Details
-
Bug
-
Status: Closed
-
Major
-
Resolution: Cannot Reproduce
-
3.8.4
-
None
-
None
-
CentOS 6 (32- and 64-bit), CentOS 7 (64-bit)
Description
Occasionally, but not always, during the course of closing the CMS connection, a segmentation fault occurs. This issue is eerily similar to that described in AMQCPP-534. However, the issue in AMQCPP-534 was supposedly fixed for ActiveMQ CMS 3.8.4, and this just so happens to be the version I'm using. I also witnessed the same issue when using version 3.8.3.
The client application I'm a running is essentially opening and closing connections at a rapid pace (i.e. a stress test). A transport listener has been set up, so as to receive notifications as to when the connection has been interrupted (broken). Whether I call upon ActiveMQConnection::removeTransportListener() or not, the segmentation fault occurs.
Here's the stack trace:
#0 0x00007f74aa4373af in activemq::transport::TransportFilter::onCommand(decaf::lang::Pointer<activemq::commands::Command, decaf::util::concurrent::atomic::AtomicRefCounter>) ()
from /home/nightrun/gmsec/SUPPORT/activemq384/lib/libactivemq-cpp.so.18
#1 0x00007f74aa435b5c in activemq::transport::IOTransport::fire(decaf::lang::Pointer<activemq::commands::Command, decaf::util::concurrent::atomic::AtomicRefCounter>) ()
from /home/nightrun/gmsec/SUPPORT/activemq384/lib/libactivemq-cpp.so.18
#2 0x00007f74aa436da1 in activemq::transport::IOTransport::run() () from /home/nightrun/gmsec/SUPPORT/activemq384/lib/libactivemq-cpp.so.18
#3 0x00007f74aa57ab82 in (anonymous namespace)::runCallback(void*) () from /home/nightrun/gmsec/SUPPORT/activemq384/lib/libactivemq-cpp.so.18
#4 0x00007f74aa57b48c in (anonymous namespace)::threadEntryMethod(void*) () from /home/nightrun/gmsec/SUPPORT/activemq384/lib/libactivemq-cpp.so.18
#5 0x00007f74aba32df3 in start_thread () from /lib64/libpthread.so.0
#6 0x00007f74aaf413dd in clone () from /lib64/libc.so.6
If additional information is required, please let me know.