So, it looks like we hit the condition where we have the group leader in zombie state and one of its thread (last one) is still in running state. Due to this, there is no SIGCHLD from this child process to the parent process (which is PGHD). 14130 14130 0000000000000000 0000000000000000 0000000000001000 00000003800044ac Zl 00:00:00 exit [ip-fastpath-epb] <defunct> ip-fastpath-epb <defunct> 14130 16430 0000000000000000 0000000000000000 0000000000001000 00000003800044ac Rl 00:00:22 - [ip-fastpath-epb] <defunct> CelloIpsockiEt_ The problem here is………….. As part of processor restart sequence, PGHD (parent process for all the application programs) sends SIGKILL to every application spawned by it and waits to receive SIGCHLD from all of its childs. If all the programs are not terminated in 10 secs, then Board-Manager (a thread inside BOS program) will send SIGABRT to PGHD which will result in dump generation and target reboot (i.e., a crash). Could you please tell us why the last thread (CelloIpsockiEt_) is still running despite of receiving SIGKILL by the application ? Is it possible that this thread is executing some routine which is task uninterruptible and is not responding to SIGKILL ? If so, how do we know if the thread is in this state ?
