bahman/scylla
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

db/batchlog_manager: bypass cache when scanning batchlog table

Browse Source 
Scans should not pollute the cache with cold data, in general. In the
case of the batchlog table, there is another reason to bypass the cache:
this table can have a lot of partition tombstones, which currently are
not purged from the cache. So in certain cases, using the cache can make
batch replay very slow, because it has to scan past tombstones of
already replayed batches.
This commit is contained in:
Botond Dénes
2024-06-19 09:43:01 -04:00
parent 29f610d861
commit 31c0fa07d8
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
db/batchlog_manager.cc
View File

@@ -134,7 +134,7 @@ future<> db::batchlog_manager::stop() {
}
future<size_t> db::batchlog_manager::count_all_batches() const {
sstring query = format("SELECT count(*) FROM {}.{}", system_keyspace::NAME, system_keyspace::BATCHLOG);
sstring query = format("SELECT count(*) FROM {}.{} BYPASS CACHE", system_keyspace::NAME, system_keyspace::BATCHLOG);
return _qp.execute_internal(query, cql3::query_processor::cache_internal::yes).then([](::shared_ptr<cql3::untyped_result_set> rs) {
return size_t(rs->one().get_as<int64_t>("count"));
});
@@ -260,7 +260,7 @@ future<> db::batchlog_manager::replay_all_failed_batches() {
return seastar::with_gate(_gate, [this, batch = std::move(batch)] () mutable {
blogger.debug("Started replayAllFailedBatches (cpu {})", this_shard_id());
return _qp.query_internal(
format("SELECT id, data, written_at, version FROM {}.{})", system_keyspace::NAME, system_keyspace::BATCHLOG),
format("SELECT id, data, written_at, version FROM {}.{} BYPASS CACHE", system_keyspace::NAME, system_keyspace::BATCHLOG),
db::consistency_level::ONE,
{},
page_size,