问题说明
我们的应用程序使用Java绑定ZMQ，使用Pub-Sub模型。子程序间歇性但频繁地崩溃，错误跟踪（在崩溃期间生成的核心转储中看到）指向ZMQ崩溃。
问题出现在我们的生产中，负载很高。
环境
版本：jzmq（3.1.0） 操作系统：Linux
最小测试代码/重现问题的步骤
将jzmq（3.1.0）与Pub-Sub模型一起使用。继续从Pub发送消息一段时间。
实际结果是什么？ （包括断言消息和调用堆栈，如果适用）
ZMQ崩溃并发出以下消息。
核心转储痕迹：
崩溃 1 ：
0 0x00002b0169da7fc5 in raise () from /lib64/libc.so.6
1 0x00002b0169da9a70 in abort () from /lib64/libc.so.6
2 0x00002b0190f1dce9 in zmq::zmq_abort (errmsg_=0x4409 <Address 0x4409 out of bounds>) at err.cpp:76
3 0x00002b0190f344c8 in zmq::stream_engine_t::write (this=, data_=, size_=) at stream_engine.cpp:517
4 0x00002b0190f34709 in zmq::stream_engine_t::out_event (this=0x1865cbf0) at stream_engine.cpp:275
5 0x00002b0190f1d3ba in zmq::epoll_t::loop (this=0x2b017f34d490) at epoll.cpp:158
6 0x00002b0190f3b28c in thread_routine (arg_=0x2b017f34d500) at thread.cpp:83
7 0x00002b016974783d in start_thread () from /lib64/libpthread.so.0
8 0x00002b0169e4d19d in clone () from /lib64/libc.so.6
崩溃 2 ：
0 0x00002b121f0444bb in memcpy () from /lib64/libc.so.6
1 0x00002b12420f5a02 in zmq::encoder_base_tzmq::v1_encoder_t::get_data (this=0x1cdd1110, data_=0x1cafc958, size_=0x1cafc960, offset_=0x0) at encoder.hpp:117
2 0x00002b12420e86f2 in zmq::stream_engine_t::out_event (this=0x1cafc910) at stream_engine.cpp:261
3 0x00002b12420e1202 in zmq::session_base_t::read_activated (this=0x1cafcb50, pipe_=0x1be54720) at session_base.cpp:264
4 0x00002b12420d27e9 in zmq::io_thread_t::in_event (this=) at io_thread.cpp:75
5 0x00002b12420d13da in zmq::epoll_t::loop (this=0x2b1234fea0b0) at epoll.cpp:162
6 0x00002b12420ef28c in thread_routine (arg_=0x2b1234fea120) at thread.cpp:83
7 0x00002b121e99883d in start_thread () from /lib64/libpthread.so.0
8 0x00002b121f09e19d in clone () from /lib64/libc.so.6
(gdb) bt
崩溃 3 ：
0 0x00002b10f9bcdfc5 in raise () from /lib64/libc.so.6
1 0x00002b10f9bcfa70 in abort () from /lib64/libc.so.6
2 0x00002b1124d92ce9 in zmq::zmq_abort (errmsg_=0x3d41 <Address 0x3d41 out of bounds>) at err.cpp:76
3 0x00002b1124da94c8 in zmq::stream_engine_t::write (this=, data_=, size_=) at stream_engine.cpp:517
4 0x00002b1124da9709 in zmq::stream_engine_t::out_event (this=0x8adb440) at stream_engine.cpp:275
5 0x00002b1124d923ba in zmq::epoll_t::loop (this=0x2b111c2cf500) at epoll.cpp:158
6 0x00002b1124db028c in thread_routine (arg_=0x2b111c2cf570) at thread.cpp:83
7 0x00002b10f956d83d in start_thread () from /lib64/libpthread.so.0
8 0x00002b10f9c7319d in clone () from /lib64/libc.so.6
预期结果是什么？
它应该继续运行，没有任何崩溃。
我们是否应该转移到基于libzmq 4.1.7的Jeromq（https://github.com/zeromq/jeromq） - 因为最新版本会修复这些问题吗？
你能建议/帮我解决这个问题吗？