Add support for multiple databases with rocksdb

Former-commit-id: cfa3b760b00776876134692d42e25a60fcd8dea9

src/IStorage.h

@@ -1,6 +1,13 @@
 #pragma once
 #include <functional>
 
+class IStorageFactory
+{
+public:
+    virtual ~IStorageFactory() {}
+    virtual class IStorage *create(int db) = 0;
+};
+
 class IStorage
 {
 public:
@@ -17,6 +24,8 @@ public:
     virtual void beginWriteBatch() {} // NOP
     virtual void endWriteBatch() {} // NOP
 
+    virtual void flush();
+
     /* This is permitted to be a shallow clone */
     virtual const IStorage *clone() const = 0;
 };

src/Makefile

@@ -208,7 +208,7 @@ endif
 
 REDIS_SERVER_NAME=keydb-server
 REDIS_SENTINEL_NAME=keydb-sentinel
-REDIS_SERVER_OBJ=adlist.o quicklist.o ae.o anet.o dict.o server.o sds.o zmalloc.o lzf_c.o lzf_d.o pqsort.o zipmap.o sha1.o ziplist.o release.o networking.o util.o object.o db.o replication.o rdb.o t_string.o t_list.o t_set.o t_zset.o t_hash.o config.o aof.o pubsub.o multi.o debug.o sort.o intset.o syncio.o cluster.o crc16.o endianconv.o slowlog.o scripting.o bio.o rio.o rand.o memtest.o crc64.o bitops.o sentinel.o notify.o setproctitle.o blocked.o hyperloglog.o latency.o sparkline.o redis-check-rdb.o redis-check-aof.o geo.o lazyfree.o module.o evict.o expire.o geohash.o geohash_helper.o childinfo.o defrag.o siphash.o rax.o t_stream.o listpack.o localtime.o acl.o storage.o rdb-s3.o fastlock.o new.o tracking.o AsyncWorkQueue.o snapshot.o storage/rocksdb.o $(ASM_OBJ)
+REDIS_SERVER_OBJ=adlist.o quicklist.o ae.o anet.o dict.o server.o sds.o zmalloc.o lzf_c.o lzf_d.o pqsort.o zipmap.o sha1.o ziplist.o release.o networking.o util.o object.o db.o replication.o rdb.o t_string.o t_list.o t_set.o t_zset.o t_hash.o config.o aof.o pubsub.o multi.o debug.o sort.o intset.o syncio.o cluster.o crc16.o endianconv.o slowlog.o scripting.o bio.o rio.o rand.o memtest.o crc64.o bitops.o sentinel.o notify.o setproctitle.o blocked.o hyperloglog.o latency.o sparkline.o redis-check-rdb.o redis-check-aof.o geo.o lazyfree.o module.o evict.o expire.o geohash.o geohash_helper.o childinfo.o defrag.o siphash.o rax.o t_stream.o listpack.o localtime.o acl.o storage.o rdb-s3.o fastlock.o new.o tracking.o AsyncWorkQueue.o snapshot.o storage/rocksdb.o storage/rocksdbfactory.o $(ASM_OBJ)
 REDIS_CLI_NAME=keydb-cli
 REDIS_CLI_OBJ=anet.o adlist.o dict.o redis-cli.o redis-cli-cpphelper.o zmalloc.o release.o anet.o ae.o crc64.o siphash.o crc16.o storage-lite.o fastlock.o new.o $(ASM_OBJ)
 REDIS_BENCHMARK_NAME=keydb-benchmark
@@ -298,6 +298,7 @@ dict-benchmark: dict.cpp zmalloc.cpp sds.c siphash.c
 
 clean:
 	rm -rf $(REDIS_SERVER_NAME) $(REDIS_SENTINEL_NAME) $(REDIS_CLI_NAME) $(REDIS_BENCHMARK_NAME) $(REDIS_CHECK_RDB_NAME) $(REDIS_CHECK_AOF_NAME) *.o *.gcda *.gcno *.gcov redis.info lcov-html Makefile.dep dict-benchmark
+	rm -rf storage/*.o
 
 .PHONY: clean
 

src/db.cpp

@@ -649,20 +649,27 @@ void randomkeyCommand(client *c) {
     decrRefCount(key);
 }
 
+static bool FEvictedDE(dictEntry *de)
+{
+    return (de != nullptr) && dictGetVal(de) == nullptr;
+}
 
 bool redisDbPersistentData::iterate(std::function<bool(const char*, robj*)> fn)
 {
     dictIterator *di = dictGetSafeIterator(m_pdict);
     dictEntry *de = nullptr;
     bool fResult = true;
-    while((de = dictNext(di)) != nullptr)
+    while(fResult && ((de = dictNext(di)) != nullptr))
     {
+        bool fEvicted = FEvictedDE(de);
+
         ensure((const char*)dictGetKey(de), &de);
         if (!fn((const char*)dictGetKey(de), (robj*)dictGetVal(de)))
-        {
             fResult = false;
-            break;
-        }
+
+        // re-evict the key so we don't OOM
+        if (fEvicted)
+            removeCachedValue((const char*)dictGetKey(de));
     }
     dictReleaseIterator(di);
 
@@ -679,10 +686,14 @@ bool redisDbPersistentData::iterate(std::function<bool(const char*, robj*)> fn)
                 return true;
 
             // Alright it's a key in the use keyspace, lets ensure it and then pass it off
+            bool fEvicted = FEvictedDE(de);
             ensure(key);
             deCurrent = dictFind(m_pdict, key);
-            return fn(key, (robj*)dictGetVal(deCurrent));
-        });
+            bool fResult =  fn(key, (robj*)dictGetVal(deCurrent));
+            if (fEvicted)
+                removeCachedValue(key);
+            return fResult;
+        }, true /*fKeyOnly*/);
     }
     
     return fResult;
@@ -713,7 +724,7 @@ void keysCommandCore(client *cIn, const redisDbPersistentDataSnapshot *db, sds p
             decrRefCount(keyobj);
         }
         return !(cIn->flags.load(std::memory_order_relaxed) & CLIENT_CLOSE_ASAP);
-    });
+    }, true /*fKeyOnly*/);
     
     setDeferredArrayLen(c,replylen,numkeys);
 
@@ -1906,7 +1917,6 @@ void redisDbPersistentData::setStorageProvider(IStorage *pstorage)
     });
 }
 
-IStorage *create_rocksdb_storage(const char *dbfile);
 void redisDb::initialize(int id)
 {
     redisDbPersistentData::initialize();
@@ -1918,8 +1928,8 @@ void redisDb::initialize(int id)
     this->avg_ttl = 0;
     this->last_expire_set = 0;
     this->defrag_later = listCreate();
-    if (id == 0)
-        this->setStorageProvider(create_rocksdb_storage("/tmp/rocks.db"));
+    if (g_pserver->m_pstorageFactory != nullptr)
+        this->setStorageProvider(g_pserver->m_pstorageFactory->create(id));
 }
 
 bool redisDbPersistentData::insert(char *key, robj *o)

src/server.cpp

@@ -61,6 +61,7 @@
 #include <uuid/uuid.h>
 #include <mutex>
 #include "aelocker.h"
+#include "storage/rocksdbfactory.h"
 
 int g_fTestMode = false;
 
@@ -2187,7 +2188,7 @@ void beforeSleep(struct aeEventLoop *eventLoop) {
     else {
         fFirstRun = false;
     }
-    
+
     aeReleaseLock();
 
     for (auto &pair : vecchanges)
@@ -3015,6 +3016,9 @@ void initServer(void) {
     signal(SIGPIPE, SIG_IGN);
     setupSignalHandlers();
 
+    // Create The Storage Factory (if necessary)
+    g_pserver->m_pstorageFactory = CreateRocksDBStorageFactory("/tmp/rocks.db", cserver.dbnum);
+
     zfree(g_pserver->db);   // initServerConfig created a dummy array, free that now
     g_pserver->db = (redisDb**)zmalloc(sizeof(redisDb*)*cserver.dbnum, MALLOC_LOCAL);
 
@@ -3891,6 +3895,7 @@ int prepareForShutdown(int flags) {
         delete g_pserver->db[idb];
         g_pserver->db[idb] = nullptr;
     }
+    delete g_pserver->m_pstorageFactory;
 
     serverLog(LL_WARNING,"%s is now ready to exit, bye bye...",
         g_pserver->sentinel_mode ? "Sentinel" : "KeyDB");

src/server.h

@@ -1381,7 +1381,7 @@ protected:
 public:
     bool FWillFreeChildDebug() const { return m_spdbSnapshotHOLDER != nullptr; }
 
-    bool iterate_threadsafe(std::function<bool(const char*, robj_roptr o)> fn) const;
+    bool iterate_threadsafe(std::function<bool(const char*, robj_roptr o)> fn, bool fKeyOnly = false) const;
     using redisDbPersistentData::createSnapshot;
     using redisDbPersistentData::endSnapshot;
     using redisDbPersistentData::end;
@@ -2274,6 +2274,8 @@ struct redisServer {
 
     GarbageCollector<redisDbPersistentDataSnapshot> garbageCollector;
 
+    IStorageFactory *m_pstorageFactory = nullptr;
+
     bool FRdbSaveInProgress() const { return rdbThreadVars.fRdbThreadActive; }
 };
 

src/snapshot.cpp

@@ -261,7 +261,7 @@ dict_iter redisDbPersistentDataSnapshot::find_threadsafe(const char *key) const
     return dict_iter(de);
 }
 
-bool redisDbPersistentDataSnapshot::iterate_threadsafe(std::function<bool(const char*, robj_roptr o)> fn) const
+bool redisDbPersistentDataSnapshot::iterate_threadsafe(std::function<bool(const char*, robj_roptr o)> fn, bool fKeyOnly) const
 {
     dictEntry *de = nullptr;
     bool fResult = true;
@@ -272,14 +272,22 @@ bool redisDbPersistentDataSnapshot::iterate_threadsafe(std::function<bool(const
     volatile size_t celem = size();
 
     dictIterator *di = dictGetSafeIterator(m_pdict);
-    while((de = dictNext(di)) != nullptr)
+    while(fResult && ((de = dictNext(di)) != nullptr))
     {
         --celem;
-        if (!fn((const char*)dictGetKey(de), (robj*)dictGetVal(de)))
+        robj *o = (robj*)dictGetVal(de);
+        if (o == nullptr && !fKeyOnly)
         {
-            fResult = false;
-            break;
+            m_spstorage->retrieve((sds)dictGetKey(de), sdslen((sds)dictGetKey(de)), [&](const char *, size_t, const void *data, size_t cb){
+                o = deserializeStoredObject(data, cb);
+            });
         }
+
+        if (!fn((const char*)dictGetKey(de), o))
+            fResult = false;
+
+        if (o != nullptr && dictGetVal(de) == nullptr)
+            decrRefCount(o);
     }
     dictReleaseIterator(di);
 
@@ -301,7 +309,7 @@ bool redisDbPersistentDataSnapshot::iterate_threadsafe(std::function<bool(const
             // Alright it's a key in the use keyspace, lets ensure it and then pass it off
             --celem;
             return fn(key, o);
-        });
+        }, fKeyOnly);
     }
 
     serverAssert(!fResult || celem == 0);
@@ -354,7 +362,7 @@ void redisDbPersistentDataSnapshot::consolidate_children(redisDbPersistentData *
             incrRefCount(o);
         dictAdd(spdb->m_pdict, sdsdup(key), o.unsafe_robjcast());
         return true;
-    });
+    }, true /*fKeyOnly*/);
     spdb->m_spstorage = m_pdbSnapshot->m_spstorage;
 
     spdb->m_pdict->iterators++;

src/storage/rocksdb.cpp

@@ -1,62 +1,11 @@
-#include "../IStorage.h"
-#include <rocksdb/db.h>
+#include "rocksdb.h"
 #include <string>
 #include <sstream>
 
-class RocksDBStorageProvider : public IStorage
-{
-    std::shared_ptr<rocksdb::DB> m_spdb;
-    std::unique_ptr<rocksdb::WriteBatch> m_spbatch;
-    const rocksdb::Snapshot *m_psnapshot = nullptr;
-    rocksdb::ReadOptions m_readOptionsTemplate;
-
-public:
-    RocksDBStorageProvider(const char *path);
-    ~RocksDBStorageProvider();
-
-    virtual void insert(const char *key, size_t cchKey, void *data, size_t cb) override;
-    virtual void erase(const char *key, size_t cchKey) override;
-    virtual void retrieve(const char *key, size_t cchKey, callback fn) const override;
-    virtual size_t clear() override;
-    virtual void enumerate(callback fn) const override;
-
-    virtual const IStorage *clone() const override;
-
-    virtual void beginWriteBatch() override;
-    virtual void endWriteBatch() override;
-
-    size_t count() const;
-
-protected:
-    RocksDBStorageProvider(std::shared_ptr<rocksdb::DB> &spdb);
-
-    const rocksdb::ReadOptions &ReadOptions() const { return m_readOptionsTemplate; }
-    rocksdb::WriteOptions WriteOptions() const;
-};
-
-IStorage *create_rocksdb_storage(const char *dbfile)
-{
-    return new RocksDBStorageProvider(dbfile);
-}
-
-RocksDBStorageProvider::RocksDBStorageProvider(const char *path)
-{
-    rocksdb::Options options;
-    options.create_if_missing = true;
-    rocksdb::DB *db = nullptr;
-    auto status = rocksdb::DB::Open(options, path, &db);
-    if (!status.ok())
-        throw status.ToString();
-    m_spdb = std::shared_ptr<rocksdb::DB>(db);
-
-    m_readOptionsTemplate = rocksdb::ReadOptions();
-}
-
-RocksDBStorageProvider::RocksDBStorageProvider(std::shared_ptr<rocksdb::DB> &spdb)
-    : m_spdb(spdb)
+RocksDBStorageProvider::RocksDBStorageProvider(std::shared_ptr<rocksdb::DB> &spdb, std::shared_ptr<rocksdb::ColumnFamilyHandle> &spcolfam, const rocksdb::Snapshot *psnapshot)
+    : m_spdb(spdb), m_psnapshot(psnapshot), m_spcolfamily(spcolfam)
 {
     m_readOptionsTemplate = rocksdb::ReadOptions();
-    m_psnapshot = spdb->GetSnapshot();
     m_readOptionsTemplate.snapshot = m_psnapshot;
 }
 
@@ -64,9 +13,9 @@ void RocksDBStorageProvider::insert(const char *key, size_t cchKey, void *data,
 {
     rocksdb::Status status;
     if (m_spbatch != nullptr)
-        status = m_spbatch->Put(rocksdb::Slice(key, cchKey), rocksdb::Slice((const char*)data, cb));
+        status = m_spbatch->Put(m_spcolfamily.get(), rocksdb::Slice(key, cchKey), rocksdb::Slice((const char*)data, cb));
     else
-        status = m_spdb->Put(WriteOptions(), rocksdb::Slice(key, cchKey), rocksdb::Slice((const char*)data, cb));
+        status = m_spdb->Put(WriteOptions(), m_spcolfamily.get(), rocksdb::Slice(key, cchKey), rocksdb::Slice((const char*)data, cb));
     if (!status.ok())
         throw status.ToString();
 }
@@ -75,9 +24,9 @@ void RocksDBStorageProvider::erase(const char *key, size_t cchKey)
 {
     rocksdb::Status status;
     if (m_spbatch != nullptr)
-        status = m_spbatch->Delete(rocksdb::Slice(key, cchKey));
+        status = m_spbatch->Delete(m_spcolfamily.get(), rocksdb::Slice(key, cchKey));
     else
-        status = m_spdb->Delete(WriteOptions(), rocksdb::Slice(key, cchKey));
+        status = m_spdb->Delete(WriteOptions(), m_spcolfamily.get(), rocksdb::Slice(key, cchKey));
     if (!status.ok())
         throw status.ToString();
 }
@@ -85,7 +34,7 @@ void RocksDBStorageProvider::erase(const char *key, size_t cchKey)
 void RocksDBStorageProvider::retrieve(const char *key, size_t cchKey, callback fn) const
 {
     std::string value;
-    auto status = m_spdb->Get(ReadOptions(), rocksdb::Slice(key, cchKey), &value);
+    auto status = m_spdb->Get(ReadOptions(), m_spcolfamily.get(), rocksdb::Slice(key, cchKey), &value);
     if (!status.ok())
         throw status.ToString();
     fn(key, cchKey, value.data(), value.size());
@@ -94,7 +43,13 @@ void RocksDBStorageProvider::retrieve(const char *key, size_t cchKey, callback f
 size_t RocksDBStorageProvider::clear()
 {
     size_t celem = count();
-    auto status = m_spdb->DropColumnFamily(m_spdb->DefaultColumnFamily());
+    auto status = m_spdb->DropColumnFamily(m_spcolfamily.get());
+    auto strName = m_spcolfamily->GetName();
+
+    rocksdb::ColumnFamilyHandle *handle = nullptr;
+    m_spdb->CreateColumnFamily(rocksdb::ColumnFamilyOptions(), strName, &handle);
+    m_spcolfamily = std::shared_ptr<rocksdb::ColumnFamilyHandle>(handle);
+
     if (!status.ok())
         throw status.ToString();
     return celem;
@@ -103,7 +58,7 @@ size_t RocksDBStorageProvider::clear()
 size_t RocksDBStorageProvider::count() const
 {
     std::string strelem;
-    if (!m_spdb->GetProperty(rocksdb::DB::Properties::kEstimateNumKeys, &strelem))
+    if (!m_spdb->GetProperty(m_spcolfamily.get(), rocksdb::DB::Properties::kEstimateNumKeys, &strelem))
         throw "Failed to get database size";
     std::stringstream sstream(strelem);
     size_t count;
@@ -113,7 +68,7 @@ size_t RocksDBStorageProvider::count() const
 
 void RocksDBStorageProvider::enumerate(callback fn) const
 {
-    std::unique_ptr<rocksdb::Iterator> it = std::unique_ptr<rocksdb::Iterator>(m_spdb->NewIterator(ReadOptions()));
+    std::unique_ptr<rocksdb::Iterator> it = std::unique_ptr<rocksdb::Iterator>(m_spdb->NewIterator(ReadOptions(), m_spcolfamily.get()));
     for (it->SeekToFirst(); it->Valid(); it->Next()) {
         fn(it->key().data(), it->key().size(), it->value().data(), it->value().size());
     }
@@ -122,7 +77,8 @@ void RocksDBStorageProvider::enumerate(callback fn) const
 
 const IStorage *RocksDBStorageProvider::clone() const
 {
-    return new RocksDBStorageProvider(const_cast<RocksDBStorageProvider*>(this)->m_spdb);
+    const rocksdb::Snapshot *psnapshot = const_cast<RocksDBStorageProvider*>(this)->m_spdb->GetSnapshot();
+    return new RocksDBStorageProvider(const_cast<RocksDBStorageProvider*>(this)->m_spdb, const_cast<RocksDBStorageProvider*>(this)->m_spcolfamily, psnapshot);
 }
 
 RocksDBStorageProvider::~RocksDBStorageProvider()
@@ -150,4 +106,9 @@ void RocksDBStorageProvider::endWriteBatch()
 {
     m_spdb->Write(WriteOptions(), m_spbatch.get());
     m_spbatch = nullptr;
+}
+
+void RocksDBStorageProvider::flush()
+{
+    m_spdb->SyncWAL();
 }

src/storage/rocksdb.h

@@ -0,0 +1,38 @@
+#pragma once
+
+#include <memory>
+#include "../IStorage.h"
+#include <rocksdb/db.h>
+
+class RocksDBStorageProvider : public IStorage
+{
+    std::shared_ptr<rocksdb::DB> m_spdb;    // Note: This must be first so it is deleted last
+    std::unique_ptr<rocksdb::WriteBatch> m_spbatch;
+    const rocksdb::Snapshot *m_psnapshot = nullptr;
+    std::shared_ptr<rocksdb::ColumnFamilyHandle> m_spcolfamily;
+    rocksdb::ReadOptions m_readOptionsTemplate;
+
+public:
+    RocksDBStorageProvider(std::shared_ptr<rocksdb::DB> &spdb, std::shared_ptr<rocksdb::ColumnFamilyHandle> &spcolfam, const rocksdb::Snapshot *psnapshot);
+    ~RocksDBStorageProvider();
+
+    virtual void insert(const char *key, size_t cchKey, void *data, size_t cb) override;
+    virtual void erase(const char *key, size_t cchKey) override;
+    virtual void retrieve(const char *key, size_t cchKey, callback fn) const override;
+    virtual size_t clear() override;
+    virtual void enumerate(callback fn) const override;
+
+    virtual const IStorage *clone() const override;
+
+    virtual void beginWriteBatch() override;
+    virtual void endWriteBatch() override;
+
+    virtual void flush() override;
+
+    size_t count() const;
+
+protected:
+
+    const rocksdb::ReadOptions &ReadOptions() const { return m_readOptionsTemplate; }
+    rocksdb::WriteOptions WriteOptions() const;
+};

src/storage/rocksdbfactory.cpp

@@ -0,0 +1,63 @@
+#include "rocksdb.h"
+
+class RocksDBStorageFactory : public IStorageFactory
+{
+    std::shared_ptr<rocksdb::DB> m_spdb;    // Note: This must be first so it is deleted last
+    std::vector<std::unique_ptr<rocksdb::ColumnFamilyHandle>> m_vecspcols;
+
+public:
+    RocksDBStorageFactory(const char *dbfile, int dbnum);
+    ~RocksDBStorageFactory();
+
+    virtual IStorage *create(int db) override;
+};
+
+IStorageFactory *CreateRocksDBStorageFactory(const char *path, int dbnum)
+{
+    return new RocksDBStorageFactory(path, dbnum);
+}
+
+RocksDBStorageFactory::RocksDBStorageFactory(const char *dbfile, int dbnum)
+{
+    // Get the count of column families in the actual database
+    std::vector<std::string> vecT;
+    auto status = rocksdb::DB::ListColumnFamilies(rocksdb::Options(), dbfile, &vecT);
+    // RocksDB requires we know the count of col families before opening, if the user only wants to see less
+    //  we still have to make room for all column family handles regardless
+    if (status.ok() && (int)vecT.size() > dbnum)
+        dbnum = (int)vecT.size();
+
+    std::vector<rocksdb::ColumnFamilyDescriptor> veccoldesc;
+    veccoldesc.push_back(rocksdb::ColumnFamilyDescriptor(rocksdb::kDefaultColumnFamilyName, rocksdb::ColumnFamilyOptions()));  // ignore default col family
+    
+    for (int idb = 0; idb < dbnum; ++idb)
+    {
+        veccoldesc.push_back(rocksdb::ColumnFamilyDescriptor(std::to_string(idb), rocksdb::ColumnFamilyOptions()));
+    }
+
+    rocksdb::Options options;
+    options.create_if_missing = true;
+    options.create_missing_column_families = true;
+    rocksdb::DB *db = nullptr;
+    
+    std::vector<rocksdb::ColumnFamilyHandle*> handles;
+    status = rocksdb::DB::Open(options, dbfile, veccoldesc, &handles, &db);
+    if (!status.ok())
+        throw status.ToString();
+
+    for (auto handle : handles)
+        m_vecspcols.emplace_back(handle);
+    m_spdb = std::shared_ptr<rocksdb::DB>(db);
+}
+
+RocksDBStorageFactory::~RocksDBStorageFactory()
+{
+    m_spdb->SyncWAL();
+}
+
+IStorage *RocksDBStorageFactory::create(int db)
+{
+    ++db;   // skip default col family
+    std::shared_ptr<rocksdb::ColumnFamilyHandle> spcolfamily(m_vecspcols[db].release());
+    return new RocksDBStorageProvider(m_spdb, spcolfamily, nullptr);
+}

src/storage/rocksdbfactory.h

@@ -0,0 +1,3 @@
+#pragma once
+
+class IStorageFactory *CreateRocksDBStorageFactory(const char *path, int dbnum);