dat
dat copied to clipboard
Check existing data when using dat clone
It's much easier to jump into dat if you can import your existing data sets.
Use case:
Currently using rsync to mirror 20K files totaling 10TB from a single server to a handful of client servers. Data is already in sync.
Cloning a dat from the central server onto an existing directory will use existing client data where possible and remove inconsistencies.
Great idea. We should support this I agree
Any updates? I think we really need it
Thanks for pinging this one. Definitely a good idea.
Note to self on implementation from mafintosh:
we need a .verifyData method in hypercore that checks that data in it's feed matches the hashes stored
On implementation, for future, I'm not sure if this is related, but I can't help but dream about ways to optimize these transfers, if there are multiple locations with partial data, there might be some sneaky optimizations possible using the tricks hidden in the Pedersen Commit Scheme, relaxing verification a bit, if one is somewhat confident that the current graph is secure from prying eyes, but I may be terribly off-base. I haven't looked into protocol incompatibilities at all. Feel free to ignore.
Referencing this here: https://github.com/mafintosh/hyperdrive/issues/203