(OSI prod) add file gives database conflict error

[Jim B. Glenn]

reported via error monitor:

ENTRY
Mon Aug 9 14:27:57 2010
Exception when processing https://karl.soros.org/communities/hepatitis-c/files/demand-access/community-educational-materials/add_file.html
Traceback (most recent call last):
  File "/opt/karl/osi/3.20.2-0/src/karl/karl/errorlog.py", line 15, in middleware
    return app(environ, start_response)
  File "/opt/karl/osi/3.20.2-0/eggs/repoze.zodbconn-0.11-py2.5.egg/repoze/zodbconn/connector.py", line 21, in __call__
    result = self.next_app(environ, start_response)
  File "/opt/karl/osi/3.20.2-0/eggs/repoze.retry-0.9.4-py2.5.egg/repoze/retry/__init__.py", line 88, in __call__
    app_iter = self.application(environ, replace_start_response)
  File "/opt/karl/osi/3.20.2-0/eggs/repoze.tm2-1.0a5-py2.5.egg/repoze/tm/__init__.py", line 47, in __call__
    self.commit()
  File "/opt/karl/osi/3.20.2-0/eggs/repoze.tm2-1.0a5-py2.5.egg/repoze/tm/__init__.py", line 52, in commit
    t.commit()
  File "/opt/karl/osi/3.20.2-0/eggs/ZODB3-3.8.5-py2.5-linux-x86_64.egg/transaction/_transaction.py", line 325, in commit
    self._commitResources()
  File "/opt/karl/osi/3.20.2-0/eggs/ZODB3-3.8.5-py2.5-linux-x86_64.egg/transaction/_transaction.py", line 424, in _commitResources
    rm.commit(self)
  File "/opt/karl/osi/3.20.2-0/eggs/ZODB3-3.8.5-py2.5-linux-x86_64.egg/ZODB/Connection.py", line 545, in commit
    self._commit(transaction)
  File "/opt/karl/osi/3.20.2-0/eggs/ZODB3-3.8.5-py2.5-linux-x86_64.egg/ZODB/Connection.py", line 582, in _commit
    raise ConflictError(object=obj)
ConflictError: database conflict error (oid 0x4f93fa, class persistent.mapping.PersistentMapping)

[Paul Everitt (paul-agendaless)]

Giving to Chris as part of the thinking about what conflict errors to care about. Feel free to close this as invalid.

[Chris Rossi (chris-archimedeanco)]

I will at least double check that conflict errors don't make it into the log unless we use up all of the tries in repoze.retry. It may be simple matter of moving repoze.retry to be deeper in the wsgi stack than the error monitor.

[Chris Rossi (chris-archimedeanco)]

I double checked and via examination of the code it does not seem possible for a conflict error to get logged unless the retry limit in repoze.retry is exceeded. This seems to be an isolated case of getting more than 3 conflict errors in a row, and thus exceeding the number of exceptions repoze.retry is willing to swallow.