpoint 06: Reactor loop will be started next.
> 13-02-12 15:26:55 I: zloop: polling for 3600000 msec
> « NOW AN ENDLESS STALL with nothing from the »
> « front end or the back end. »
>
>
> Is it obvious to anyone why the broker is not recognizing anything from the
> client threads or the worker threads?
>
> Like I said, I have tried several of these examples. They all end up
> unresponsive though not crashed.
> In most cases I am building the samples without ANY modification at all.
> I've tried with IPC and with TCP.
>
> Thanks for any insights,
>
> Alan
>
>
>
>
> _______________________________________________
> zeromq-dev mailing list
> zeromq-dev at lists.zeromq.org
> http://lists.zeromq.org/mailman/listinfo/zeromq-dev
>



More information about the zeromq-dev mailing list