<Original message>
In the EE department here at TAMU we have two SparcServer 1000's running
Solaris 2.4. Today around 12:00 on one of them, we began getting about 1
zombie (<defunct>) process about every two seconds. Then, about an hour
later, the same thing began happening on our other server. By about 2:30
one server was completely useless (i.e. no more processes) and the other
was well on its way. The zombie processes belong to everybody from root to
people not even logged in. We could not kill the processes using kill -9.
So my question is, "What could have caused this and what can we do?"
We ended up having to reboot both machines. This is not the solution we
wanted, but it is the only solution we could come to. We can't really
understand the problem becuase they are two distinct machines, yet both
suferred the same fate. Any help -killing these processes or identifying
the cause- would be appreciated.
SunOS foo 5.4 Generic_Patch sun4d sparc
--------------------
Those who responded:
From: Dave Mitchell <D.Mitchell@dcs.shef.ac.uk>
From: roy@mss.com (Rahul Roy)
From: Casper Dik <casper@Holland.Sun.COM>
From: perryh@pluto.rain.com (Perry Hutchison)
From: raoul@MIT.EDU
Well, even though I did not include the info I really should have, I still
got some very helpful responses. I should have mentioned the fact that the
parent process for all of the zombies was init. This probably would have
been most helpful and I apologize. Anyway, the following is what I learned:
(1) Don't ever use kill and zombie in the same sentence unless you want a
spanking (zombies are already dead and cannot be killed. You must kill the
parent.)
(2) try using kernel patch 101945-23 instead of the generic patch.
(3) check for possible intrusion due to the various ownership of the
zombies and because the problem manifested itself on two distinct servers
within a short period of time.
Thanks to all that responded and I will try the patch and check the wrapper
logs.
Thanks,
Jarod Jenson
-------------------------------------------------------------------
|Texas A&M University jarodj@tamu.edu |
|System Integrator/Administrator |
-------------------------------------------------------------------
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:10:28 CDT