I'm new to OSX. For my httpd sandbox, I notice "apachectl stop" has to
forcibly kill children.  When I caught one with pstack, the listener
thread was the only thing running (other than the main thread trying
to join_workers)


  thread #2
    frame #0: 0x00007fffc2baf0c2 libsystem_kernel.dylib`__accept + 10
    frame #1: 0x000000010a4c3e8c
libapr-1.0.dylib`apr_socket_accept(new=0x0000700005116e50,
sock=0x00007fc03e01ad50, connection_context=0x00007fc03e053c28) at
sockets.c:271 [opt]
    frame #2: 0x000000010a35a607
httpd`ap_unixd_accept(accepted=0x0000700005116ea0,
lr=0x00007fc03e01ad08, ptrans=<unavailable>) at unixd.c:302 [opt]
    frame #3: 0x000000010a7bd349
mod_mpm_worker.so`listener_thread(thd=0x00007fc0400051d0,
dummy=<unavailable>) at worker.c:693 [opt]
    frame #4: 0x00007fffc2c9a93b libsystem_pthread.dylib`_pthread_body + 180
    frame #5: 0x00007fffc2c9a887 libsystem_pthread.dylib`_pthread_start + 286
    frame #6: 0x00007fffc2c9a08d libsystem_pthread.dylib`thread_start + 13

It happens with both 1 or multiple listening sockets.  It doesn't
happen in my 2.4 sandbox.

Does anyone else see this (or not?) . The obvious symptom is:
[Wed Jan 10 19:11:21.204862 2018] [core:warn] [pid 68986:tid
140736609248192] AH00045: child process 68988 still did not exit,
sending a SIGTERM

-- 
Eric Covener
cove...@gmail.com

Reply via email to