Re: Significant memory leak in qpid-dispatch 1.0.0
I'm good with that list. DISPATCH-881 is already on the 1.0.x branch. I've taken the liberty to create pull requests for the backports of DISPATCH-914, 889, and 882. -K On Mon, Feb 12, 2018 at 11:36 AM, Ted Rosswrote: > I've moved eleven jiras into 1.0.1 for consideration. These are all > resolved and are candidates for backport into 1.0.1. > > DISPATCH-874 - unable to load .json or .woff2 files from local > file system from http port > DISPATCH-881 - Inbound pre-settled messages causes memory leak of > deliveries > DISPATCH-882 - router buffers messages for slow presettled receiver > DISPATCH-883 - Router crashes when it processes management request > for connections > DISPATCH-887 - Dispatch reestablishes connection inspite of > deleting the connector > DISPATCH-889 - linkRoute patterns beginning with #/string match > substrings after the / > DISPATCH-895 - qpid-dispatch crashes with a SEGFAULT in libqpid-proton > DISPATCH-900 - Memory leak when repeatedly opening and closing connections > DISPATCH-908 - Router loses dispositions over receive link on > qpid-interop-test 2-node test > DISPATCH-914 - qd_connector_t leaks mutexes > DISPATCH-920 - Enabled policy blocks inter-router links > > -Ted > > On Fri, Feb 9, 2018 at 8:03 AM, Ted Ross wrote: >> I agree. Let's do a 1.0.1 with this fix. >> >> -Ted >> >> On Wed, Feb 7, 2018 at 2:54 PM, Ken Giusti wrote: >>> Folks, >>> >>> Can we get a fix release for qpid-dispatch containing the fix to >>> https://issues.apache.org/jira/browse/DISPATCH-881 ? >>> >>> This fixes a per-delivery leak that's present in the 1.0.0 release. This >>> is a per-message leak that causes the router's memory use to climb linear >>> to the amount of traffic handled. >>> >>> thanks >>> >>> -- >>> -K > > - > To unsubscribe, e-mail: users-unsubscr...@qpid.apache.org > For additional commands, e-mail: users-h...@qpid.apache.org > -- -K - To unsubscribe, e-mail: users-unsubscr...@qpid.apache.org For additional commands, e-mail: users-h...@qpid.apache.org
Temporary queues deletion
Greetings! I have a problem with autodeleted queues. I have a service that reads amqp messages and sends replies using reply_to value. Client that sends the message generates temporary queue to use for replies using "/#" syntax. Client sends the message, waits for reply and terminates. But temporary queue, created by client, still remains in the broker until the other end (my service) is restarted. Here's how this queue is shown by qpid-stat after client has terminated: queue dur autoDel excl msg msgIn msgOut bytes bytesIn bytesOut cons bind === 821 857-349.0 Y 0 1 1 0257 257 01 Queue is marked as auto delete, number of connections is 0, but bind is 1. I guess this is the reason why it is not deleted. Is there a way to request messenger to unbind the queue as soon as the message is sent? I am using pn_messenger_* interface of proton library. Best regards, -- Michael Ivanov - To unsubscribe, e-mail: users-unsubscr...@qpid.apache.org For additional commands, e-mail: users-h...@qpid.apache.org
Re: Qpid proton cpp
Possible you've found a bug, but also possible you're hitting some C++ template type-mismatch potholes. C++ compilers often give incomprehensible error messages for such problems. Post a code snip that illustrates the problem with the compiler version you're using, and the error message you see. If you can choose your compiler for development I recommend clang++ for better than average error messages. My employer ships software built with gcc, but I sometimes switch to clang just to figure out what gcc is complaining about. On Wed, Feb 7, 2018 at 3:51 AM, Baptiste Robertwrote: > Hi everybody, > > I'm currently working with the qpid proton cpp API (0.18 version) with an > old gcc without c++11 support. I can compile it and use the examples, but > I'm building a multi-threaded client, where I'm trying to create new sender > or sending message asynchronously. So in the examples I found the > working_queue principles, but I'm facing the following difficulty : > How to create a work for the work queue properly ? > > I found this function proton::make_work, but when I'm trying to call it for > sending a message like that : proton::make_work(::sender::send, > _sender, m) it doesn't compile.. > > I'm missing something, on the asynchronous utilization ? > > Have a good day, > > -- > Baptiste Robert >
Re: Temporary queues deletion
On 15/02/18 20:40, Michael Ivanov wrote: Greetings! I have a problem with autodeleted queues. I have a service that reads amqp messages and sends replies using reply_to value. Client that sends the message generates temporary queue to use for replies using "/#" syntax. Client sends the message, waits for reply and terminates. But temporary queue, created by client, still remains in the broker until the other end (my service) is restarted. Here's how this queue is shown by qpid-stat after client has terminated: queue dur autoDel excl msg msgIn msgOut bytes bytesIn bytesOut cons bind === 821 857-349.0 Y 0 1 1 0257 257 01 Queue is marked as auto delete, number of connections is 0, but bind is 1. I guess this is the reason why it is not deleted. Do you have a simple reproducer? - To unsubscribe, e-mail: users-unsubscr...@qpid.apache.org For additional commands, e-mail: users-h...@qpid.apache.org