> I found out today that I have two zombie sshd processes on a busy server
> (dozen or so users over ssh, many other services), and I can't kill
them.
>
> sshd 38653 0.0 0.0 0 0 ?? Z Mon08AM 0:00.03
<defunct>
> sshd 75851 0.0 0.0 0 0 ?? Z 7:33PM 0:00.08
<defunct>
>
> kill -9 doesn't complain, but doesn't do the job either.
>
> Is there a way to find out what has happened and why does the situation
> occur? (I can't reboot the server for testing)
You can't kill them because they are already dead. They
are just holding state so that the parent process can know
how they died. Once the parent process wait()'s on them
(or the parent dies) they will disappear.
This is a FAQ that is over 20 years old.
Mark
--
Mark Andrews, ISC
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: +61 2 9871 4742 INTERNET: Mark_Andrews@isc.org