durusmail: quixote-users: SCGI vs. mod_python
RFC: SCGI graceful restart
2003-02-20
SCGI vs. mod_python (was RFC: SCGI graceful restart)
2003-02-20
SCGI vs. mod_python (was RFC: SCGI graceful restart)
SCGI vs. mod_python (was RFC: SCGI graceful restart)
2003-02-20
2003-02-20
SCGI vs. mod_python
2003-02-21
2003-02-21
2003-02-21
2003-02-21
2003-02-21
2003-03-01
2003-02-21
2003-02-21
2003-02-21
2003-02-21
2003-02-21
2003-02-21
Re: RFC: SCGI graceful restart
2003-02-20
SCGI vs. mod_python
akuchlin@mems-exchange.org
2003-02-21
On Thu, Feb 20, 2003 at 03:38:53PM -0700, Jonathan Corbet wrote:
>[And yes, before somebody points it out, there is clearly a memory leak in
>there somewhere.  Concentrating the traffic on one process has brought that
>out in an very clear way.  I'm not quite sure how I'm going to track that
>down... you're not supposed to have to deal with memory leaks in Python...]

If you have objects with __del__ methods, maybe you're running into
uncollectable cycles.  Assuming you're running a 2.x Python, you can
use gc.set_debug() to turn on leak detection, and then look in
gc.garbage to see if there are any uncollectable cycles building up.
You can thank Neil for that capability, too. :)

I'd also be suspicious of any extension modules you're using.  Maybe
there's a refcounting bug in the PostgreSQL module or some place like
that.  Or maybe there's a cache somewhere that never gets cleaned out?

--amk                                                    (www.amk.ca)
Tourist, Rincewind decided, meant "idiot".
      -- Terry Pratchett, _The Colour of Magic_

reply