PGM tends to take advantage of the hardware acceleration that routers often
provide, and is widely supported. So, effectively, repeatedly multicasting
some data out to many destinations with just some processor based software
solution. But because this is an intensive task that is often performed,
It's really really hard to keep documentation up to date. The style of
documentation that is provided to the community is spectacular, however
documentation is usually the icing on the cake when it comes to a free
software project.
On top of being incredibly time consuming, documentation gets
Sounds like you need malamute. Check it out, let me know if that solves
your problem.
On Aug 16, 2015 8:17 PM, Bob Stanton bobstant...@gmail.com wrote:
I need help with a system with bidirectional communication between a
Client a Server and another Client. The Server need to pass messages
Whoops, I spoke too soon. Yeah um, ignore that, I missed a part of that.
On Aug 16, 2015 11:48 PM, Kenneth Adam Miller kennethadammil...@gmail.com
wrote:
I'm pretty sure that bind_out_sock should be set up with a connect call,
and the bind_ prefixes might be left off.
On Aug 14, 2015 10:38 AM
I'm pretty sure that bind_out_sock should be set up with a connect call,
and the bind_ prefixes might be left off.
On Aug 14, 2015 10:38 AM, Jason Sia jsi...@gmail.com wrote:
Hi,
I'm using zeromq 2.2.0 version due to some library dependency, I created a
streamer device push pull , for some
Can you run: file on the libzmq .so library that was installed by your apt
command and tell the output here? It may be that the target that was
compiled for is not suitable for your architecture.
On Aug 12, 2015 5:14 PM, Kelly Beard kenverybigl...@gmail.com wrote:
Running on a Raspberry Pi 2
Did you remember to set the option before you make your bind/connect?
Did you also enable zmq_req_correlate?
Do you have a code sample? Also, other people have had trouble with this
feature before too, so you might try reading what they wrote last time.
On Aug 12, 2015 4:07 PM, Ilya Kulakov
of the problem is in the makefile generation
infrastructure with automake-that has to be told so that the makefiles that
it generates will be correct.
On Fri, Aug 7, 2015 at 6:20 PM, Pieter Hintjens p...@imatix.com wrote:
Ah... Sorry. It's make dist that fails.
On 7 Aug 2015 21:11, Kenneth Adam Miller
Where can I checkout something to debug what you've got? I'll work on it
for you.
On Fri, Aug 7, 2015 at 4:29 AM, Pieter Hintjens p...@imatix.com wrote:
Hi guys,
I'm trying to fix issue #1505 by adding Makefile.am into each builds/
subdirectory.
This fails miserably with make[1]: *** No
.
:/
On Wed, Jul 29, 2015 at 4:55 AM, Bjorn Reese bre...@mail1.stofanet.dk
wrote:
On 07/29/2015 01:28 AM, Kenneth Adam Miller wrote:
But the problem that this solves brings up a different problem: if the
single threaded programs poll to write, then the cycle could deadlock
where every side
I can't use ZeroMQ for this, because I know if I could this wouldn't be a
problem. But since there are so many systems engineers on the channel and
fundamentally it's what ZMQ aims to solve, I think it's the appropriate
place.
For a scalability benchmark meant to emulate our real world scenario,
you've still got
pending outbound data to write, and then you will face those three choices
I mentioned.
On Tue, Jul 28, 2015, at 05:07 PM, Kenneth Adam Miller wrote:
Thank you, by the way.
On Tue, Jul 28, 2015 at 8:07 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
There's
The way this deadlock all came about was reluctance to isolate that polling
code for the fact that the code base is large.
On Tue, Jul 28, 2015 at 8:24 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
That's ok, this is only for a simulation test. I'm not focused on
congestion
Thank you, by the way.
On Tue, Jul 28, 2015 at 8:07 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
There's no need to take any of those options because in my case, because
the producer polls every time before sending.
I think I realize with your first sentence what I missed when
, at 04:28 PM, Kenneth Adam Miller wrote:
I can't use ZeroMQ for this, because I know if I could this wouldn't be a
problem. But since there are so many systems engineers on the channel and
fundamentally it's what ZMQ aims to solve, I think it's the appropriate
place.
For a scalability
AM, Leonard Michelet
leonard.miche...@openwide.fr wrote:
I use IPC socket on linux and it use named pipe (the address socket must
be something like ipc://path/to/the/pipe).
I hope it answer the question...
- Mail original -
De: Kenneth Adam Miller kennethadammil...@gmail.com
À
Does inproc or ipc use fifo pipes in ferrying the data between sockets in
the underlying library?
___
zeromq-dev mailing list
zeromq-dev@lists.zeromq.org
http://lists.zeromq.org/mailman/listinfo/zeromq-dev
How could a vtkRenderWindow possibly be sent between a socket over machines
or even processes? You know that the C++ object representation of a
vtkRenderWindow is set of contiguous member data right?
Bad address looks like you passed a pointer to zmq that wasn't valid.
Likely you should debug it
Well one shortcut you could use in malamute that would get you off the
ground immediately is to alter the endpoint routes for each of the
respective groups so that you can interject a process chain. In this way,
you can delegate the management of those clients by having malamute reflect
that
I don't have any problem or question explicitly stated, or even that there
is some needed improvement that you don't have, so it's hard to know how to
be helpful. The one thing I would say that I have experience with and that
I could also help with is the idea of your external service managing
malamute. I'll take a look at it.
Thanks,
Koren
On Mon, Jul 20, 2015, 00:21 Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
I don't have any problem or question explicitly stated, or even that
there is some needed improvement that you don't have, so it's hard to know
how
software today, and I'm even still working on that. I
have edits to like 4-7 open source projects sitting around with my trying
to get everything pristine before shipping. Sorry! I'm shipping ocaml-zmq
changes at this moment.
On Fri, Jun 26, 2015 at 8:42 PM, Kenneth Adam Miller
kennethadammil
Oh, I have something to merge in to allow the newest version to work well
together, let me do a merge request before you execute this please.
On Wed, Jun 17, 2015 at 6:43 PM, Pieter Hintjens p...@imatix.com wrote:
Sounds good to me. It never developed as a separate project, did it.
On Thu,
you will often get resistance
for reasons that have nothing to do with the technology itself.
-Pieter
On Thu, May 14, 2015 at 6:33 AM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
So, recently in the workforce the discussion came up that despite all of
the nice work that has
Great, these are excellent points. I can use this!
On Thu, May 14, 2015 at 9:34 AM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
Oh thank god! Someone that understands!
On Thu, May 14, 2015 at 4:11 AM, Pieter Hintjens p...@imatix.com wrote:
Why even use an operating system? It's
So, recently in the workforce the discussion came up that despite all of
the nice work that has been accomplished with ZMQ, it should be removed
because using it as a component for distributed analysis creates
complication. Much to my dismay, they cannot be convinced otherwise, and
see
For malamute, are service requests dispatched per recipient client recv
call, or sent in bulk divided equally to all connected workers upon request
reception?
___
zeromq-dev mailing list
zeromq-dev@lists.zeromq.org
to a minimal
test case. Thanks!
On Sun, Apr 12, 2015 at 9:27 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
I mean a github issue on the malamute repo.
On Sun, Apr 12, 2015 at 3:27 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
It is reproduced in that git repo
stuff
in it that makes it impossible for me to use as-is.
Cheers
Pieter
On Sat, Apr 11, 2015 at 8:39 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
Allow me to clarify:
The unit test doesn't specifically fail, it in fact just hangs. I don't
know
what to do make sure
I mean a github issue on the malamute repo.
On Sun, Apr 12, 2015 at 3:27 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
It is reproduced in that git repo that I linked at the bottom of my last
email, and the main file for that is very short. I've been struggling with
subtly
Hintjens p...@imatix.com wrote:
On Sun, Apr 12, 2015 at 7:59 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
Key here is the message client A sent is persisted, but if this were
swapped
out for service semantics APIs, the tool would hang at client B reading.
Hmm, I think
having this issue?
On Fri, Apr 10, 2015 at 8:23 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
I've found that, with similar unit tests, there no way to ensure that
before a send occur that there is a client on the other side in order to
make sure that the broker will create a queue
*);
mlm_client_recvx(client, topic, content, NULL);
assert(!strcmp(content, gotit));
mlm_client_destroy(client);
shutDown();
}
start up and shut down just create an in process malamute actor and bind it
to port .
On Fri, Apr 10, 2015 at 2:41 PM, Kenneth Adam Miller
kennethadammil...@gmail.com
connect asynchronously, each subscribed to what the other side
produces, in is implausible to prevent messages being lost. If messages get
lost this way, since I can't timeout on receive operations, the clients
deadlock.
On Fri, Apr 10, 2015 at 6:27 PM, Kenneth Adam Miller
kennethadammil...@gmail.com
test. Thanks so much for your help!
On Wed, Apr 1, 2015 at 3:10 AM, Pieter Hintjens p...@imatix.com wrote:
You have to at the lest:
- start the actor
- tell it what endpoint to bind on OR tell it what config file to load
-Pieter
On Wed, Apr 1, 2015 at 5:56 AM, Kenneth Adam Miller
On Wed, Apr 1, 2015 at 5:42 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
Oh ok... I'm trying to find out how to start an actor.
I tried telling the broker what endpoint to bind on in source with:
zstr_sendx(server, mlm_server, BIND, endpoint, tcp://*:,
NULL
So, I wanted to create a specific broker on a dedicated thread that I could
start up and shut down easily in order to facilitate testing. Here's what I
had for that, consulting mlm_shell, malamute.c and mlm_tutorial:
void startUP(bool verbose=true, zactor_t **server=global_server) {
std::cout
Since you have acknowledged that a UDP stack is needed on android in C, do
you have a public repository where you're testing and working on it?
On Mon, Mar 23, 2015 at 9:42 AM, Arnaud Loonstra arn...@sphaero.org wrote:
On 2015-03-20 14:41, Pieter Hintjens wrote:
Could you provide me a spec
. That was one thing we took
home from the workshops. This means Bluetooth in C on Android...
On Fri, Mar 13, 2015 at 4:08 AM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
Found this:
http://zguide.zeromq.org/php:chapter8
On Thu, Mar 12, 2015 at 11:05 PM, Kenneth Adam Miller
on their findings.
On Thu, Mar 12, 2015 at 6:01 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
As per title, has anyone ever gotten Zyre to run on Android?
Also, mesh networking is a research top the last I had learned. How well
do does Zyre work? How extensively has it been tested
Found this:
http://zguide.zeromq.org/php:chapter8
On Thu, Mar 12, 2015 at 11:05 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
Hey, I saw hydra and had a tab open because it was interesting, but I
couldn't find any meta information on it. Can you tell me about it? I knew
it had
As per title, has anyone ever gotten Zyre to run on Android?
Also, mesh networking is a research top the last I had learned. How well do
does Zyre work? How extensively has it been tested?
___
zeromq-dev mailing list
zeromq-dev@lists.zeromq.org
I have replied inline.
On Wed, Mar 11, 2015 at 8:09 AM, Riskybiz riskybizl...@live.com wrote:
If I an application uses zmq_poll() before sending or receiving messages
to check socket(s) for these events;
*ZMQ_POLLIN*
For ØMQ sockets, at least one message may be received from the *socket*
Oh whoops, wrong link. Here, read this:
http://zeromq.org/whitepapers:architecture
On Tue, Mar 10, 2015 at 10:45 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
I'm pretty sure that's exactly what happens. ZMQ handles all of the
buffering in the background so your application can
I'm pretty sure that's exactly what happens. ZMQ handles all of the
buffering in the background so your application can look clean and good.
Check out this question, fielded by Pieter himself:
http://stackoverflow.com/questions/10528659/what-is-the-rationale-behind-zeromq-context
On Tue, Mar
Ok, well when you have some code let me know and I will review.
On Tue, Mar 10, 2015 at 10:18 AM, Yassine Lamgarchal boucek...@gmail.com
wrote:
Hi, inline :).
2015-03-09 19:25 GMT+01:00 Kenneth Adam Miller
kennethadammil...@gmail.com:
Inline.
On Mon, Mar 9, 2015 at 7:12 AM, Yassine
at your above response. This seems to be a newer, different topic.
Please explain what you mean more.
--
Yassine
2015-03-09 3:16 GMT+01:00 Kenneth Adam Miller kennethadammil...@gmail.com
:
Replied inline
On Sun, Mar 8, 2015 at 7:55 PM, Yassine Lamgarchal boucek...@gmail.com
wrote
Replied inline
On Sun, Mar 8, 2015 at 7:55 PM, Yassine Lamgarchal boucek...@gmail.com
wrote:
Hello all o/,
I'm currently designing a distributed application which is composed of
clients
and a remote server.
I struggle with the asynchronous client api... I want it to be the most
simple
, Kenneth Adam Miller kennethadammil...@gmail.com wrote:
On your example, are you running the requisite request receiver, eg the
python script that binds to port in order that the server can have
something to connect to? Just checking.
Well it depends on what all you want it to do in what
-
Am 06.03.2015 um 17:29 schrieb Kenneth Adam Miller
kennethadammil...@gmail.com:
I notice that in your public code, you wait until you get a request on
the sub socket, and then publish on the workers socket. The workers socket
is a pub socket, and just receiving a request on the sub socket
I notice that in your public code, you wait until you get a request on the
sub socket, and then publish on the workers socket. The workers socket is a
pub socket, and just receiving a request on the sub socket doesn't mean
that all workers have connected. Why not, before you start publishing at
What if instead of sending from the tail, you just used a push socket that
the head could wait on in a loop. The REP socket could send a reply to the
client immediately after having received from the tail on a pull socket. In
this way, you still get what you want, and don't have to deal with the
...@gmail.com wrote:
Hi Kenneth, Thank you. Some comments/questions below.
On Fri, Mar 6, 2015 at 11:59 AM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
What if instead of sending from the tail, you just used a push socket
that the head could wait on in a loop. The REP socket could
brokers.
On Tue, Mar 3, 2015 at 6:21 AM, Pieter Hintjens p...@imatix.com wrote:
:-) I appreciate your work on learning how this works. Sorry I don't
already have a tutorial written. Things are a bit busy. I'll push that
to top priority.
On Tue, Mar 3, 2015 at 10:20 AM, Kenneth Adam Miller
if the reader
is a microsecond too late, it will miss the message.
On Tue, Mar 3, 2015 at 7:32 AM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
This is what I currently have right now-it seems to lock up sometimes...
char * exchange_addresses(std::string consumer_topic, std
Ok, a guide written by you would be really good, thanks. I just wanted to
help.
On Tue, Mar 3, 2015 at 3:38 AM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
Ok, I understand.
I was thinking maybe it would better to just loop and continuously send
requests for the service until
So, I have the following code:
void connectToFrontEnd() {
std::cout connectToFrontEnd std::endl;
zsock_t *inproc_await = zsock_new_rep(inproc://frontend);
std::cout frontend exchange_addresses std::endl;
char * servrAddr = exchange_addresses(backendEndpoints,
frontendEndpoints,
Wait, is it because each of the peers have specified the same mailbox?
On Mon, Mar 2, 2015 at 4:12 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
But only one side gets a message from the broker. The other side just
freezes.
On Mon, Mar 2, 2015 at 3:54 PM, Pieter Hintjens p
But only one side gets a message from the broker. The other side just
freezes.
On Mon, Mar 2, 2015 at 3:54 PM, Pieter Hintjens p...@imatix.com wrote:
Sure, it'd work as subjects.
On Mon, Mar 2, 2015 at 8:07 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
What are the two
, Mar 2, 2015 at 10:25 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
Yeah that fixed it!
Now I just have to iron out what precisely is concurrent.
On Mon, Mar 2, 2015 at 4:24 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
Wait, is it because each of the peers
Is there any limit to the number of mailboxes on the malamute broker? How
do I manage mailbox lifetimes, or is that something that I need to consider?
On Mon, Mar 2, 2015 at 8:21 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
Ah, now that I considered it in the context of each side
...@imatix.com wrote:
There's no limit on mailboxes, and you don't need to consider
lifetimes. Consider these like email addresses. Mailboxes will at some
point have to be saved to disk (they're memory only now.)
On Tue, Mar 3, 2015 at 4:15 AM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote
almost there though!
On Tue, Mar 3, 2015 at 1:20 AM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
As it turns out, I think that there was some misunderstanding in terms of
address exchange semantics. I think you thought it necessary to persist the
messages, hence the requirement
the service
semantics, and the lookup service can talk to the broker over inproc
or tcp as it wants (it could be a thread in the same process, or a
separate process).
On Sun, Mar 1, 2015 at 9:00 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
So, in order to manage a mutual exchange
Oh you mean with mlm_client_set_worker! Do I do set_worker on each side
with different service names? How does a client get a specific service?
On Mon, Mar 2, 2015 at 11:26 AM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
Service semantics? I don't know what those are...
I read what
;
}
Problem is, both threads block at mlm_client_recvx... As per example, it
looks correct.
On Mon, Mar 2, 2015 at 11:30 AM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
Oh you mean with mlm_client_set_worker! Do I do set_worker on each side
with different service names? How does
I got it to work by setting the subscribed topic to inproc* on
mlm_client_set_worker call.
On Mon, Mar 2, 2015 at 12:07 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
Ok after looking at mlm_client.c, I have the following:
Two concurrent calls to exchange addresses
the replies, which end in a client's
mailbox
- read the replies using the recv method
For this to work, peers need to specify a mailbox address in the connect
method.
If you like I'll write an article and make examples.
-Pieter
On Mon, Mar 2, 2015 at 6:45 PM, Kenneth Adam Miller
examples.
-Pieter
On Mon, Mar 2, 2015 at 6:45 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
I got it to work by setting the subscribed topic to inproc* on
mlm_client_set_worker call.
On Mon, Mar 2, 2015 at 12:07 PM, Kenneth Adam Miller
kennethadammil...@gmail.com
On your example, are you running the requisite request receiver, eg the
python script that binds to port in order that the server can have
something to connect to? Just checking.
Well it depends on what all you want it to do in what scenario. If these
are two typical user machines, you might
So, in order to manage a mutual exchange of address between two concurrent
parties, I thought that on each side I would have a producer produce to a
topic that the opposite side was subscribed to. That means that each side
is both a producer and a consumer.
I have the two entities running in
I'm having difficulty - I run the malamute command and it reads up the
malamute.cfg because it announces that it's binding to tcp://*:
I then have the following code with regards to establishing a client
connection to it:
mlm_client_t *frontend_reader = mlm_client_new();
Ah that helped! I think I can make progress now!!
On Sun, Mar 1, 2015 at 1:12 PM, Pieter Hintjens p...@imatix.com wrote:
On Sun, Mar 1, 2015 at 6:57 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
I'm having difficulty - I run the malamute command and it reads up
It's difficult to use the broker, even with the test that is included in
the project tree. I am trying to write my own test just to try and exchange
two strings between pairs of clients...
In particular, I'm confused about this line:
mlm_client_sendto (client, server, something, NULL, 1000,
tutorials later for using Malamute. Take a look at
mlm_tutorial.c and mshell.c perhaps.
On Sat, Feb 28, 2015 at 3:03 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
It's difficult to use the broker, even with the test that is included in
the
project tree. I am trying to write my own
Boy I wish I knew about this sooner, before writing my own broker from
scratch and all that lol I guess a lot of people have written brokers.
On Fri, Feb 27, 2015 at 5:21 AM, Pieter Hintjens p...@imatix.com wrote:
Hi all,
I've moved the Malamute broker project into the ZeroMQ organization on
Wait-is this a broker for which a lot of tests establish that it's already
working, or is this a brand new project?
On Fri, Feb 27, 2015 at 4:50 PM, Pieter Hintjens p...@imatix.com wrote:
There's a specification of the protocol:
https://github.com/zeromq/malamute/blob/master/src/mlm_proto.bnf
So, as per the title, I have two regular machines working underneath
different firewalls, and I need to be able to get each machine talking to
one another. They each talk to a broker that helps them exchange their
address information-but just calling connect on a machine that is
potentially under
Hello,
Did you ever get this resolved?
On Thu, Jan 8, 2015 at 10:29 PM, Gang Liu gangban@gmail.com wrote:
Dear all,
I am new to zeromq. I have some questions about message
queuing broker (msgqueue.c) which use zmq_proxy() forward msgs between
frontend and backend.
problem and alternative solutions like that.
No-one here can do that for you.
On Mon, Jan 12, 2015 at 7:31 AM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
Instead of routing all information through the broker and requiring an
intermediary hop, I'd like to consider an approach where
architectural questions, you won't often get answers here.
On Mon, Jan 12, 2015 at 3:43 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
Oh I apologize. I was just asking if what I was thinking of was possible
with zeromq.
On Mon, Jan 12, 2015 at 4:29 AM, Pieter Hintjens p
Instead of routing all information through the broker and requiring an
intermediary hop, I'd like to consider an approach where the address
information that the req socket parses out on the side that first sends
ready is used in order to manage a simple mutual connection facilitator in
ZMQ...
:
You can change anything you like. Be clear about the semantics
however. The load balancing broker is sending requests for action to
workers. It's not a symmetric flow.
On Fri, Jan 9, 2015 at 6:24 AM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
Can the mechanism of load balancing
:55 AM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
I thought that the tidbits that were being received on the end of the
broker at the front end and backend (I'm talking about how in the guide it
refers to some of what is received as address and empty) were required
because
, 2015 at 8:39 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
Does the broker demonstrated in the manual under:
http://zguide.zeromq.org/page:all#A-Load-Balancing-Message-Broker
demonstrate that
A) when each end makes a connection request, after they link up from the
broker
Does the broker demonstrated in the manual under:
http://zguide.zeromq.org/page:all#A-Load-Balancing-Message-Broker
demonstrate that
A) when each end makes a connection request, after they link up from the
broker, their messages route directly between one another
In this scenario, the messages
I'm using an inproc xpub/xsub setup to manage graceful shutdown of a
multi-threaded system. Here's a diagram:
T1 T2 .
/ |\ / |\
s1 s2 s3 - s1 s2 s3 -
^ /
Adam Miller
kennethadammil...@gmail.com wrote:
I'm using an inproc xpub/xsub setup to manage graceful shutdown of a
multi-threaded system. Here's a diagram:
T1 T2 .
/ |\ / |\
s1 s2 s3
Fixed it-rather than connect and immediately send, I just kept the
publisher and reused it.
On Fri, Jan 2, 2015 at 3:06 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
I added pub-sub tracing, and I can see that when they miss messages,
nothing is sent at all on the network
investigate, except I'm deep into other stuff.
On Tue, Dec 30, 2014 at 7:08 AM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
Well, it doesn't throw the exception when I comment the proxy line out.
And
it does when I uncomment it. That one change back to back is
differentiating
);
}
compiled with:
g++ -std=c++11 test.cpp -lzmq -o test
On Tue, Dec 30, 2014 at 10:42 AM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
Didn't think you were, was just explaining my reasoning. Thank you!
I'll do my best to create a workable, minimal reproduction of the issue
Need to de-reference the pointers; the prototype given for the zmq wrapper
is void * for all parameters... so it doesn't differentiate at compile time.
On Tue, Dec 30, 2014 at 11:42 AM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
This triggers it:
//test.cpp
#include zmq.hpp
There's no really complete example in the guide with source for usage of an
XSUB/XPUB combination that I could find. So I have some questions:
Suppose the following is how I set up the XPub XSub sockets:
_xpub = new zmq::socket_t(_context, ZMQ_XPUB);
_xpub-bind(inproc://killpub);
_xsub =
On Mon, Dec 29, 2014 at 11:03 AM, Pieter Hintjens p...@imatix.com wrote:
On Mon, Dec 29, 2014 at 4:51 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
Now my questions: for xpub and xsub do you do the following:
pub-setsockopt(ZMQ_SUBSCRIBE, , 0);
No, this works only on SUB
.
What do I do to get it to work?
On Thu, Dec 25, 2014 at 10:51 AM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
I have a dockerized ZMQ instance where I am trying to develop an app. I
have duplicate source both inside the docker instance and at the host
level.
I can compile
, 2014 at 9:25 AM, Pieter Hintjens p...@imatix.com wrote:
Is this using the same version of ZeroMQ in both cases?
On Mon, Dec 29, 2014 at 1:02 AM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
I found out that doing a socket connection on an inproc that hasn't been
bound to yet
If no publishers are currently connected to an xpub/xsub setup, but
subscribers start connecting and setting up subscriptions, will the
publishers get the subscriptions when they finally do connect? Or is the
XPUB/XSUB setup something that holds onto subscription information
internally somehow?
On Mon, Dec 29, 2014 at 1:34 PM, Dmitri Toubelis
dmitri.toube...@alkeron.com wrote:
I have a dockerized ZMQ instance where I am trying to develop an app. I
have duplicate source both inside the docker instance and at the host
level. I can compile both inside and out with duplicate compiler
, _xpub, NULL); //throws here
On Mon, Dec 29, 2014 at 5:08 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
On Mon, Dec 29, 2014 at 11:03 AM, Pieter Hintjens p...@imatix.com wrote:
On Mon, Dec 29, 2014 at 4:51 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
Now my
, Dec 30, 2014 at 12:24 AM, Pieter Hintjens p...@imatix.com wrote:
There is nothing in the proxy code that would reject a XPUB or XSUB socket.
On Mon, Dec 29, 2014 at 11:50 PM, Kenneth Adam Miller
kennethadammil...@gmail.com wrote:
On Mon, Dec 29, 2014 at 3:23 PM, Pieter Hintjens p
1 - 100 of 142 matches
Mail list logo