In the SharedMemory example, sometimes the publisher and subscriber topic cannot be matched

[foreverkid]

Is there an already existing issue for this?

• I have searched the existing issues

Expected behavior

When a publisher and subscriber have the same topic, they should match every time.

Current behavior

Sometimes publisher and subscriber cannot match and communicate

Steps to reproduce

• Compile the QNX version of example fast-dds/examples/C++/DDS/HelloWorldExampleSharedMem
• Open two terminals in the qnx system to run ./DDSHelloWorldExampleSharedMem publisher and ./DDSHelloWorldExampleSharedMem subscriber respectively
• Keep restarting the QNX system and run the previous step
• Sometimes publisher and subscriber cannot matched and communicate

Fast DDS version/commit

2.6.X

Platform/Architecture

Other. Please specify in Additional context section.

Transport layer

UDPv4, Shared Memory Transport (SHM), Zero copy

Additional context

Platform: QNX7

I turned on the debug function of dds and compared the difference between connected and unconnected. Before printing out the "Subscriber matched." message, they differ in the following

Connected:
01.0f.ce.6a.1e.90.de.ae.01.00.00.00|0.0.1.c1 at MTTLoc: {UNICAST:[SHM:[]:7414]} DefLoc:{UNICAST:[SHM:[]:7415]} -> Function onNewCacheChangeAdded
[RTPS_TRANSPORT_SHM Info] (ID:5) Opening fastrtps_port7414 -> Function open_port_internal
[RTPS_TRANSPORT_SHM Info] (ID:5) Port 7414 (0d449125890005a2) OpenedWrite -> Function open_port_internal

Disconnected:
01.0f.ea.21.1e.10.f4.ef.01.00.00.00|0.0.1.c1 at MTTLoc: {} DefLoc:{} -> Function onNewCacheChangeAdded

XML configuration file

No response

Relevant log output

No response

Network traffic capture

No response

[JLBuenoLopez]

Thanks for your contribution @foreverkid,

QNX is not a Tier 1 platform. Currently is considered Tier 3 supported and consequently, as per Fast DDS CONTRIBUTING guidelines, I am moving this issue to the Unofficial support discussion forum.

Recently, some long standing Shared Memory issues where fixed related to ports stuck due to non-cleanly closed applications. These fixes have been backported to 2.6.x but have not been merged yet (#3695, #3699, #3711). You might try with a branch taking those fixes and see if your issue is still happening.