self.__block.acquire(blocking) hanging Plone instance using LDAP

[Troy D. <tr...@me...>]

Hello Everyone,

I've been trying to resolve a deadlock situation on a Plone instance  
that uses ldap for authentication, and thus python-ldap. The ldap  
server is on a Fedora Core 2 machine, openldap-servers-2.1.29-1 rpm.  
It has an SSL cert from GoDaddy, and LDAPUserFolder is configured to  
use SSL.

The system running Plone is Fedora Core 5, python-2.4.3-8.FC5 rpm,  
and up until tonight I had python-ldap rpm 2.0.6-5.2.1. But I'm  
hoping that upgrading to 2.2.0 will help, I downloaded and installed  
it a few minutes ago.  Plone is 2.5, zope is 2.9.4-final,  
LDAPUserFolder 2.7, LDAPMultiPlugins 1.3.

Does any of this send up red flags? Here's a traceback I got from  
zope's DeadlockDebugger:

http://rafb.net/paste/results/lHaHty24.html

I think the relevant section is here:

File "/opt/zope/instance1/Products/LDAPUserFolder/LDAPDelegate.py",  
line 189, in connect
     conn.simple_bind_s(user_dn, user_pwd)
   File "/opt/zope/instance1/Products/LDAPUserFolder/ 
SharedResource.py", line 110, in __call__
     self._lock.acquire()
   File "/usr/lib/python2.4/threading.py", line 100, in acquire
     rc = self.__block.acquire(blocking)


After doing a little research, I came across the same errors in this  
message:

http://www.dataflake.org/tracker/issue_00423

And that led me to a message from this list last year:

http://sourceforge.net/mailarchive/message.php?msg_id=11183313

So I think I've found the right place...

Hopefully I'll write in tomorrow saying that the problem is solved  
and the Plone instance is stable again. If so, I hope this message  
guides other folks that encounter this problem toward a solution.

Crossing fingers & toes,
Troy