Does anyone experience or know about memory leak problems with propel 1.x when handling large amounts of complex join queries during data synchronisation? How do you address these problems? Is it useful to disable the Propel cache using disableInstancePooling() during data export?
The issues are not the joins, but loading a lot of different entities that are kept in the instance pool. You might face the issue because every joined entity/table will make the usage of the instance cache even worse.
So yes we have deactivated the instance pooling for exports / imports that interact with a lot of entities in a single process.
Just use
\Propel\Runtime\Propel::disableInstancePooling()
at the beginning of your process
The issues are not the joins, but loading a lot of different entities that are kept in the instance pool. You might face the issue because every joined entity/table will make the usage of the instance cache even worse.
So yes we have deactivated the instance pooling for exports / imports that interact with a lot of entities in a single process.
Just use
\Propel\Runtime\Propel::disableInstancePooling()
at the beginning of your process
Thank you for your feedback. π