The in-process IPFS nodes currently spawned in the testbed are configured to use an in-memory datastore. This is preventing us from being able to run experiments with large datasets (https://github.com/adlrocha/beyond-bitswap/blob/bcbbdfd7bdf20ea65e8d7a256bcc95272ee2059a/testbed/utils/ipfs.go#L172).
This can be fixed by using, for instance, a flat-fs datastore. I performed a simple test using a flatfs repo (without dependency injection) and it fixes the problem. Fixing this for the configurable in-process node with dependency injection needs a bit more of work, and I am not going to address this right now.
The in-process IPFS nodes currently spawned in the testbed are configured to use an in-memory datastore. This is preventing us from being able to run experiments with large datasets (https://github.com/adlrocha/beyond-bitswap/blob/bcbbdfd7bdf20ea65e8d7a256bcc95272ee2059a/testbed/utils/ipfs.go#L172).
This can be fixed by using, for instance, a flat-fs datastore. I performed a simple test using a flatfs repo (without dependency injection) and it fixes the problem. Fixing this for the configurable in-process node with dependency injection needs a bit more of work, and I am not going to address this right now.