Support for read-only slaves. Semantical fixes.

[redis.git] / src / db.c

diff --git a/src/db.c b/src/db.c
index 47994df20eb7bcfb6fce90f890cbc163bfb401b3..7ca2a5f66f76cee13a2d24c85b5be5f553edf308 100644 (file)
--- a/src/db.c
+++ b/src/db.c
@@ -1,6 +1,7 @@
 #include "redis.h"
 
 #include <signal.h>
 #include "redis.h"
 
 #include <signal.h>

+#include <ctype.h>

 
 void SlotToKeyAdd(robj *key);
 void SlotToKeyDel(robj *key);
 
 void SlotToKeyAdd(robj *key);
 void SlotToKeyDel(robj *key);


@@ -34,24 +35,29 @@ void SlotToKeyDel(robj *key);
 robj *lookupKey(redisDb *db, robj *key) {
     dictEntry *de = dictFind(db->dict,key->ptr);
     if (de) {
 robj *lookupKey(redisDb *db, robj *key) {
     dictEntry *de = dictFind(db->dict,key->ptr);
     if (de) {

-        robj *val = dictGetEntryVal(de);
+        robj *val = dictGetVal(de);

 
         /* Update the access time for the aging algorithm.
          * Don't do it if we have a saving child, as this will trigger
          * a copy on write madness. */
 
         /* Update the access time for the aging algorithm.
          * Don't do it if we have a saving child, as this will trigger
          * a copy on write madness. */

-        if (server.bgsavechildpid == -1 && server.bgrewritechildpid == -1)
+        if (server.rdb_child_pid == -1 && server.aof_child_pid == -1)

             val->lru = server.lruclock;
             val->lru = server.lruclock;

-        server.stat_keyspace_hits++;

         return val;
     } else {
         return val;
     } else {

-        server.stat_keyspace_misses++;

         return NULL;
     }
 }
 
 robj *lookupKeyRead(redisDb *db, robj *key) {
         return NULL;
     }
 }
 
 robj *lookupKeyRead(redisDb *db, robj *key) {

+    robj *val;
+

     expireIfNeeded(db,key);
     expireIfNeeded(db,key);

-    return lookupKey(db,key);
+    val = lookupKey(db,key);
+    if (val == NULL)
+        server.stat_keyspace_misses++;
+    else
+        server.stat_keyspace_hits++;
+    return val;

 }
 
 robj *lookupKeyWrite(redisDb *db, robj *key) {
 }
 
 robj *lookupKeyWrite(redisDb *db, robj *key) {


@@ -79,8 +85,7 @@ void dbAdd(redisDb *db, robj *key, robj *val) {
     sds copy = sdsdup(key->ptr);
     int retval = dictAdd(db->dict, copy, val);
 
     sds copy = sdsdup(key->ptr);
     int retval = dictAdd(db->dict, copy, val);
 

-    redisAssert(retval == REDIS_OK);
-    if (server.cluster_enabled) SlotToKeyAdd(key);
+    redisAssertWithInfo(NULL,key,retval == REDIS_OK);

  }
 
 /* Overwrite an existing key with a new value. Incrementing the reference
  }
 
 /* Overwrite an existing key with a new value. Incrementing the reference


@@ -91,7 +96,7 @@ void dbAdd(redisDb *db, robj *key, robj *val) {
 void dbOverwrite(redisDb *db, robj *key, robj *val) {
     struct dictEntry *de = dictFind(db->dict,key->ptr);
     
 void dbOverwrite(redisDb *db, robj *key, robj *val) {
     struct dictEntry *de = dictFind(db->dict,key->ptr);
     

-    redisAssert(de != NULL);
+    redisAssertWithInfo(NULL,key,de != NULL);

     dictReplace(db->dict, key->ptr, val);
 }
 
     dictReplace(db->dict, key->ptr, val);
 }
 


@@ -109,7 +114,7 @@ void setKey(redisDb *db, robj *key, robj *val) {
     }
     incrRefCount(val);
     removeExpire(db,key);
     }
     incrRefCount(val);
     removeExpire(db,key);

-    touchWatchedKey(db,key);
+    signalModifiedKey(db,key);

 }
 
 int dbExists(redisDb *db, robj *key) {
 }
 
 int dbExists(redisDb *db, robj *key) {


@@ -130,7 +135,7 @@ robj *dbRandomKey(redisDb *db) {
         de = dictGetRandomKey(db->dict);
         if (de == NULL) return NULL;
 
         de = dictGetRandomKey(db->dict);
         if (de == NULL) return NULL;
 

-        key = dictGetEntryKey(de);
+        key = dictGetKey(de);

         keyobj = createStringObject(key,sdslen(key));
         if (dictFind(db->expires,key)) {
             if (expireIfNeeded(db,keyobj)) {
         keyobj = createStringObject(key,sdslen(key));
         if (dictFind(db->expires,key)) {
             if (expireIfNeeded(db,keyobj)) {


@@ -148,7 +153,6 @@ int dbDelete(redisDb *db, robj *key) {
      * the key, because it is shared with the main dictionary. */
     if (dictSize(db->expires) > 0) dictDelete(db->expires,key->ptr);
     if (dictDelete(db->dict,key->ptr) == DICT_OK) {
      * the key, because it is shared with the main dictionary. */
     if (dictSize(db->expires) > 0) dictDelete(db->expires,key->ptr);
     if (dictDelete(db->dict,key->ptr) == DICT_OK) {

-        if (server.cluster_enabled) SlotToKeyDel(key);

         return 1;
     } else {
         return 0;
         return 1;
     } else {
         return 0;


@@ -209,11 +213,17 @@ void flushallCommand(redisClient *c) {
     signalFlushedDb(-1);
     server.dirty += emptyDb();
     addReply(c,shared.ok);
     signalFlushedDb(-1);
     server.dirty += emptyDb();
     addReply(c,shared.ok);

-    if (server.bgsavechildpid != -1) {
-        kill(server.bgsavechildpid,SIGKILL);
-        rdbRemoveTempFile(server.bgsavechildpid);
+    if (server.rdb_child_pid != -1) {
+        kill(server.rdb_child_pid,SIGKILL);
+        rdbRemoveTempFile(server.rdb_child_pid);
+    }
+    if (server.saveparamslen > 0) {
+        /* Normally rdbSave() will reset dirty, but we don't want this here
+         * as otherwise FLUSHALL will not be replicated nor put into the AOF. */
+        int saved_dirty = server.dirty;
+        rdbSave(server.rdb_filename);
+        server.dirty = saved_dirty;

     }
     }

-    rdbSave(server.dbfilename);

     server.dirty++;
 }
 
     server.dirty++;
 }
 


@@ -242,10 +252,6 @@ void existsCommand(redisClient *c) {
 void selectCommand(redisClient *c) {
     int id = atoi(c->argv[1]->ptr);
 
 void selectCommand(redisClient *c) {
     int id = atoi(c->argv[1]->ptr);
 

-    if (server.cluster_enabled && id != 0) {
-        addReplyError(c,"SELECT is not allowed in cluster mode");
-        return;
-    }

     if (selectDb(c,id) == REDIS_ERR) {
         addReplyError(c,"invalid DB index");
     } else {
     if (selectDb(c,id) == REDIS_ERR) {
         addReplyError(c,"invalid DB index");
     } else {


@@ -276,7 +282,7 @@ void keysCommand(redisClient *c) {
     di = dictGetIterator(c->db->dict);
     allkeys = (pattern[0] == '*' && pattern[1] == '\0');
     while((de = dictNext(di)) != NULL) {
     di = dictGetIterator(c->db->dict);
     allkeys = (pattern[0] == '*' && pattern[1] == '\0');
     while((de = dictNext(di)) != NULL) {

-        sds key = dictGetEntryKey(de);
+        sds key = dictGetKey(de);

         robj *keyobj;
 
         if (allkeys || stringmatchlen(pattern,plen,key,sdslen(key),0)) {
         robj *keyobj;
 
         if (allkeys || stringmatchlen(pattern,plen,key,sdslen(key),0)) {


@@ -321,13 +327,28 @@ void typeCommand(redisClient *c) {
 }
 
 void shutdownCommand(redisClient *c) {
 }
 
 void shutdownCommand(redisClient *c) {

-    if (prepareForShutdown() == REDIS_OK)
-        exit(0);
+    int flags = 0;
+
+    if (c->argc > 2) {
+        addReply(c,shared.syntaxerr);
+        return;
+    } else if (c->argc == 2) {
+        if (!strcasecmp(c->argv[1]->ptr,"nosave")) {
+            flags |= REDIS_SHUTDOWN_NOSAVE;
+        } else if (!strcasecmp(c->argv[1]->ptr,"save")) {
+            flags |= REDIS_SHUTDOWN_SAVE;
+        } else {
+            addReply(c,shared.syntaxerr);
+            return;
+        }
+    }
+    if (prepareForShutdown(flags) == REDIS_OK) exit(0);

     addReplyError(c,"Errors trying to SHUTDOWN. Check logs.");
 }
 
 void renameGenericCommand(redisClient *c, int nx) {
     robj *o;
     addReplyError(c,"Errors trying to SHUTDOWN. Check logs.");
 }
 
 void renameGenericCommand(redisClient *c, int nx) {
     robj *o;

+    long long expire;

 
     /* To use the same key as src and dst is probably an error */
     if (sdscmp(c->argv[1]->ptr,c->argv[2]->ptr) == 0) {
 
     /* To use the same key as src and dst is probably an error */
     if (sdscmp(c->argv[1]->ptr,c->argv[2]->ptr) == 0) {


@@ -339,16 +360,18 @@ void renameGenericCommand(redisClient *c, int nx) {
         return;
 
     incrRefCount(o);
         return;
 
     incrRefCount(o);

+    expire = getExpire(c->db,c->argv[1]);

     if (lookupKeyWrite(c->db,c->argv[2]) != NULL) {
         if (nx) {
             decrRefCount(o);
             addReply(c,shared.czero);
             return;
         }
     if (lookupKeyWrite(c->db,c->argv[2]) != NULL) {
         if (nx) {
             decrRefCount(o);
             addReply(c,shared.czero);
             return;
         }

-        dbOverwrite(c->db,c->argv[2],o);
-    } else {
-        dbAdd(c->db,c->argv[2],o);
+        /* Overwrite: delete the old key before creating the new one with the same name. */
+        dbDelete(c->db,c->argv[2]);

     }
     }

+    dbAdd(c->db,c->argv[2],o);
+    if (expire != -1) setExpire(c->db,c->argv[2],expire);

     dbDelete(c->db,c->argv[1]);
     signalModifiedKey(c->db,c->argv[1]);
     signalModifiedKey(c->db,c->argv[2]);
     dbDelete(c->db,c->argv[1]);
     signalModifiedKey(c->db,c->argv[1]);
     signalModifiedKey(c->db,c->argv[2]);


@@ -369,11 +392,6 @@ void moveCommand(redisClient *c) {
     redisDb *src, *dst;
     int srcid;
 
     redisDb *src, *dst;
     int srcid;
 

-    if (server.cluster_enabled) {
-        addReplyError(c,"MOVE is not allowed in cluster mode");
-        return;
-    }
-

     /* Obtain source and target DB pointers */
     src = c->db;
     srcid = c->db->id;
     /* Obtain source and target DB pointers */
     src = c->db;
     srcid = c->db->id;


@@ -419,22 +437,23 @@ void moveCommand(redisClient *c) {
 int removeExpire(redisDb *db, robj *key) {
     /* An expire may only be removed if there is a corresponding entry in the
      * main dict. Otherwise, the key will never be freed. */
 int removeExpire(redisDb *db, robj *key) {
     /* An expire may only be removed if there is a corresponding entry in the
      * main dict. Otherwise, the key will never be freed. */

-    redisAssert(dictFind(db->dict,key->ptr) != NULL);
+    redisAssertWithInfo(NULL,key,dictFind(db->dict,key->ptr) != NULL);

     return dictDelete(db->expires,key->ptr) == DICT_OK;
 }
 
     return dictDelete(db->expires,key->ptr) == DICT_OK;
 }
 

-void setExpire(redisDb *db, robj *key, time_t when) {
-    dictEntry *de;
+void setExpire(redisDb *db, robj *key, long long when) {
+    dictEntry *kde, *de;

 
     /* Reuse the sds from the main dict in the expire dict */
 
     /* Reuse the sds from the main dict in the expire dict */

-    de = dictFind(db->dict,key->ptr);
-    redisAssert(de != NULL);
-    dictReplace(db->expires,dictGetEntryKey(de),(void*)when);
+    kde = dictFind(db->dict,key->ptr);
+    redisAssertWithInfo(NULL,key,kde != NULL);
+    de = dictReplaceRaw(db->expires,dictGetKey(kde));
+    dictSetSignedIntegerVal(de,when);

 }
 
 /* Return the expire time of the specified key, or -1 if no expire
  * is associated with this key (i.e. the key is non volatile) */
 }
 
 /* Return the expire time of the specified key, or -1 if no expire
  * is associated with this key (i.e. the key is non volatile) */

-time_t getExpire(redisDb *db, robj *key) {
+long long getExpire(redisDb *db, robj *key) {

     dictEntry *de;
 
     /* No expire? return ASAP */
     dictEntry *de;
 
     /* No expire? return ASAP */


@@ -443,8 +462,8 @@ time_t getExpire(redisDb *db, robj *key) {
 
     /* The entry was found in the expire dict, this means it should also
      * be present in the main dict (safety check). */
 
     /* The entry was found in the expire dict, this means it should also
      * be present in the main dict (safety check). */

-    redisAssert(dictFind(db->dict,key->ptr) != NULL);
-    return (time_t) dictGetEntryVal(de);
+    redisAssertWithInfo(NULL,key,dictFind(db->dict,key->ptr) != NULL);
+    return dictGetSignedIntegerVal(de);

 }
 
 /* Propagate expires into slaves and the AOF file.
 }
 
 /* Propagate expires into slaves and the AOF file.


@@ -458,11 +477,12 @@ time_t getExpire(redisDb *db, robj *key) {
 void propagateExpire(redisDb *db, robj *key) {
     robj *argv[2];
 
 void propagateExpire(redisDb *db, robj *key) {
     robj *argv[2];
 

-    argv[0] = createStringObject("DEL",3);
+    argv[0] = shared.del;

     argv[1] = key;
     argv[1] = key;

-    incrRefCount(key);
+    incrRefCount(argv[0]);
+    incrRefCount(argv[1]);

 
 

-    if (server.appendonly)
+    if (server.aof_state != REDIS_AOF_OFF)

         feedAppendOnlyFile(server.delCommand,db->id,argv,2);
     if (listLength(server.slaves))
         replicationFeedSlaves(server.slaves,db->id,argv,2);
         feedAppendOnlyFile(server.delCommand,db->id,argv,2);
     if (listLength(server.slaves))
         replicationFeedSlaves(server.slaves,db->id,argv,2);


@@ -472,7 +492,7 @@ void propagateExpire(redisDb *db, robj *key) {
 }
 
 int expireIfNeeded(redisDb *db, robj *key) {
 }
 
 int expireIfNeeded(redisDb *db, robj *key) {

-    time_t when = getExpire(db,key);
+    long long when = getExpire(db,key);

 
     if (when < 0) return 0; /* No expire for this key */
 
 
     if (when < 0) return 0; /* No expire for this key */
 


@@ -491,7 +511,7 @@ int expireIfNeeded(redisDb *db, robj *key) {
     }
 
     /* Return when this key has not expired */
     }
 
     /* Return when this key has not expired */

-    if (time(NULL) <= when) return 0;
+    if (mstime() <= when) return 0;

 
     /* Delete the key */
     server.stat_expiredkeys++;
 
     /* Delete the key */
     server.stat_expiredkeys++;


@@ -503,26 +523,51 @@ int expireIfNeeded(redisDb *db, robj *key) {
  * Expires Commands
  *----------------------------------------------------------------------------*/
 
  * Expires Commands
  *----------------------------------------------------------------------------*/
 

-void expireGenericCommand(redisClient *c, robj *key, robj *param, long offset) {
+/* Given an string object return true if it contains exactly the "ms"
+ * or "MS" string. This is used in order to check if the last argument
+ * of EXPIRE, EXPIREAT or TTL is "ms" to switch into millisecond input/output */
+int stringObjectEqualsMs(robj *a) {
+    char *arg = a->ptr;
+    return tolower(arg[0]) == 'm' && tolower(arg[1]) == 's' && arg[2] == '\0';
+}
+
+void expireGenericCommand(redisClient *c, long long offset, int unit) {

     dictEntry *de;
     dictEntry *de;

-    long seconds;
+    robj *key = c->argv[1], *param = c->argv[2];
+    long long milliseconds;

 
 

-    if (getLongFromObjectOrReply(c, param, &seconds, NULL) != REDIS_OK) return;
+    if (getLongLongFromObjectOrReply(c, param, &milliseconds, NULL) != REDIS_OK)
+        return;

 
 

-    seconds -= offset;
+    if (unit == UNIT_SECONDS) milliseconds *= 1000;
+    milliseconds -= offset;

 
     de = dictFind(c->db->dict,key->ptr);
     if (de == NULL) {
         addReply(c,shared.czero);
         return;
     }
 
     de = dictFind(c->db->dict,key->ptr);
     if (de == NULL) {
         addReply(c,shared.czero);
         return;
     }

-    if (seconds <= 0 && !server.loading) {
-        if (dbDelete(c->db,key)) server.dirty++;
-        addReply(c, shared.cone);
+    /* EXPIRE with negative TTL, or EXPIREAT with a timestamp into the past
+     * should never be executed as a DEL when load the AOF or in the context
+     * of a slave instance.
+     *
+     * Instead we take the other branch of the IF statement setting an expire
+     * (possibly in the past) and wait for an explicit DEL from the master. */
+    if (milliseconds <= 0 && !server.loading && !server.masterhost) {
+        robj *aux;
+
+        redisAssertWithInfo(c,key,dbDelete(c->db,key));
+        server.dirty++;
+
+        /* Replicate/AOF this as an explicit DEL. */
+        aux = createStringObject("DEL",3);
+        rewriteClientCommandVector(c,2,aux,key);
+        decrRefCount(aux);

         signalModifiedKey(c->db,key);
         signalModifiedKey(c->db,key);

+        addReply(c, shared.cone);

         return;
     } else {
         return;
     } else {

-        time_t when = time(NULL)+seconds;
+        long long when = mstime()+milliseconds;

         setExpire(c->db,key,when);
         addReply(c,shared.cone);
         signalModifiedKey(c->db,key);
         setExpire(c->db,key,when);
         addReply(c,shared.cone);
         signalModifiedKey(c->db,key);


@@ -532,22 +577,42 @@ void expireGenericCommand(redisClient *c, robj *key, robj *param, long offset) {
 }
 
 void expireCommand(redisClient *c) {
 }
 
 void expireCommand(redisClient *c) {

-    expireGenericCommand(c,c->argv[1],c->argv[2],0);
+    expireGenericCommand(c,0,UNIT_SECONDS);

 }
 
 void expireatCommand(redisClient *c) {
 }
 
 void expireatCommand(redisClient *c) {

-    expireGenericCommand(c,c->argv[1],c->argv[2],time(NULL));
+    expireGenericCommand(c,mstime(),UNIT_SECONDS);

 }
 
 }
 

-void ttlCommand(redisClient *c) {
-    time_t expire, ttl = -1;
+void pexpireCommand(redisClient *c) {
+    expireGenericCommand(c,0,UNIT_MILLISECONDS);
+}
+
+void pexpireatCommand(redisClient *c) {
+    expireGenericCommand(c,mstime(),UNIT_MILLISECONDS);
+}
+
+void ttlGenericCommand(redisClient *c, int output_ms) {
+    long long expire, ttl = -1;

 
     expire = getExpire(c->db,c->argv[1]);
     if (expire != -1) {
 
     expire = getExpire(c->db,c->argv[1]);
     if (expire != -1) {

-        ttl = (expire-time(NULL));
+        ttl = expire-mstime();

         if (ttl < 0) ttl = -1;
     }
         if (ttl < 0) ttl = -1;
     }

-    addReplyLongLong(c,(long long)ttl);
+    if (ttl == -1) {
+        addReplyLongLong(c,-1);
+    } else {
+        addReplyLongLong(c,output_ms ? ttl : ((ttl+500)/1000));
+    }
+}
+
+void ttlCommand(redisClient *c) {
+    ttlGenericCommand(c, 0);
+}
+
+void pttlCommand(redisClient *c) {
+    ttlGenericCommand(c, 1);

 }
 
 void persistCommand(redisClient *c) {
 }
 
 void persistCommand(redisClient *c) {


@@ -638,35 +703,3 @@ int *zunionInterGetKeys(struct redisCommand *cmd,robj **argv, int argc, int *num
     *numkeys = num;
     return keys;
 }
     *numkeys = num;
     return keys;
 }

-
-/* Slot to Key API. This is used by Redis Cluster in order to obtain in
- * a fast way a key that belongs to a specified hash slot. This is useful
- * while rehashing the cluster. */
-void SlotToKeyAdd(robj *key) {
-    unsigned int hashslot = keyHashSlot(key->ptr,sdslen(key->ptr));
-
-    zslInsert(server.cluster.slots_to_keys,hashslot,key);
-    incrRefCount(key);
-}
-
-void SlotToKeyDel(robj *key) {
-    unsigned int hashslot = keyHashSlot(key->ptr,sdslen(key->ptr));
-
-    zslDelete(server.cluster.slots_to_keys,hashslot,key);
-}
-
-unsigned int GetKeysInSlot(unsigned int hashslot, robj **keys, unsigned int count) {
-    zskiplistNode *n;
-    zrangespec range;
-    int j = 0;
-
-    range.min = range.max = hashslot;
-    range.minex = range.maxex = 0;
-    
-    n = zslFirstInRange(server.cluster.slots_to_keys, range);
-    while(n && n->score == hashslot && count--) {
-        keys[j++] = n->obj;
-        n = n->level[0].forward;
-    }
-    return j;
-}