ShadowDb

ShadowDb

Overview

ShadowDb is a wrapper for any StateDb operations. It runs all operations on two StateDbs simultaneously hence slowing down the command itself.

Using ShadowDb without existing StateDb

To run for example runvm with ShadowDb we need to specify we want to use it with flag --shadow-db then we specify implementation with db-shadow-impl (carmen, geth...) and variant with --db-shadow-variant (go-file, cpp-file...). Using --keep-db will keep both prime and shadow StateDb in structure path/to/state/db/tmp/prime and path/to/state/db/tmp/shadow.

Using ShadowDb with existing StateDb

To run for example apireplay with ShadowDb we need to respect expected structure. First we specify using ShadowDb with --shadow-db. Then we specify path to StateDb and ShadowDb with --db-src in which we have to have two StateDbs directories. One named prime other named shadow. Implementation and Variant are both read from statedb_info.json.