Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Defunct processes

Re: Defunct processes

From: Billy Verreynne <vslabs_at_onwe.co.za>
Date: 25 Nov 2003 23:04:40 -0800
Message-ID: <1a75df45.0311252304.299003e5@posting.google.com>


joel-garry_at_home.com (Joel Garry) wrote

> Sounds like you answered your own question.

Is that good or bad? :-)

> I'd guess running out of swap is preventing the children from saying "bye-bye."

Yeah.. possibly (do not know enough about kernel internals and how signals are using user mem and swap).

> I take it you don't have a huge file-based secondary swap for these users?

Me not sysadmin on any of the HP boxes we use.  

> Once they are defunct, I don't think PMON can do anything a kill -9
> can't do. Which is nothing.

Yes - that's what I understand. A defunct process does not exist as a process at all. It is simply taking up a slot in the list of processes struct because it has an oustanding signal not yet handled by the parent. A kill -9 on it will do nothing.

> http://ftp.cerias.purdue.edu/pub/tools/unix/sysutils/lsof/ might help,
> especially if you are being misdirected to look at pmon.

Thanks. Did not know of this.

There's a fuser command on HP-UX that is similar (it takes file/dev as param). I will have a look. The defunct's PPID is PMON though, thus PMON is the parent and, according to what is said about zombies, not processing (or picking up) the defunct child's signal.

> Oh, of course, check patches.

Huh? Patches! What are those? ;-)

AFAIK the the HP-UX 11.00 patch level matches 9i's requirements.

--
Billy
Received on Wed Nov 26 2003 - 01:04:40 CST

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US