Comment #2 on issue 154 by bananastalktome: pid file out of sync (created before socket binding)
http://code.google.com/p/memcached/issues/detail?id=154
Although not the OP, I agree with them that this is an issue. This has thrown me off several times, as I have run a script to start memcached when it is already started and, from then on, been unable to stop it through that script given the pidfile change. It would make sense to me that if memcached successfully starts, it should write to the pidfile. If it does not start successfully, it should not. Given a different port, it is up to the user to specify a different pidfile. If they do not do so, it is user error and not up to memcached to determine. However, if the daemon fails to start, there should not be a pidfile written.
