orphaned processes on 2650's running RH 7.3

Brian K. Jones jonesy at CS.Princeton.EDU
Fri Jan 31 11:49:00 CST 2003


Hi,

I'm trying to find out all I can about why the following occurs and how
I might be able to avoid it.  

As an admin, I'm also interested in --stable and reliable-- ways of
cleaning this mess up in an automated way - but first and foremost is
the WHY.

I have 3 identically configured PE2650's running RH 7.3 with a
customized 2.4.20 kernel (which rocks, by the way).  These are general
purpose servers and can be used for just about anything.  

What a lot of users do is 'ssh -X' to the box, and either shoot back an
xterm, from which they launch vim or emacs, or shoot back gvim or
xemacs.  

When they're done editing, they just click the 'x' in the top right
corner of the window (usually) to close the window instead of exiting
the editor, and then exiting the xterm or whatever.  

This leaves me with lots of orphaned *vim and *emacs processes (owned by
PPID '1') which wind up eating away at the cpu cycles.  At any given
time, leaving these processes orphaned for 2 days regularly drives 'load
average' up over 5.x.  Killing them immediately drops the load back down
to 0.01 or so.  

This seems to only really happen with these two apps.  Certainly people
use the same tactic with windows running 'pine' or 'mutt' as well, but
these never top the list or even cause a stir.  

The vim and emacs versions came stock with 7.3.  Testing with 8.0 shows
the same problem.  

Any advice welcome and hereby solicited. 
brian.  




More information about the Linux-PowerEdge mailing list