[native] Adapt new async cache config

prestodb · Jun 1, 2024 · fea27e0 · fea27e0
1 parent 6514b46
commit fea27e0
Show file tree

Hide file tree

Showing 4 changed files with 46 additions and 23 deletions.
diff --git a/presto-native-execution/presto_cpp/main/PrestoServer.cpp b/presto-native-execution/presto_cpp/main/PrestoServer.cpp
@@ -651,6 +651,29 @@ void PrestoServer::initializeThreadPools() {
   }
 }
 
+std::unique_ptr<cache::SsdCache> PrestoServer::setupSsdCache() {
+  VELOX_CHECK_NULL(cacheExecutor_);
+  auto* systemConfig = SystemConfig::instance();
+  if (systemConfig->asyncCacheSsdGb() == 0) {
+    return nullptr;
+  }
+
+  constexpr int32_t kNumSsdShards = 16;
+  cacheExecutor_ = std::make_unique<folly::IOThreadPoolExecutor>(kNumSsdShards);
+  cache::SsdCache::Config cacheConfig(
+      systemConfig->asyncCacheSsdPath(),
+      systemConfig->asyncCacheSsdGb() << 30,
+      kNumSsdShards,
+      cacheExecutor_.get(),
+      systemConfig->asyncCacheSsdCheckpointGb() << 30,
+      systemConfig->asyncCacheSsdDisableFileCow(),
+      systemConfig->ssdCacheChecksumEnabled(),
+      systemConfig->ssdCacheReadVerificationEnabled());
+  PRESTO_STARTUP_LOG(INFO) << "Initializing SSD cache with "
+                           << cacheConfig.toString();
+  return std::make_unique<cache::SsdCache>(cacheConfig);
+}
+
 void PrestoServer::initializeVeloxMemory() {
   auto* systemConfig = SystemConfig::instance();
   const uint64_t memoryGb = systemConfig->systemMemoryGb();
@@ -695,29 +718,7 @@ void PrestoServer::initializeVeloxMemory() {
                            << memory::memoryManager()->toString();
 
   if (systemConfig->asyncDataCacheEnabled()) {
-    std::unique_ptr<cache::SsdCache> ssd;
-    const auto asyncCacheSsdGb = systemConfig->asyncCacheSsdGb();
-    if (asyncCacheSsdGb > 0) {
-      constexpr int32_t kNumSsdShards = 16;
-      cacheExecutor_ =
-          std::make_unique<folly::IOThreadPoolExecutor>(kNumSsdShards);
-      auto asyncCacheSsdCheckpointGb =
-          systemConfig->asyncCacheSsdCheckpointGb();
-      auto asyncCacheSsdDisableFileCow =
-          systemConfig->asyncCacheSsdDisableFileCow();
-      PRESTO_STARTUP_LOG(INFO)
-          << "Initializing SSD cache with capacity " << asyncCacheSsdGb
-          << "GB, checkpoint size " << asyncCacheSsdCheckpointGb
-          << "GB, file cow "
-          << (asyncCacheSsdDisableFileCow ? "DISABLED" : "ENABLED");
-      ssd = std::make_unique<cache::SsdCache>(
-          systemConfig->asyncCacheSsdPath(),
-          asyncCacheSsdGb << 30,
-          kNumSsdShards,
-          cacheExecutor_.get(),
-          asyncCacheSsdCheckpointGb << 30,
-          asyncCacheSsdDisableFileCow);
-    }
+    std::unique_ptr<cache::SsdCache> ssd = setupSsdCache();
     std::string cacheStr =
         ssd == nullptr ? "AsyncDataCache" : "AsyncDataCache with SSD";
     cache_ = cache::AsyncDataCache::create(

diff --git a/presto-native-execution/presto_cpp/main/PrestoServer.h b/presto-native-execution/presto_cpp/main/PrestoServer.h
@@ -202,6 +202,8 @@ class PrestoServer {
 
   void registerSystemConnector();
 
+  std::unique_ptr<velox::cache::SsdCache> setupSsdCache();
+
   const std::string configDirectoryPath_;
 
   std::shared_ptr<CoordinatorDiscoverer> coordinatorDiscoverer_;

diff --git a/presto-native-execution/presto_cpp/main/common/Configs.cpp b/presto-native-execution/presto_cpp/main/common/Configs.cpp
@@ -438,6 +438,14 @@ bool SystemConfig::asyncCacheSsdDisableFileCow() const {
   return optionalProperty<bool>(kAsyncCacheSsdDisableFileCow).value();
 }
 
+bool SystemConfig::ssdCacheChecksumEnabled() const {
+  return optionalProperty<bool>(kSsdCacheChecksumEnabled).value();
+}
+
+bool SystemConfig::ssdCacheReadVerificationEnabled() const {
+  return optionalProperty<bool>(kSsdCacheReadVerificationEnabled).value();
+}
+
 std::string SystemConfig::shuffleName() const {
   return optionalProperty(kShuffleName).value();
 }

diff --git a/presto-native-execution/presto_cpp/main/common/Configs.h b/presto-native-execution/presto_cpp/main/common/Configs.h
@@ -320,6 +320,14 @@ class SystemConfig : public ConfigBase {
   /// option to disable cow for cache files.
   static constexpr std::string_view kAsyncCacheSsdDisableFileCow{
       "async-cache-ssd-disable-file-cow"};
+  /// When enabled, a CRC-based checksum is calculated for each cache entry
+  /// written to SSD. The checksum is stored in the next checkpoint file.
+  static constexpr std::string_view kSsdCacheChecksumEnabled{
+      "ssd-cache-checksum-enabled"};
+  /// When enabled, the checksum is recalculated and verified against the stored
+  /// value when cache data is loaded from the SSD.
+  static constexpr std::string_view kSsdCacheReadVerificationEnabled{
+      "ssd-cache-read-verification-enabled"};
   static constexpr std::string_view kEnableSerializedPageChecksum{
       "enable-serialized-page-checksum"};
 
@@ -625,6 +633,10 @@ class SystemConfig : public ConfigBase {
 
   bool asyncCacheSsdDisableFileCow() const;
 
+  bool ssdCacheChecksumEnabled() const;
+
+  bool ssdCacheReadVerificationEnabled() const;
+
   std::string shuffleName() const;
 
   bool enableSerializedPageChecksum() const;