Roman Zeyde

OTOH, maybe it's even simpler to just store a "giant map" (using a new RocksDB column family), mapping: ``` transaction_global_index -> transaction_byte_offset_within_its_block ``` - `transaction_global_index` is computed by enumerating all...

The additional storage requirements are not expected to be large, since the "map" is sorted by its key (so it should compress well). The value (`transaction_byte_offset_within_its_block`) can be represented using...

https://github.com/romanz/electrs/compare/status-cache seems to improve subsequent large wallet loading significantly by using an ephemeral cache. Can be tested by loading the following mainnet wallet using Sparrow wallet: ``` seed: "all all...

Latest push https://github.com/romanz/electrs/commit/176fd4ca6f43c28dcc82cb3e8b856080af8a63c3 uses a separate RocksDB instance (instead of a HashMap) for persistance. Since the DB contents are derived from the addresses being queried, make sure to store it...

It would be interesting to test it on a Raspberry Pi :) CC: @ejose19 @Transisto

IIUC, we can use `ramfs` to store the cache DB: - https://wiki.debian.org/ramfs - https://unix.stackexchange.com/a/325421. WDYT?

If we use [SQLite](https://docs.rs/sqlite/) for cache DB, we can default to in-memory storage: ``` If the filename is ":memory:", then a private, temporary in-memory database is created for the connection....

Many thanks for reporting this issue! > ... when I went to query it. I got a "hidden service not found" or something similar to that effect. Could you please...

CC: @lukechilds

I can suggest using Prometheus interface for checking when the initial indexing is over (https://github.com/romanz/electrs/issues/505): ``` $ bitcoin-cli getblockchaininfo | jq .blocks 703805 $ curl --no-progress-meter localhost:4224 | grep index_height...