milenkovicm commented on issue #3941: URL: https://github.com/apache/arrow-datafusion/issues/3941#issuecomment-1291849770
Thanks @alamb, As I said in previous comment, the biggest problem of integrating aggregation with memory manager is `async` at `MemoryConsumer::try_grow` function. `async` is there to facilitate `async spill(..)` additional memory can't be acquired. easy ways to work around `async` would be either to: * de-couple `try_grow` and `spill`, removing `async` from `try_grow`, do manual `spill` in case of failure. * component to use memory manager directly `memory_manager().can_grow_directly` (we need to make this function public), do manual spill in case of failure wdyt? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
