[Zope] Re: [Zope-DB] ERROR(200) Conflict Resolution Unexpected error)

[Dieter Maurer]

I redirected the response to "zope at zope.org" where is it better placed.

Gerhard Schmidt wrote at 2004-1-14 16:19 +0100:
>we are running an zope system with one ZEO and six Zope Frontends
>
>The system ist running for one Month without a problem under 
>zope2.7b3
>
>Today I had several errors in the zeo.log. 
>
>2004-01-14T16:01:34 ERROR(200) Conflict Resolution Unexpected error
>Traceback (most recent call last):
>  File "/usr/local/www/Zope/lib/python/ZODB/ConflictResolution.py", line 109, in tryT
>oResolveConflict
>    old = state(self, oid, oldSerial, prfactory)
>  File "/usr/local/www/Zope/lib/python/ZODB/ConflictResolution.py", line 38, in state
>    p = p or self.loadSerial(oid, serial)
>  File "/usr/local/www/Zope/lib/python/ZODB/FileStorage.py", line 712, in loadSerial
>    raise POSKeyError(serial)
>POSKeyError: 0000000000000000

This can happen when you recently packed the ZODB.
In this case, the original version of the object may no longer
be available. Beside the log entry above, this should have no
other effect.

>Soon all fronends run out of service and this error is logged. 

Do you see any log entries in the frontends log files?

The exception in the server should cause a "ConflictError"
in one of its clients. This client should retry its request.
All other clients should be unaffected.

Or, to say it differently, your frontends should not die
due to the above problem.

>I have rebuild the Data.fs with fsrecover.py. No Errors reported 
>but its runnig again. But still some of the above errors are logged 
>in zeo.log. 
>
>Anybody some Idea what caused this failure. 

Do you pack your database continuously?
Usually (infrequent packing), one would expect that the problem
above would occur with a very small probability...

-- 
Dieter