Open maltemoeser opened 5 years ago
Currently the RPC parser uses the getblock
command and fetches the whole block, not just the header (getblockheader
was added in Bitcoin Core v0.12). It then keeps all that data in memory, which exceeds the memory on our machine.
In the future we may want to patch the Bitcoin CPP API to include the getblockheader
command and do the initial block header fetch with block headers only, and then lazily load the information from the node if we need it for the parse.
Parsing Bitcoin Core over RPC runs out of memory during fetching of block headers.
Output of
/var/log/syslog
:System Information
BlockSci version: v0.6 Using AMI: no Total memory: 60 GB