When building a node, do I have to sync its block data from scratch or can a load some kind of snapshot and finish syncing from there?
It's possible to start with tezos-node
data from another node and finish the sync from there. I do this by running the new tezos-node
briefly to have it create ~/.tezos-node
and the usual files in that. Then I stop tezos-node
and do the following:
cd ~/.tezos-node
rm -r context store
curl http://quicksync.tzdutch.com/latest.tar.lz4 | lz4 -d | tar xf -
The download and expand step can take several hours. Then I restart tezos-node
and let it finish syncing the data.
See https://www.tzdutch.com/quicksync/ about the source of the context and store data.
-
How hard would it be to create a proof that such a tarball matches all blocks for some height, and the put them on ipfs on some regular basis? 🤔 I have been downloading the tzdutch tarball e few times too, and it's great! But seems we should/could have a more general solution for this? – asbjornenge Jan 29 '19 at 21:33