622 Commits

This Branch

This Branch

All Branches

Author	SHA1	Message	Date
Roman Zeyde	774d75bd97	Use match statement for daemon_dir construction	2018-07-16 09:18:48 +03:00
Lawrence Nahum	a85b6ba0c8	add support for regtest	2018-07-15 23:10:29 +02:00
Roman Zeyde	86e5cbda2e	Don't run cargo with --verbose on Travis	2018-07-16 00:00:03 +03:00
Roman Zeyde	de15dbd392	Fix formatting	2018-07-15 23:53:19 +03:00
Roman Zeyde	2d74d6d4f0	Rename a few Config members following #19 ... rpc_addr -> electrum_rpc_addr daemon_rpc_url -> daemon_rpc_addr	2018-07-15 22:39:19 +03:00
Lawrence Nahum	408372cba4	make listen port and bitcoin rpc configurable	2018-07-15 19:08:15 +02:00
Roman Zeyde	ea65239141	Trace received Electrum RPC requests	2018-07-15 14:49:51 +03:00
Roman Zeyde	5507ee1a8b	Simplify main loop using Option<RPC>	2018-07-15 14:05:17 +03:00
Roman Zeyde	d1f5c61def	Use correct txids for getting mempool transactions ... If block is found after "getmempoolentry()" and before "gettransactions()", there will be more transactions at `txs` than in `entries` - causing a panic.	2018-07-15 09:28:50 +03:00
Roman Zeyde	ea24d6a01f	Add release notes	2018-07-15 09:14:45 +03:00
Roman Zeyde	ab4747ce7e	Add cmake to installation instructions	2018-07-15 07:49:23 +03:00
Roman Zeyde	849c83ed64	Bump version	2018-07-14 18:04:27 +03:00
Roman Zeyde	00de0c028c	Log daemon's network info ... (including its version string)	2018-07-14 17:13:34 +03:00
Roman Zeyde	253e244a86	Fix cargo fmt for Travis ... Stable Rust doesn't support `cargo fmt --check` (yet).	2018-07-14 13:02:28 +03:00
Roman Zeyde	197e5792f8	Improve error message for daemon disconnection	2018-07-14 12:42:21 +03:00
Roman Zeyde	ac9e300226	Run 'cargo fmt --check' on Travis	2018-07-14 10:14:35 +03:00
Roman Zeyde	2852020c58	Fix leftovers at examples/load.rs	2018-07-13 22:24:14 +03:00
Roman Zeyde	daee3fbe4b	Replace "Vec::new()" by shorter "vec![]"	2018-07-13 21:30:30 +03:00
Roman Zeyde	1a9b977dad	Update TODOs ... - implemented Daemon::gettransactions() API - implemented multithreaded indexing workers - SSTable injection requires non-overlapping ranges (= pre-sorting = full compaction) - HTTP parsing can be improved	2018-07-13 21:12:26 +03:00
Roman Zeyde	48cd929a99	Remove StoreOptions and enable manual compactions explicitly	2018-07-13 14:23:10 +03:00
Roman Zeyde	9a89e6a2d8	Update README with latest performance metrics	2018-07-12 23:26:37 +03:00
Roman Zeyde	03790d9977	Add logging for bulk indexing	2018-07-12 21:30:33 +03:00
Roman Zeyde	4d322d4819	Substract signed integers	2018-07-12 19:38:39 +03:00
Roman Zeyde	1e258f4b3d	Allow vectorized fetch for mempool transactions	2018-07-12 19:38:16 +03:00
Roman Zeyde	cdacf451a9	Log delta of mempool transactions	2018-07-12 18:58:38 +03:00
Roman Zeyde	e09ac477bb	Skip bulk indexing if it's already over	2018-07-12 16:52:25 +03:00
Roman Zeyde	b9a4cb4318	Move bulk reader and indexer to separate module	2018-07-12 16:47:39 +03:00
Roman Zeyde	4c7413db9d	Increase open files limit explicitly ... Otherwise, full compaction may fail.	2018-07-12 14:44:10 +03:00
Roman Zeyde	c7b8f57228	Keep default parallism settings	2018-07-12 12:47:05 +03:00
Roman Zeyde	125bfe0e15	Log latest indexed block	2018-07-12 10:47:25 +03:00
Roman Zeyde	62a78f18ad	Put "finish" marker after compaction is over	2018-07-12 10:46:33 +03:00
Roman Zeyde	f92dfc808d	Log indexed blk*.dat file path	2018-07-12 09:25:17 +03:00
Roman Zeyde	7b5679db70	Panic if DB exists before running bulk load benchmark	2018-07-11 21:26:12 +03:00
Roman Zeyde	b32e503db2	Rename regular indexing benchmark	2018-07-11 21:26:12 +03:00
Roman Zeyde	dd70a1cc8a	Persist last indexed block marker	2018-07-11 21:26:12 +03:00
Roman Zeyde	6609d6e008	Store indexed blocks' hashes	2018-07-11 21:01:54 +03:00
Roman Zeyde	2158fa1961	Sort blk*.dat files	2018-07-11 21:00:57 +03:00
Roman Zeyde	10919035bd	Rename bulk load benchmark	2018-07-11 20:44:05 +03:00
Roman Zeyde	4da521b3cd	Refactor bulk load benchmark ... - bulk writes take ~1h (reading blk*.dat files at ~48MB/s) - full compaction takes ~45m (writing 37GB of SSTables)	2018-07-11 20:42:57 +03:00
Roman Zeyde	1f979baf8f	Increase compaction readahead to 2MB ... This should improve compaction throughput on HDDs.	2018-07-11 20:33:09 +03:00
Roman Zeyde	219a962145	Use sort_unstable_by() instead of implementing Ord for Row	2018-07-11 20:31:44 +03:00
Roman Zeyde	ef7a7eac8a	Sort rows before writing to DB ... Can happen in parallel, before sending the rows to the DB writer thread.	2018-07-11 16:26:06 +03:00
Roman Zeyde	49ec4d1c38	Support SSTable ingestion into DB	2018-07-11 13:48:57 +03:00
Roman Zeyde	e69686f988	Use multiple bulk index workers	2018-07-10 23:52:26 +03:00
Roman Zeyde	4b83c6536f	Add SSTableWriter to store module	2018-07-10 16:09:31 +03:00
Roman Zeyde	fb372b2915	Split reading from indexing blk*.dat files ... Also, use a single Vec<Row> for write operation	2018-07-10 15:37:56 +03:00
Roman Zeyde	56229409a0	Simplify SIGINT polling	2018-07-10 15:29:56 +03:00
Roman Zeyde	314978f30a	Simplify bulk indexing flow ... Use separate thread per blk*.dat file.	2018-07-10 15:28:04 +03:00
Roman Zeyde	45f3c7d178	Make store::Row sortable (by key) and clonable	2018-07-09 19:07:33 +03:00
Roman Zeyde	baf2d8624d	Rename bulk_load() -> bulk_index()	2018-07-09 18:02:49 +03:00