[Zope] Re: Zope hanging (poss. threads-related)

[Marcus Collins]

Hi,

First of all, thanks to all the folk who have contributed valuable
information to this thread!

I did originally increase the pool_size to 20 connections (I'm using
ZMySQLDA 1.1.3, patched with MySQLdb 0.1.0; MySQL 3.22.29 on remote host),
so I was running Zope with the same number of threads as there were
connections, at the time of the hanging. My testing of Zope 2.1.6 is under
these same conditions.

I think the message of Jim Fulton's to which Michel refers is
http://lists.zope.org/pipermail/zope/1999-December/015455.html

Lately, I've seen reports of Zope restarting when manage_* is used; just to
clarify, in my case Zope is not restarting: a single thread is hanging
(deadlocked for some reason, perhaps?). This occurs in manage_* as well as
on our live site, but only a single thread is affected at a time (other
users could still log in while this thread was in its quiescent state.

Michel Pelletier wrote:
> I suspect this problem *might* be unrelated to the threadlock 
> discussed so far, in the case of the reported lock, 2 or more 
> threads cause instability.  In your case, you report 4 is stable.

Are you referring to the global python threadlock, or the ZODB lock to
prevent conflict errors? Writes to ZODB are particularly infrequent, only
occurring when I import changes.

I'm trying to procure another similar machine for testing; unfortunately my
personal machine's hard drive just packed up, so that's no good...  If you
have *any* suggestions as to where to start debugging (Pavlos Christoforou
mentioned the async loop -- is that ZServer/medusa/asyncore.py?), please let
me know. I'll start by slowly upping NUMBER_OF_THREADS until something
screams. I'm really keen to get to the bottom of this... It's a pity we're
about eight hours ahead of you down here in South Africa...

Off to try grokking ZServer/*,

-- Marcus