Perl problem

[Paul Howarth]

On Mon, 2005-02-07 at 16:49 -0800, Rick Stevens wrote:
> Bas Bosman wrote:
> > Hello I have a really strange problem which I can’t trace.
> > 
> > After my fedora core 2 server has been running for a while the load goes 
> > up above 1.XX without a clear reason.
> > 
> > When I run the TOP command I get this:
> > 
> >  
> > 
> > Tasks: 179 total,   3 running, 144 sleeping,   0 stopped,  32 zombie
> > 
> > Cpu(s): 83.6% us, 15.8% sy,  0.0% ni,  0.0% id,  0.0% wa,  0.7% hi,  0.0% si
> > 
> > Mem:    256024k total,   252912k used,     3112k free,    43184k buffers
> > 
> > Swap:  1044216k total,   286548k used,   757668k free,    14916k cached
> > 
> >  
> > 
> >   PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
> > 
> >  1574 apache    25   0  8588 3360 1880 R 88.5  1.3 903:27.45 perl
> > 
> >  
> > 
> > So I guess 32 zombies is a bit much;)
> > 
> >  
> > 
> > Anyone has any idea how I can trace this problem?
> 
> Do an lsof or gdb on the zombies and see what parent they were spawned
> off of.

Given that the perl process is running as user "apache", it looks like a
CGI or mod-perl script on the webserver is the problem.

Paul.
-- 
Paul Howarth <paul at city-fan.org>