Shared-memory buffer memory-leak in /dev/shm/

[ciandonovan]

When an rmw_zenoh node crashes or otherwise doesn't gracefully shutdown, the associated shmem_XXXXXX file in /dev/shm/ isn't cleaned up. Nor is it cleaned up the next time a node instance launches.

This leads to persistent useless memory usage that won't be resolved until the host reboots or the files are manually deleted. If there is an issue with a node that causes it to crash often and it is automatically restarted by a service manager, this could lead to memory exhaustion which Linux does not deal with well, causing system lockups and freezes.