Benjamin,
Could you provide a minimal test case that reproduces the problem, and
perhaps file an issue on the github tracker, thanks.
-Pieter
On Tue, Aug 10, 2010 at 8:34 AM, MinRK benjami...@gmail.com wrote:
Hello,
I'm using ZMQ devices for parallel computing in IPython. One of our devices
Hello!
I need a simple request/response interaction, though I am not happy with
ZMQ_REQ/ZMQ_REP pattern.
The main reason is it is unclear what I supposed to do if a response from a
server is delayed or lost.
Do we have an example covers e.g. reopening the socket in this case?
Is it reasonable
On Tue, Aug 10, 2010 at 2:30 AM, Matt Weinstein mattweinst...@gmail.com wrote:
What I'd prefer is a zmq_abort(socket) that kills the most recent
train of packets, as long as a SNDMORE == 0 packet has not been sent.
These are likely to be sitting in a ypipe somewhere along the chain.
I suspect
Hi,
You are running into a classic problem and yes, the solution is to use
XREQ/XREP. However the necessary message format is not documented
yet. You can look at the code for REQ and see how it constructs the
request, then you can do the same in your client before calling XREQ.
I hope to have
Jon,
We know how to upstream the subscriptions from SUB to PUB but it
involves a fair chunk of work internally to 0MQ.
-Pieter
On Tue, Aug 10, 2010 at 11:07 AM, j...@totient.demon.co.uk wrote:
Hi
Does anyone have any pointers/design sketch on what needs to be changed to
get the pub/sub
Pieter,
p...@imatix.com said:
Attached is a change I'd like to make to de-emphasize non-portable
device names and point people rather towards binding to * for
starters.
Reason: to reduce avoidable mistakes.
This change is fine, but the patch you sent also includes an extraneous
change to
p...@imatix.com said:
On Tue, Aug 10, 2010 at 12:28 PM, Martin Lucina m...@kotelna.sk wrote:
This change is fine, but the patch you sent also includes an extraneous
change to builds/redhat/zeromq.spec. I presume that would be a separate
commit.
Ah, sorry, the lack of a robust procedure
COBOL on AIX... eyes glaze over, PTSD kicks in...
Oops. Sorry about that; I am stuck with this POS.
Gonzalo, sounds like a very interesting project. You'll have seen the
discussion on reliability at
http://www.zeromq.org/blog:requirements-for-reliability?
Yes, and I am thinking
Why not port the COBOL to a real VMS platform instead of Charon VAX or
whatever you are using? What kind of application is this and are the
sources still available?
It was this way initially, and we stepped away from hardware-based VMS,
it was deemed too risky. Not that Charon is less risky,
On Tue, Aug 10, 2010 at 2:17 PM, gonzalo diethelm gdieth...@dcv.cl wrote:
COBOL on AIX... eyes glaze over, PTSD kicks in...
Oops. Sorry about that; I am stuck with this POS.
Oh, not that, just that in the 80's I must have written half a million
lines of Cobol.
Yes, and I am thinking pretty
Yes. This along with net-neutrality concerns, are very troublesome.
Ever since I read Martin's post, the word federation has been stuck in my
mind. This one word accurately describes part of the vision I have for my
data laboratory project.
And naturally, thanks to 0mq it makes a federation of
Hello!
I am observing rather strange zmq_poll behavior.
It is triggered by timeout and works fine if -1.
The client looks like
==
zmq::context_t ctx (1);
zmq::socket_t s(ctx, ZMQ_REQ);
// Connect to the server.
s.connect (tcp://localhost:23001);
for (int i = 0; i != 20;
Hi,
ilej...@narod.ru said:
I am observing rather strange zmq_poll behavior.
It is triggered by timeout and works fine if -1.
This is a known issue with the current zmq_poll() implementation. The
zmq_poll(3) documentation mentions this, albeit briefly:
Important
The zmq_poll()
On Tue, Aug 10, 2010 at 3:17 PM, Peter Alexander vel.ac...@gmail.com wrote:
Sorry for this ot post, but I just had to say thanks for mentioning ditaa.
I've been looking for a little gem like this for a long time and don't know
why I had not found it myself earlier.
Yes, it's elegant. For
Does this have security implications, or is it _just_ for performance?
If for performance, you can use hints to the upstream on a pub/sub
channel from the endpoints to the upstream servers.
There are lots of simple protocols, like using a GUID per client and
sending an update or a delta to
matt_weinst...@yahoo.com said:
This is the same problem as pthread condition variables have, e.g.:
http://www.justsoftwaresolutions.co.uk/threading/condition-variable-spurious-wakes.html
Just think of the revents as condition variables to check...
Does someone want to write a quick
ZMQ_FD and ZMQ_EVENTS appear to return the current status of a socket,
am I right?
I looked at the poll() code but it looks pretty vanilla, list of
timers ... except for the cross-thread counter, which would be nice,
is that what I'm seeing ...?
On Aug 10, 2010, at 10:17 AM, Martin Lucina
Martin, Matt,
thank you very much for clarification
and for speedy answering.
To be honest I don't fill why spurious wakeups are unavoidable in 0mq ...
though it doesn't matter.
Here is a mockup of client application
==
// One I/O thread in the thread pool will do.
zmq::context_t ctx
Hi Benjamin,
Thanks for posting the test case, it makes it clear.
At the least the documentation for zmq_device(3) should state what
socket types are safe to use. Do any other socket types prepend an
identity apart from XREP?
-Pieter
On Tue, Aug 10, 2010 at 10:25 AM, Pieter Hintjens
COBOL on AIX... eyes glaze over, PTSD kicks in...
Oops. Sorry about that; I am stuck with this POS.
Oh, not that, just that in the 80's I must have written half a million
lines of Cobol.
Poor you.
So you want a reliable pipeline. What are your reasons for not using
xreq/xrep and
Basically -- reverse the normal server paradigm, and use an XREP
socket on the device side, then have each server send a single message
when it wakes up on its REQ socket. This message identifies the
server to the device side by UUID. The server will be back to
alternative recv-send
On Tue, Aug 10, 2010 at 7:33 PM, gonzalo diethelm gdieth...@dcv.cl wrote:
What would I gain from xreq/xrep, that could not be done with push/pull?
Mind you, I have not yet used xreq/xrep, so my question is out of sheer
ignorance.
With push/pull there's no reply, so no way to tell the sender
XREQ/XREP are used if your requester or servicer needs to wait for a
response or request, i.e. if you're in a threaded model.
If you're in an asynchronous model, you're going to be doing all the
housekeeping yourself and anything will work as long as it transports
packets.
On Aug 10, 2010,
On Tue, Aug 10, 2010 at 10:14, Pieter Hintjens p...@imatix.com wrote:
Hi Benjamin,
Thanks for posting the test case, it makes it clear.
At the least the documentation for zmq_device(3) should state what
socket types are safe to use. Do any other socket types prepend an
identity apart from
On 11 August 2010 03:00, Dana Leonard dleo...@gmail.com wrote:
Is this a standard behavior for IPC sockets on Unix or does the socket
need additional configuration to allow different users to use it? I have
tried changing the permissions on the socket to 777 to no avail. Even root
cannot
25 matches
Mail list logo
