Cleanup of node data (to fix 'Bus error: 10' error)

Question

There are already two unanswered questions about this subject but this is important and I decided to ask a slightly different one. The subject is: sometimes, the node gets in a bad state and it won't run, only failing with "Bus error: 10". I personally saw it after a power outage for example.

Is there a way to cleanup / reset a node which would solve the issue or do something somewhat similar to these steps (which seem to solve the issue)?

1. ./tezos-node snapshot export MySnapshot
2. Move away your node ~/.tezos-node
3. ./tezos-node identity generate (I don't fully care about this step)
4. ./tezos-node snapshot import MySnapshot

Bus Error 10 - Node Won't Start

Starting Tezos node gives "Bus error: 10"

Answer 1

When the node is brutally stopped it should often result in a corrupted storage as the node may be interrupted when writing something in the storage. Such a corruption is quite hard to repair and one should favor restarting the node on "a fresh storage".

If your storage is corrupted, it is very likely that you could not even export a snapshot from it. You should rather get a recent snapshot from somewhere else. Before importing a snapshot you should remove your node's data directory (~/.tezos-node by default) but you can keep the identity.json as it can take a while to be generated. Then, you should import a snapshot as usual.