Using force kill when we want to stop a container should only be
called after we left some time to make sure nothing has been killing
the container process. Indeed, signals can sometime take a few seconds
to terminate a process. In this case, we don't want to SIGKILL this
process, otherwise it would return with the exit code 137. Instead, we
want to leave it some time (10 seconds) so that it can properly end.
After those 10 seconds, we consider the previous signal didn't get the
expected effect and we force kill the container process. Indeed, we
still want to try to kill the container.
sameo
commented
6 years ago 
Using force kill when we want to stop a container should only be called after we left some time to make sure nothing has been killing the container process. Indeed, signals can sometime take a few seconds to terminate a process. In this case, we don't want to SIGKILL this process, otherwise it would return with the exit code 137. Instead, we want to leave it some time (10 seconds) so that it can properly end.
After those 10 seconds, we consider the previous signal didn't get the expected effect and we force kill the container process. Indeed, we still want to try to kill the container.
Fixes #606
Depends-on: github.com/clearcontainers/runtime#1003
Signed-off-by: Sebastien Boeuf sebastien.boeuf@intel.com