[Zope-CMF] zope spins when try to update workflow

Edmund Moseley edmund at unc.edu
Fri Dec 29 17:13:54 EST 2006


Quoting Andreas Jung <lists at andreas-jung.com>:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
>
>
> - --On 29. Dezember 2006 11:17:27 -0500 Edmund Moseley <edmund at unc.edu>
> wrote:
>
>> Quoting Andreas Jung <lists at andreas-jung.com>:
>>
>>> -----BEGIN PGP SIGNED MESSAGE-----
>>> Hash: SHA1
>>>
>>>
>>>
>>> - --On 29. Dezember 2006 10:56:50 -0500 Edmund Moseley <edmund at unc.edu>
>>> wrote:
>>>> Jens,
>>>> Thanks for your generic answer. What details about my setup might help
>>>> in figureing out how I could successfully update security without my
>>>> server grinding to a halt.
>>>> Zope 2.9.5, Python 2.4.4, Plone 2.5.1
>>>> 4GB of physical memory
>>>> 4GB of swap
>>>> 2 Duo Core processors
>>>>
>>>> I have about 35,000 objects in the ZODB, but I'm sure this doesn't come
>>>> close to the total amount of memory.
>>>>
>>>> We have a ZEO server and client on the same machine.
>>>>
>>>
>>> Check out the Advanced tab of your portal_catalog. There is configuration
>>> setting (threshold or so) to enable subtransactions/savepoints.
>>>
>>> - -aj
>>
>> Thanks for this tip. Subtransactions are already enabled, and the
>> threshold is set to  10000. Should I drop this number? If so, by how much?
>
> 10000 should be fine in most cases....but it depends on the size of your
> objects. Try to reduce it to 1000 and try again. Don't forget to restart
> before reindexing..Zope/Python won't free any memory (in most cases).
>
> - -aj

Is there a similar configuration for portal_workflow?

Thanks,
Edmund


More information about the Zope-CMF mailing list