From b44d3fc796484a50cd6854f20c9b46e5fddedc9d Mon Sep 17 00:00:00 2001 From: Alex Date: Wed, 8 Jun 2022 10:01:44 +0200 Subject: Abstract database behind generic interface and implement alternative drivers (#322) - [x] Design interface - [x] Implement Sled backend - [x] Re-implement the SledCountedTree hack ~~on Sled backend~~ on all backends (i.e. over the abstraction) - [x] Convert Garage code to use generic interface - [x] Proof-read converted Garage code - [ ] Test everything well - [x] Implement sqlite backend - [x] Implement LMDB backend - [ ] (Implement Persy backend?) - [ ] (Implement other backends? (like RocksDB, ...)) - [x] Implement backend choice in config file and garage server module - [x] Add CLI for converting between DB formats - Exploit the new interface to put more things in transactions - [x] `.updated()` trigger on Garage tables Fix #284 **Bugs** - [x] When exporting sqlite, trees iterate empty?? - [x] LMDB doesn't work **Known issues for various back-ends** - Sled: - Eats all my RAM and also all my disk space - `.len()` has to traverse the whole table - Is actually quite slow on some operations - And is actually pretty bad code... - Sqlite: - Requires a lock to be taken on all operations. The lock is also taken when iterating on a table with `.iter()`, and the lock isn't released until the iterator is dropped. This means that we must be VERY carefull to not do anything else inside a `.iter()` loop or else we will have a deadlock! Most such cases have been eliminated from the Garage codebase, but there might still be some that remain. If your Garage-over-Sqlite seems to hang/freeze, this is the reason. - (adapter uses a bunch of unsafe code) - Heed (LMDB): - Not suited for 32-bit machines as it has to map the whole DB in memory. - (adpater uses a tiny bit of unsafe code) **My recommendation:** avoid 32-bit machines and use LMDB as much as possible. **Converting databases** is actually quite easy. For example from Sled to LMDB: ```bash cd src/db cargo run --features cli --bin convert -- -i path/to/garage/meta/db -a sled -o path/to/garage/meta/db.lmdb -b lmdb ``` Then, just add this to your `config.toml`: ```toml db_engine = "lmdb" ``` Co-authored-by: Alex Auvolat Reviewed-on: https://git.deuxfleurs.fr/Deuxfleurs/garage/pulls/322 Co-authored-by: Alex Co-committed-by: Alex --- src/model/s3/block_ref_table.rs | 21 ++++++++++++--------- src/model/s3/object_table.rs | 12 ++++++++++-- src/model/s3/version_table.rs | 13 +++++++++++-- 3 files changed, 33 insertions(+), 13 deletions(-) (limited to 'src/model/s3') diff --git a/src/model/s3/block_ref_table.rs b/src/model/s3/block_ref_table.rs index 9b3991bf..9589b4aa 100644 --- a/src/model/s3/block_ref_table.rs +++ b/src/model/s3/block_ref_table.rs @@ -1,6 +1,8 @@ use serde::{Deserialize, Serialize}; use std::sync::Arc; +use garage_db as db; + use garage_util::data::*; use garage_table::crdt::Crdt; @@ -51,21 +53,22 @@ impl TableSchema for BlockRefTable { type E = BlockRef; type Filter = DeletedFilter; - fn updated(&self, old: Option<&Self::E>, new: Option<&Self::E>) { - #[allow(clippy::or_fun_call)] - let block = &old.or(new).unwrap().block; + fn updated( + &self, + tx: &mut db::Transaction, + old: Option<&Self::E>, + new: Option<&Self::E>, + ) -> db::TxOpResult<()> { + let block = old.or(new).unwrap().block; let was_before = old.map(|x| !x.deleted.get()).unwrap_or(false); let is_after = new.map(|x| !x.deleted.get()).unwrap_or(false); if is_after && !was_before { - if let Err(e) = self.block_manager.block_incref(block) { - warn!("block_incref failed for block {:?}: {}", block, e); - } + self.block_manager.block_incref(tx, block)?; } if was_before && !is_after { - if let Err(e) = self.block_manager.block_decref(block) { - warn!("block_decref failed for block {:?}: {}", block, e); - } + self.block_manager.block_decref(tx, block)?; } + Ok(()) } fn matches_filter(entry: &Self::E, filter: &Self::Filter) -> bool { diff --git a/src/model/s3/object_table.rs b/src/model/s3/object_table.rs index 3d9a89f7..62f5d8d9 100644 --- a/src/model/s3/object_table.rs +++ b/src/model/s3/object_table.rs @@ -2,6 +2,8 @@ use serde::{Deserialize, Serialize}; use std::collections::BTreeMap; use std::sync::Arc; +use garage_db as db; + use garage_util::background::BackgroundRunner; use garage_util::data::*; @@ -232,7 +234,12 @@ impl TableSchema for ObjectTable { type E = Object; type Filter = ObjectFilter; - fn updated(&self, old: Option<&Self::E>, new: Option<&Self::E>) { + fn updated( + &self, + _tx: &mut db::Transaction, + old: Option<&Self::E>, + new: Option<&Self::E>, + ) -> db::TxOpResult<()> { let version_table = self.version_table.clone(); let old = old.cloned(); let new = new.cloned(); @@ -259,7 +266,8 @@ impl TableSchema for ObjectTable { } } Ok(()) - }) + }); + Ok(()) } fn matches_filter(entry: &Self::E, filter: &Self::Filter) -> bool { diff --git a/src/model/s3/version_table.rs b/src/model/s3/version_table.rs index ad096772..881c245a 100644 --- a/src/model/s3/version_table.rs +++ b/src/model/s3/version_table.rs @@ -1,6 +1,8 @@ use serde::{Deserialize, Serialize}; use std::sync::Arc; +use garage_db as db; + use garage_util::background::BackgroundRunner; use garage_util::data::*; @@ -137,7 +139,12 @@ impl TableSchema for VersionTable { type E = Version; type Filter = DeletedFilter; - fn updated(&self, old: Option<&Self::E>, new: Option<&Self::E>) { + fn updated( + &self, + _tx: &mut db::Transaction, + old: Option<&Self::E>, + new: Option<&Self::E>, + ) -> db::TxOpResult<()> { let block_ref_table = self.block_ref_table.clone(); let old = old.cloned(); let new = new.cloned(); @@ -160,7 +167,9 @@ impl TableSchema for VersionTable { } } Ok(()) - }) + }); + + Ok(()) } fn matches_filter(entry: &Self::E, filter: &Self::Filter) -> bool { -- cgit v1.2.3