Cluster branch merged to unstable.

11 files changed, 1727 insertions, 9 deletions

diff --git a/CLUSTER b/CLUSTER
new file mode 100644
index 000000000..bfe5234aa
--- /dev/null
+++ b/CLUSTER
@@ -0,0 +1,131 @@
+TODO
+
+- disconnect FAIL clients after some pong idle time.
+
+---------------------------------
+
+* Majority rule: the cluster con continue when there are all the hash slots covered AND when there are the majority of masters.
+* Shutdown on request rule: when a node sees many connections closed or even a timeout longer than usual on almost all the other nodes, it will usually wait for the normal timeout before to change the state, unless it receives a query from a client: in such a case it will put itself into error status.
+
+--------------------------------
+
+* When asked for a key that is not in a node's business it will reply:
+
+    -ASK 1.2.3.4:6379   (in case we want the client to ask just one time)
+    -MOVED <slotid> 1.2.3.4:6379  (in case the hash slot is permanently moved)
+
+So with -ASK a client should just retry the query against this new node, a single time.
+
+With -MOVED the client should update its hash slots table to reflect the fact that now the specified node is the one to contact for the specified hash slot.
+
+* Nodes communicate using a binary protocol.
+
+* Node failure detection.
+
+    1) Every node contains information about all the other nodes:
+        - If this node is believed to work ok or not
+        - The hash slots for which this node is responsible
+        - If the node is a master or a slave
+        - If it is a slave, the slave of which node
+        - if it is a master, the list of slave nodes
+        - The slaves are ordered for "<ip>:<port>" string from lower to higher
+          ordered lexicographically. When a master is down, the cluster will
+          try to elect the first slave in the list.
+
+    2) Every node also contains the unix time where every other node was
+       reported to work properly (that is, it replied to a ping or any other
+       protocol request correctly). For every node we also store the timestamp
+       at which we sent the latest ping, so we can easily compute the current
+       lag.
+
+    3) From time to time a node pings a random node, selected among the nodes
+       with the least recent "alive" time stamp. Three random nodes are selected
+       and the one with lower alive time stamp is pinged.
+
+    4) The ping packet contains also information about a few random nodes
+       alive time stamp. So that the receiver of the ping will update the
+       alive table if the received alive timestamp is more recent the
+       one present in the node local table.
+
+       In the ping packet every node "gossip" information is somethig like
+       this:
+
+       <ip>:<port>:<status>:<pingsent_timestamp>:<pongreceived_timestamp>
+
+       status is OK, POSSIBLE_FAILURE, FAILURE.
+
+    5) The node replies to ping with a pong packet, that also contains a random
+       selections of nodes timestamps.
+
+A given node thinks another node may be in a failure state once there is a
+ping timeout bigger than 30 seconds (configurable).
+
+When a possible failure is detected the node performs the following action:
+
+    1) Is the average between all the other nodes big? For instance bigger
+       than 30 seconds / 2 = 15 seconds? Probably *we* are disconnected.
+       In such a case we don't trust our lag data, and reset all the
+       timestamps of sent ping to zero. This way when we'll reconnect there
+       is no risk that we'll claim many nodes are down, taking inappropriate
+       actions.
+
+    2) Messages from nodes marked as failed are *always* ignored by the other
+       nodes. A new node needs to be "introduced" by a good online node.
+
+    3) If we are well connected (that is, condition "1" is not true) and a
+       node timeout is > 30 seconds, we mark the node as POSSIBLE_FAILURE
+       (a flat in the cluster node structure). Every time we sent a ping
+       to another node we inform this other nodes that we detected this
+       condition, as already stated.
+
+    4) Once a node receives a POSSIBLE_FAILURE status for a node that is
+       already marked as POSSIBLE_FAILURE locally, it sends a message
+       to all the other nodes of type NODE_FAILURE_DETECTED, communicating the
+       ip/port of the specified node.
+
+       All the nodes need to update the status of this node setting it into
+       FAILURE.
+
+    5) If the computer in FAILURE state is a master node, what is needed is
+       to perform a Slave Election.
+
+SLAVE ELECTION
+
+    1) The slave election is performed by the first slave (with slaves ordered
+       lexicographically). Actually it is the first functioning slave, so if
+       the first slave is marked as failing the next slave will perform the
+       election and so forth. Such a slave is called the "Successor".
+
+    2) The Successor starts checking that all the nodes in the cluster already
+       marked the master in FAILURE state. If at least one node does not agree
+       no action is performed.
+
+    3) If all the nodes agree that the master is failing, the Successor does
+       the following:
+
+       a) It will send a SUCCESSION message to all the other nodes, that will
+          upgrade the hash slot tables accordingly. It will make sure that all
+          the nodes are updated and if some node did not received the message
+          it will keep trying.
+       b) Once all nodes not marked as FAILURE accepted the SUCCESSION message
+          it will update his own table and will start acting as a master
+          accepting write queries.
+       c) Every node receiving the succession message, if not already informed
+          of the change will broadcast the same message to other three random
+          nodes. No action is performed if the specified host was already marked
+          as the master node.
+       d) A node that was a slave of the original master that failed will
+          switch master to the new one once the SUCCESSION message is received.
+
+RANDOM
+
+    1) When selecting a slave, the system will try to pick one with an IP different than the master and other slaves, if possible.
+
+    2) The PING packet also contains information about the local configuration checksum. This is the SHA1 of the current configuration, without the bits that normally change form one node to another (like latest ping reply, failure status of nodes, and so forth). From time to time the local config SHA1 is checked against the list of the other nodes, and if there is a mismatch between our configuration and the most common one that lasts for more than N seconds, the most common configuration is asked and retrieved from another node. The event is logged.
+
+    3) Every time a node updates its internal cluster configuration, it dumps such a config in the cluster.conf file. On startup the configuration is reloaded.
+    Nodes can share the cluster configuration when needed (for instance if SHA1 does not match) using this exact same format.
+
+CLIENTS
+
+    - Clients may be configured to use slaves to perform reads, when read-after-write consistency is not required.
diff --git a/src/Makefile b/src/Makefile
index acbc9b5f0..f99be4b9c 100644
--- a/src/Makefile
+++ b/src/Makefile
@@ -25,7 +25,7 @@ PREFIX= /usr/local
 INSTALL_BIN= $(PREFIX)/bin
 INSTALL= cp -p
 
-OBJ = adlist.o ae.o anet.o dict.o redis.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 dscache.o pubsub.o multi.o debug.o sort.o intset.o syncio.o diskstore.o endian.o
+OBJ = adlist.o ae.o anet.o dict.o redis.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 dscache.o pubsub.o multi.o debug.o sort.o intset.o syncio.o diskstore.o cluster.o crc16.o endian.o
 BENCHOBJ = ae.o anet.o redis-benchmark.o sds.o adlist.o zmalloc.o
 CLIOBJ = anet.o sds.o adlist.o redis-cli.o zmalloc.o release.o
 CHECKDUMPOBJ = redis-check-dump.o lzf_c.o lzf_d.o
@@ -105,6 +105,8 @@ t_zset.o: t_zset.c redis.h fmacros.h config.h ae.h sds.h dict.h adlist.h \
   zmalloc.h anet.h zipmap.h ziplist.h intset.h version.h
 util.o: util.c redis.h fmacros.h config.h ae.h sds.h dict.h adlist.h \
   zmalloc.h anet.h zipmap.h ziplist.h intset.h version.h
+cluster.o: redis.h fmacros.h config.h ae.h sds.h dict.h adlist.h \
+  zmalloc.h anet.h zipmap.h ziplist.h intset.h version.h
 ziplist.o: ziplist.c zmalloc.h ziplist.h
 zipmap.o: zipmap.c zmalloc.h
 zmalloc.o: zmalloc.c config.h
diff --git a/src/cluster.c b/src/cluster.c
new file mode 100644
index 000000000..aac52578b
--- /dev/null
+++ b/src/cluster.c
@@ -0,0 +1,1299 @@
+#include "redis.h"
+
+#include <arpa/inet.h>
+
+void clusterAcceptHandler(aeEventLoop *el, int fd, void *privdata, int mask);
+void clusterReadHandler(aeEventLoop *el, int fd, void *privdata, int mask);
+void clusterSendPing(clusterLink *link, int type);
+void clusterSendFail(char *nodename);
+void clusterUpdateState(void);
+int clusterNodeGetSlotBit(clusterNode *n, int slot);
+
+/* -----------------------------------------------------------------------------
+ * Initialization
+ * -------------------------------------------------------------------------- */
+
+void clusterGetRandomName(char *p) {
+    FILE *fp = fopen("/dev/urandom","r");
+    char *charset = "0123456789abcdef";
+    int j;
+
+    if (!fp) {
+        redisLog(REDIS_WARNING,
+            "Unrecovarable error: can't open /dev/urandom:%s" ,strerror(errno));
+        exit(1);
+    }
+    fread(p,REDIS_CLUSTER_NAMELEN,1,fp);
+    for (j = 0; j < REDIS_CLUSTER_NAMELEN; j++)
+        p[j] = charset[p[j] & 0x0F];
+    fclose(fp);
+}
+
+int clusterLoadConfig(char *filename) {
+    FILE *fp = fopen(filename,"r");
+    
+    if (fp == NULL) return REDIS_ERR;
+    fclose(fp);
+
+    redisLog(REDIS_NOTICE,"Node configuration loaded, I'm %.40s",
+        server.cluster.myself->name);
+    return REDIS_OK;
+
+fmterr:
+    redisLog(REDIS_WARNING,"Unrecovarable error: corrupted cluster.conf file.");
+    fclose(fp);
+    exit(1);
+}
+
+void clusterInit(void) {
+    server.cluster.myself = createClusterNode(NULL,REDIS_NODE_MYSELF);
+    server.cluster.state = REDIS_CLUSTER_FAIL;
+    server.cluster.nodes = dictCreate(&clusterNodesDictType,NULL);
+    server.cluster.node_timeout = 15;
+    memset(server.cluster.migrating_slots_to,0,
+        sizeof(server.cluster.migrating_slots_to));
+    memset(server.cluster.importing_slots_from,0,
+        sizeof(server.cluster.importing_slots_from));
+    memset(server.cluster.slots,0,
+        sizeof(server.cluster.slots));
+    if (clusterLoadConfig("cluster.conf") == REDIS_ERR) {
+        /* No configuration found. We will just use the random name provided
+         * by the createClusterNode() function. */
+        redisLog(REDIS_NOTICE,"No cluster configuration found, I'm %.40s",
+            server.cluster.myself->name);
+    }
+    clusterAddNode(server.cluster.myself);
+    /* We need a listening TCP port for our cluster messaging needs */
+    server.cfd = anetTcpServer(server.neterr,
+            server.port+REDIS_CLUSTER_PORT_INCR, server.bindaddr);
+    if (server.cfd == -1) {
+        redisLog(REDIS_WARNING, "Opening cluster TCP port: %s", server.neterr);
+        exit(1);
+    }
+    if (aeCreateFileEvent(server.el, server.cfd, AE_READABLE,
+        clusterAcceptHandler, NULL) == AE_ERR) oom("creating file event");
+}
+
+/* -----------------------------------------------------------------------------
+ * CLUSTER communication link
+ * -------------------------------------------------------------------------- */
+
+clusterLink *createClusterLink(clusterNode *node) {
+    clusterLink *link = zmalloc(sizeof(*link));
+    link->sndbuf = sdsempty();
+    link->rcvbuf = sdsempty();
+    link->node = node;
+    link->fd = -1;
+    return link;
+}
+
+/* Free a cluster link, but does not free the associated node of course.
+ * Just this function will make sure that the original node associated
+ * with this link will have the 'link' field set to NULL. */
+void freeClusterLink(clusterLink *link) {
+    if (link->fd != -1) {
+        aeDeleteFileEvent(server.el, link->fd, AE_WRITABLE);
+        aeDeleteFileEvent(server.el, link->fd, AE_READABLE);
+    }
+    sdsfree(link->sndbuf);
+    sdsfree(link->rcvbuf);
+    if (link->node)
+        link->node->link = NULL;
+    close(link->fd);
+    zfree(link);
+}
+
+void clusterAcceptHandler(aeEventLoop *el, int fd, void *privdata, int mask) {
+    int cport, cfd;
+    char cip[128];
+    clusterLink *link;
+    REDIS_NOTUSED(el);
+    REDIS_NOTUSED(mask);
+    REDIS_NOTUSED(privdata);
+
+    cfd = anetTcpAccept(server.neterr, fd, cip, &cport);
+    if (cfd == AE_ERR) {
+        redisLog(REDIS_VERBOSE,"Accepting cluster node: %s", server.neterr);
+        return;
+    }
+    redisLog(REDIS_VERBOSE,"Accepted cluster node %s:%d", cip, cport);
+    /* We need to create a temporary node in order to read the incoming
+     * packet in a valid contest. This node will be released once we
+     * read the packet and reply. */
+    link = createClusterLink(NULL);
+    link->fd = cfd;
+    aeCreateFileEvent(server.el,cfd,AE_READABLE,clusterReadHandler,link);
+}
+
+/* -----------------------------------------------------------------------------
+ * Key space handling
+ * -------------------------------------------------------------------------- */
+
+/* We have 4096 hash slots. The hash slot of a given key is obtained
+ * as the least significant 12 bits of the crc16 of the key. */
+unsigned int keyHashSlot(char *key, int keylen) {
+    return crc16(key,keylen) & 0x0FFF;
+}
+
+/* -----------------------------------------------------------------------------
+ * CLUSTER node API
+ * -------------------------------------------------------------------------- */
+
+/* Create a new cluster node, with the specified flags.
+ * If "nodename" is NULL this is considered a first handshake and a random
+ * node name is assigned to this node (it will be fixed later when we'll
+ * receive the first pong).
+ *
+ * The node is created and returned to the user, but it is not automatically
+ * added to the nodes hash table. */
+clusterNode *createClusterNode(char *nodename, int flags) {
+    clusterNode *node = zmalloc(sizeof(*node));
+
+    if (nodename)
+        memcpy(node->name, nodename, REDIS_CLUSTER_NAMELEN);
+    else
+        clusterGetRandomName(node->name);
+    node->flags = flags;
+    memset(node->slots,0,sizeof(node->slots));
+    node->numslaves = 0;
+    node->slaves = NULL;
+    node->slaveof = NULL;
+    node->ping_sent = node->pong_received = 0;
+    node->configdigest = NULL;
+    node->configdigest_ts = 0;
+    node->link = NULL;
+    return node;
+}
+
+int clusterNodeRemoveSlave(clusterNode *master, clusterNode *slave) {
+    int j;
+
+    for (j = 0; j < master->numslaves; j++) {
+        if (master->slaves[j] == slave) {
+            memmove(master->slaves+j,master->slaves+(j+1),
+                (master->numslaves-1)-j);
+            master->numslaves--;
+            return REDIS_OK;
+        }
+    }
+    return REDIS_ERR;
+}
+
+int clusterNodeAddSlave(clusterNode *master, clusterNode *slave) {
+    int j;
+
+    /* If it's already a slave, don't add it again. */
+    for (j = 0; j < master->numslaves; j++)
+        if (master->slaves[j] == slave) return REDIS_ERR;
+    master->slaves = zrealloc(master->slaves,
+        sizeof(clusterNode*)*(master->numslaves+1));
+    master->slaves[master->numslaves] = slave;
+    master->numslaves++;
+    return REDIS_OK;
+}
+
+void clusterNodeResetSlaves(clusterNode *n) {
+    zfree(n->slaves);
+    n->numslaves = 0;
+}
+
+void freeClusterNode(clusterNode *n) {
+    sds nodename;
+    
+    nodename = sdsnewlen(n->name, REDIS_CLUSTER_NAMELEN);
+    redisAssert(dictDelete(server.cluster.nodes,nodename) == DICT_OK);
+    sdsfree(nodename);
+    if (n->slaveof) clusterNodeRemoveSlave(n->slaveof, n);
+    if (n->link) freeClusterLink(n->link);
+    zfree(n);
+}
+
+/* Add a node to the nodes hash table */
+int clusterAddNode(clusterNode *node) {
+    int retval;
+    
+    retval = dictAdd(server.cluster.nodes,
+            sdsnewlen(node->name,REDIS_CLUSTER_NAMELEN), node);
+    return (retval == DICT_OK) ? REDIS_OK : REDIS_ERR;
+}
+
+/* Node lookup by name */
+clusterNode *clusterLookupNode(char *name) {
+    sds s = sdsnewlen(name, REDIS_CLUSTER_NAMELEN);
+    struct dictEntry *de;
+
+    de = dictFind(server.cluster.nodes,s);
+    sdsfree(s);
+    if (de == NULL) return NULL;
+    return dictGetEntryVal(de);
+}
+
+/* This is only used after the handshake. When we connect a given IP/PORT
+ * as a result of CLUSTER MEET we don't have the node name yet, so we
+ * pick a random one, and will fix it when we receive the PONG request using
+ * this function. */
+void clusterRenameNode(clusterNode *node, char *newname) {
+    int retval;
+    sds s = sdsnewlen(node->name, REDIS_CLUSTER_NAMELEN);
+   
+    redisLog(REDIS_DEBUG,"Renaming node %.40s into %.40s",
+        node->name, newname);
+    retval = dictDelete(server.cluster.nodes, s);
+    sdsfree(s);
+    redisAssert(retval == DICT_OK);
+    memcpy(node->name, newname, REDIS_CLUSTER_NAMELEN);
+    clusterAddNode(node);
+}
+
+/* -----------------------------------------------------------------------------
+ * CLUSTER messages exchange - PING/PONG and gossip
+ * -------------------------------------------------------------------------- */
+
+/* Process the gossip section of PING or PONG packets.
+ * Note that this function assumes that the packet is already sanity-checked
+ * by the caller, not in the content of the gossip section, but in the
+ * length. */
+void clusterProcessGossipSection(clusterMsg *hdr, clusterLink *link) {
+    uint16_t count = ntohs(hdr->count);
+    clusterMsgDataGossip *g = (clusterMsgDataGossip*) hdr->data.ping.gossip;
+    clusterNode *sender = link->node ? link->node : clusterLookupNode(hdr->sender);
+
+    while(count--) {
+        sds ci = sdsempty();
+        uint16_t flags = ntohs(g->flags);
+        clusterNode *node;
+
+        if (flags == 0) ci = sdscat(ci,"noflags,");
+        if (flags & REDIS_NODE_MYSELF) ci = sdscat(ci,"myself,");
+        if (flags & REDIS_NODE_MASTER) ci = sdscat(ci,"master,");
+        if (flags & REDIS_NODE_SLAVE) ci = sdscat(ci,"slave,");
+        if (flags & REDIS_NODE_PFAIL) ci = sdscat(ci,"fail?,");
+        if (flags & REDIS_NODE_FAIL) ci = sdscat(ci,"fail,");
+        if (flags & REDIS_NODE_HANDSHAKE) ci = sdscat(ci,"handshake,");
+        if (flags & REDIS_NODE_NOADDR) ci = sdscat(ci,"noaddr,");
+        if (ci[sdslen(ci)-1] == ',') ci[sdslen(ci)-1] = ' ';
+
+        redisLog(REDIS_DEBUG,"GOSSIP %.40s %s:%d %s",
+            g->nodename,
+            g->ip,
+            ntohs(g->port),
+            ci);
+        sdsfree(ci);
+
+        /* Update our state accordingly to the gossip sections */
+        node = clusterLookupNode(g->nodename);
+        if (node != NULL) {
+            /* We already know this node. Let's start updating the last
+             * time PONG figure if it is newer than our figure.
+             * Note that it's not a problem if we have a PING already 
+             * in progress against this node. */
+            if (node->pong_received < ntohl(g->pong_received)) {
+                 redisLog(REDIS_DEBUG,"Node pong_received updated by gossip");
+                node->pong_received = ntohl(g->pong_received);
+            }
+            /* Mark this node as FAILED if we think it is possibly failing
+             * and another node also thinks it's failing. */
+            if (node->flags & REDIS_NODE_PFAIL &&
+                (flags & (REDIS_NODE_FAIL|REDIS_NODE_PFAIL)))
+            {
+                redisLog(REDIS_NOTICE,"Received a PFAIL acknowledge from node %.40s, marking node %.40s as FAIL!", hdr->sender, node->name);
+                node->flags &= ~REDIS_NODE_PFAIL;
+                node->flags |= REDIS_NODE_FAIL;
+                /* Broadcast the failing node name to everybody */
+                clusterSendFail(node->name);
+                clusterUpdateState();
+            }
+        } else {
+            /* If it's not in NOADDR state and we don't have it, we
+             * start an handshake process against this IP/PORT pairs.
+             *
+             * Note that we require that the sender of this gossip message
+             * is a well known node in our cluster, otherwise we risk
+             * joining another cluster. */
+            if (sender && !(flags & REDIS_NODE_NOADDR)) {
+                clusterNode *newnode;
+
+                redisLog(REDIS_DEBUG,"Adding the new node");
+                newnode = createClusterNode(NULL,REDIS_NODE_HANDSHAKE);
+                memcpy(newnode->ip,g->ip,sizeof(g->ip));
+                newnode->port = ntohs(g->port);
+                clusterAddNode(newnode);
+            }
+        }
+
+        /* Next node */
+        g++;
+    }
+}
+
+/* IP -> string conversion. 'buf' is supposed to at least be 16 bytes. */
+void nodeIp2String(char *buf, clusterLink *link) {
+    struct sockaddr_in sa;
+    socklen_t salen = sizeof(sa);
+
+    if (getpeername(link->fd, (struct sockaddr*) &sa, &salen) == -1)
+        redisPanic("getpeername() failed.");
+    strncpy(buf,inet_ntoa(sa.sin_addr),sizeof(link->node->ip));
+}
+
+
+/* Update the node address to the IP address that can be extracted
+ * from link->fd, and at the specified port. */
+void nodeUpdateAddress(clusterNode *node, clusterLink *link, int port) {
+}
+
+/* When this function is called, there is a packet to process starting
+ * at node->rcvbuf. Releasing the buffer is up to the caller, so this
+ * function should just handle the higher level stuff of processing the
+ * packet, modifying the cluster state if needed.
+ *
+ * The function returns 1 if the link is still valid after the packet
+ * was processed, otherwise 0 if the link was freed since the packet
+ * processing lead to some inconsistency error (for instance a PONG
+ * received from the wrong sender ID). */
+int clusterProcessPacket(clusterLink *link) {
+    clusterMsg *hdr = (clusterMsg*) link->rcvbuf;
+    uint32_t totlen = ntohl(hdr->totlen);
+    uint16_t type = ntohs(hdr->type);
+    clusterNode *sender;
+
+    redisLog(REDIS_DEBUG,"--- packet to process %lu bytes (%lu) ---",
+        (unsigned long) totlen, sdslen(link->rcvbuf));
+    if (totlen < 8) return 1;
+    if (totlen > sdslen(link->rcvbuf)) return 1;
+    if (type == CLUSTERMSG_TYPE_PING || type == CLUSTERMSG_TYPE_PONG ||
+        type == CLUSTERMSG_TYPE_MEET)
+    {
+        uint16_t count = ntohs(hdr->count);
+        uint32_t explen; /* expected length of this packet */
+
+        explen = sizeof(clusterMsg)-sizeof(union clusterMsgData);
+        explen += (sizeof(clusterMsgDataGossip)*count);
+        if (totlen != explen) return 1;
+    }
+    if (type == CLUSTERMSG_TYPE_FAIL) {
+        uint32_t explen = sizeof(clusterMsg)-sizeof(union clusterMsgData);
+
+        explen += sizeof(clusterMsgDataFail);
+        if (totlen != explen) return 1;
+    }
+
+    sender = clusterLookupNode(hdr->sender);
+    if (type == CLUSTERMSG_TYPE_PING || type == CLUSTERMSG_TYPE_MEET) {
+        redisLog(REDIS_DEBUG,"Ping packet received: %p", link->node);
+
+        /* Add this node if it is new for us and the msg type is MEET.
+         * In this stage we don't try to add the node with the right
+         * flags, slaveof pointer, and so forth, as this details will be
+         * resolved when we'll receive PONGs from the server. */
+        if (!sender && type == CLUSTERMSG_TYPE_MEET) {
+            clusterNode *node;
+
+            node = createClusterNode(NULL,REDIS_NODE_HANDSHAKE);
+            nodeIp2String(node->ip,link);
+            node->port = ntohs(hdr->port);
+            clusterAddNode(node);
+        }
+
+        /* Get info from the gossip section */
+        clusterProcessGossipSection(hdr,link);
+
+        /* Anyway reply with a PONG */
+        clusterSendPing(link,CLUSTERMSG_TYPE_PONG);
+    } else if (type == CLUSTERMSG_TYPE_PONG) {
+        int update = 0;
+
+        redisLog(REDIS_DEBUG,"Pong packet received: %p", link->node);
+        if (link->node) {
+            if (link->node->flags & REDIS_NODE_HANDSHAKE) {
+                /* If we already have this node, try to change the
+                 * IP/port of the node with the new one. */
+                if (sender) {
+                    redisLog(REDIS_WARNING,
+                        "Handshake error: we already know node %.40s, updating the address if needed.", sender->name);
+                    nodeUpdateAddress(sender,link,ntohs(hdr->port));
+                    freeClusterNode(link->node); /* will free the link too */
+                    return 0;
+                }
+
+                /* First thing to do is replacing the random name with the
+                 * right node name if this was an handshake stage. */
+                clusterRenameNode(link->node, hdr->sender);
+                redisLog(REDIS_DEBUG,"Handshake with node %.40s completed.",
+                    link->node->name);
+                link->node->flags &= ~REDIS_NODE_HANDSHAKE;
+            } else if (memcmp(link->node->name,hdr->sender,
+                        REDIS_CLUSTER_NAMELEN) != 0)
+            {
+                /* If the reply has a non matching node ID we
+                 * disconnect this node and set it as not having an associated
+                 * address. */
+                redisLog(REDIS_DEBUG,"PONG contains mismatching sender ID");
+                link->node->flags |= REDIS_NODE_NOADDR;
+                freeClusterLink(link);
+                /* FIXME: remove this node if we already have it.
+                 *
+                 * If we already have it but the IP is different, use
+                 * the new one if the old node is in FAIL, PFAIL, or NOADDR
+                 * status... */
+                return 0;
+            }
+        }
+        /* Update our info about the node */
+        link->node->pong_received = time(NULL);
+
+        /* Update master/slave info */
+        if (sender) {
+            if (!memcmp(hdr->slaveof,REDIS_NODE_NULL_NAME,
+                sizeof(hdr->slaveof)))
+            {
+                sender->flags &= ~REDIS_NODE_SLAVE;
+                sender->flags |= REDIS_NODE_MASTER;
+                sender->slaveof = NULL;
+            } else {
+                clusterNode *master = clusterLookupNode(hdr->slaveof);
+
+                sender->flags &= ~REDIS_NODE_MASTER;
+                sender->flags |= REDIS_NODE_SLAVE;
+                if (sender->numslaves) clusterNodeResetSlaves(sender);
+                if (master) clusterNodeAddSlave(master,sender);
+            }
+        }
+
+        /* Update our info about served slots if this new node is serving
+         * slots that are not served from our point of view. */
+        if (sender && sender->flags & REDIS_NODE_MASTER) {
+            int newslots, j;
+
+            newslots =
+                memcmp(sender->slots,hdr->myslots,sizeof(hdr->myslots)) != 0;
+            memcpy(sender->slots,hdr->myslots,sizeof(hdr->myslots));
+            if (newslots) {
+                for (j = 0; j < REDIS_CLUSTER_SLOTS; j++) {
+                    if (clusterNodeGetSlotBit(sender,j)) {
+                        if (server.cluster.slots[j] == sender) continue;
+                        if (server.cluster.slots[j] == NULL ||
+                            server.cluster.slots[j]->flags & REDIS_NODE_FAIL)
+                        {
+                            server.cluster.slots[j] = sender;
+                            update = 1;
+                        }
+                    }
+                }
+            }
+        }
+
+        /* Get info from the gossip section */
+        clusterProcessGossipSection(hdr,link);
+
+        /* Update the cluster state if needed */
+        if (update) clusterUpdateState();
+    } else if (type == CLUSTERMSG_TYPE_FAIL && sender) {
+        clusterNode *failing;
+
+        failing = clusterLookupNode(hdr->data.fail.about.nodename);
+        if (failing && !(failing->flags & REDIS_NODE_FAIL)) {
+            redisLog(REDIS_NOTICE,
+                "FAIL message received from %.40s about %.40s",
+                hdr->sender, hdr->data.fail.about.nodename);
+            failing->flags |= REDIS_NODE_FAIL;
+            failing->flags &= ~REDIS_NODE_PFAIL;
+            clusterUpdateState();
+        }
+    } else {
+        redisLog(REDIS_NOTICE,"Received unknown packet type: %d", type);
+    }
+    return 1;
+}
+
+/* This function is called when we detect the link with this node is lost.
+   We set the node as no longer connected. The Cluster Cron will detect
+   this connection and will try to get it connected again.
+   
+   Instead if the node is a temporary node used to accept a query, we
+   completely free the node on error. */
+void handleLinkIOError(clusterLink *link) {
+    freeClusterLink(link);
+}
+
+/* Send data. This is handled using a trivial send buffer that gets
+ * consumed by write(). We don't try to optimize this for speed too much
+ * as this is a very low traffic channel. */
+void clusterWriteHandler(aeEventLoop *el, int fd, void *privdata, int mask) {
+    clusterLink *link = (clusterLink*) privdata;
+    ssize_t nwritten;
+    REDIS_NOTUSED(el);
+    REDIS_NOTUSED(mask);
+
+    nwritten = write(fd, link->sndbuf, sdslen(link->sndbuf));
+    if (nwritten <= 0) {
+        redisLog(REDIS_NOTICE,"I/O error writing to node link: %s",
+            strerror(errno));
+        handleLinkIOError(link);
+        return;
+    }
+    link->sndbuf = sdsrange(link->sndbuf,nwritten,-1);
+    if (sdslen(link->sndbuf) == 0)
+        aeDeleteFileEvent(server.el, link->fd, AE_WRITABLE);
+}
+
+/* Read data. Try to read the first field of the header first to check the
+ * full length of the packet. When a whole packet is in memory this function
+ * will call the function to process the packet. And so forth. */
+void clusterReadHandler(aeEventLoop *el, int fd, void *privdata, int mask) {
+    char buf[1024];
+    ssize_t nread;
+    clusterMsg *hdr;
+    clusterLink *link = (clusterLink*) privdata;
+    int readlen;
+    REDIS_NOTUSED(el);
+    REDIS_NOTUSED(mask);
+
+again:
+    if (sdslen(link->rcvbuf) >= 4) {
+        hdr = (clusterMsg*) link->rcvbuf;
+        readlen = ntohl(hdr->totlen) - sdslen(link->rcvbuf);
+    } else {
+        readlen = 4 - sdslen(link->rcvbuf);
+    }
+
+    nread = read(fd,buf,readlen);
+    if (nread == -1 && errno == EAGAIN) return; /* Just no data */
+
+    if (nread <= 0) {
+        /* I/O error... */
+        redisLog(REDIS_NOTICE,"I/O error reading from node link: %s",
+            (nread == 0) ? "connection closed" : strerror(errno));
+        handleLinkIOError(link);
+        return;
+    } else {
+        /* Read data and recast the pointer to the new buffer. */
+        link->rcvbuf = sdscatlen(link->rcvbuf,buf,nread);
+        hdr = (clusterMsg*) link->rcvbuf;
+    }
+
+    /* Total length obtained? read the payload now instead of burning
+     * cycles waiting for a new event to fire. */
+    if (sdslen(link->rcvbuf) == 4) goto again;
+
+    /* Whole packet in memory? We can process it. */
+    if (sdslen(link->rcvbuf) == ntohl(hdr->totlen)) {
+        if (clusterProcessPacket(link)) {
+            sdsfree(link->rcvbuf);
+            link->rcvbuf = sdsempty();
+        }
+    }
+}
+
+/* Put stuff into the send buffer. */
+void clusterSendMessage(clusterLink *link, unsigned char *msg, size_t msglen) {
+    if (sdslen(link->sndbuf) == 0 && msglen != 0)
+        aeCreateFileEvent(server.el,link->fd,AE_WRITABLE,
+                    clusterWriteHandler,link);
+
+    link->sndbuf = sdscatlen(link->sndbuf, msg, msglen);
+}
+
+/* Build the message header */
+void clusterBuildMessageHdr(clusterMsg *hdr, int type) {
+    int totlen;
+
+    memset(hdr,0,sizeof(*hdr));
+    hdr->type = htons(type);
+    memcpy(hdr->sender,server.cluster.myself->name,REDIS_CLUSTER_NAMELEN);
+    memcpy(hdr->myslots,server.cluster.myself->slots,
+        sizeof(hdr->myslots));
+    memset(hdr->slaveof,0,REDIS_CLUSTER_NAMELEN);
+    if (server.cluster.myself->slaveof != NULL) {
+        memcpy(hdr->slaveof,server.cluster.myself->slaveof->name,
+                                    REDIS_CLUSTER_NAMELEN);
+    }
+    hdr->port = htons(server.port);
+    hdr->state = server.cluster.state;
+    memset(hdr->configdigest,0,32); /* FIXME: set config digest */
+
+    if (type == CLUSTERMSG_TYPE_FAIL) {
+        totlen = sizeof(clusterMsg)-sizeof(union clusterMsgData);
+        totlen += sizeof(clusterMsgDataFail);
+    }
+    hdr->totlen = htonl(totlen);
+    /* For PING, PONG, and MEET, fixing the totlen field is up to the caller */
+}
+
+/* Send a PING or PONG packet to the specified node, making sure to add enough
+ * gossip informations. */
+void clusterSendPing(clusterLink *link, int type) {
+    unsigned char buf[1024];
+    clusterMsg *hdr = (clusterMsg*) buf;
+    int gossipcount = 0, totlen;
+    /* freshnodes is the number of nodes we can still use to populate the
+     * gossip section of the ping packet. Basically we start with the nodes
+     * we have in memory minus two (ourself and the node we are sending the
+     * message to). Every time we add a node we decrement the counter, so when
+     * it will drop to <= zero we know there is no more gossip info we can
+     * send. */
+    int freshnodes = dictSize(server.cluster.nodes)-2;
+
+    if (link->node && type == CLUSTERMSG_TYPE_PING)
+        link->node->ping_sent = time(NULL);
+    clusterBuildMessageHdr(hdr,type);
+        
+    /* Populate the gossip fields */
+    while(freshnodes > 0 && gossipcount < 3) {
+        struct dictEntry *de = dictGetRandomKey(server.cluster.nodes);
+        clusterNode *this = dictGetEntryVal(de);
+        clusterMsgDataGossip *gossip;
+        int j;
+
+        /* Not interesting to gossip about ourself.
+         * Nor to send gossip info about HANDSHAKE state nodes (zero info). */
+        if (this == server.cluster.myself ||
+            this->flags & REDIS_NODE_HANDSHAKE) {
+                freshnodes--; /* otherwise we may loop forever. */
+                continue;
+        }
+
+        /* Check if we already added this node */
+        for (j = 0; j < gossipcount; j++) {
+            if (memcmp(hdr->data.ping.gossip[j].nodename,this->name,
+                    REDIS_CLUSTER_NAMELEN) == 0) break;
+        }
+        if (j != gossipcount) continue;
+
+        /* Add it */
+        freshnodes--;
+        gossip = &(hdr->data.ping.gossip[gossipcount]);
+        memcpy(gossip->nodename,this->name,REDIS_CLUSTER_NAMELEN);
+        gossip->ping_sent = htonl(this->ping_sent);
+        gossip->pong_received = htonl(this->pong_received);
+        memcpy(gossip->ip,this->ip,sizeof(this->ip));
+        gossip->port = htons(this->port);
+        gossip->flags = htons(this->flags);
+        gossipcount++;
+    }
+    totlen = sizeof(clusterMsg)-sizeof(union clusterMsgData);
+    totlen += (sizeof(clusterMsgDataGossip)*gossipcount);
+    hdr->count = htons(gossipcount);
+    hdr->totlen = htonl(totlen);
+    clusterSendMessage(link,buf,totlen);
+}
+
+/* Send a message to all the nodes with a reliable link */
+void clusterBroadcastMessage(void *buf, size_t len) {
+    dictIterator *di;
+    dictEntry *de;
+
+    di = dictGetIterator(server.cluster.nodes);
+    while((de = dictNext(di)) != NULL) {
+        clusterNode *node = dictGetEntryVal(de);
+
+        if (!node->link) continue;
+        if (node->flags & (REDIS_NODE_MYSELF|REDIS_NODE_NOADDR)) continue;
+        clusterSendMessage(node->link,buf,len);
+    }
+    dictReleaseIterator(di);
+}
+
+/* Send a FAIL message to all the nodes we are able to contact.
+ * The FAIL message is sent when we detect that a node is failing
+ * (REDIS_NODE_PFAIL) and we also receive a gossip confirmation of this:
+ * we switch the node state to REDIS_NODE_FAIL and ask all the other
+ * nodes to do the same ASAP. */
+void clusterSendFail(char *nodename) {
+    unsigned char buf[1024];
+    clusterMsg *hdr = (clusterMsg*) buf;
+
+    clusterBuildMessageHdr(hdr,CLUSTERMSG_TYPE_FAIL);
+    memcpy(hdr->data.fail.about.nodename,nodename,REDIS_CLUSTER_NAMELEN);
+    clusterBroadcastMessage(buf,ntohl(hdr->totlen));
+}
+
+/* -----------------------------------------------------------------------------
+ * CLUSTER cron job
+ * -------------------------------------------------------------------------- */
+
+/* This is executed 1 time every second */
+void clusterCron(void) {
+    dictIterator *di;
+    dictEntry *de;
+    int j;
+    time_t min_ping_sent = 0;
+    clusterNode *min_ping_node = NULL;
+
+    /* Check if we have disconnected nodes and reestablish the connection. */
+    di = dictGetIterator(server.cluster.nodes);
+    while((de = dictNext(di)) != NULL) {
+        clusterNode *node = dictGetEntryVal(de);
+
+        if (node->flags & (REDIS_NODE_MYSELF|REDIS_NODE_NOADDR)) continue;
+        if (node->link == NULL) {
+            int fd;
+            clusterLink *link;
+
+            fd = anetTcpNonBlockConnect(server.neterr, node->ip,
+                node->port+REDIS_CLUSTER_PORT_INCR);
+            if (fd == -1) continue;
+            link = createClusterLink(node);
+            link->fd = fd;
+            node->link = link;
+            aeCreateFileEvent(server.el,link->fd,AE_READABLE,clusterReadHandler,link);
+            /* If the node is flagged as MEET, we send a MEET message instead
+             * of a PING one, to force the receiver to add us in its node
+             * table. */
+            clusterSendPing(link, node->flags & REDIS_NODE_MEET ?
+                    CLUSTERMSG_TYPE_MEET : CLUSTERMSG_TYPE_PING);
+            /* We can clear the flag after the first packet is sent.
+             * If we'll never receive a PONG, we'll never send new packets
+             * to this node. Instead after the PONG is received and we
+             * are no longer in meet/handshake status, we want to send
+             * normal PING packets. */
+            node->flags &= ~REDIS_NODE_MEET;
+
+            redisLog(REDIS_NOTICE,"Connecting with Node %.40s at %s:%d\n", node->name, node->ip, node->port+REDIS_CLUSTER_PORT_INCR);
+        }
+    }
+    dictReleaseIterator(di);
+
+    /* Ping some random node. Check a few random nodes and ping the one with
+     * the oldest ping_sent time */
+    for (j = 0; j < 5; j++) {
+        de = dictGetRandomKey(server.cluster.nodes);
+        clusterNode *this = dictGetEntryVal(de);
+
+        if (this->link == NULL) continue;
+        if (this->flags & (REDIS_NODE_MYSELF|REDIS_NODE_HANDSHAKE)) continue;
+        if (min_ping_node == NULL || min_ping_sent > this->ping_sent) {
+            min_ping_node = this;
+            min_ping_sent = this->ping_sent;
+        }
+    }
+    if (min_ping_node) {
+        redisLog(REDIS_DEBUG,"Pinging node %40s", min_ping_node->name);
+        clusterSendPing(min_ping_node->link, CLUSTERMSG_TYPE_PING);
+    }
+
+    /* Iterate nodes to check if we need to flag something as failing */
+    di = dictGetIterator(server.cluster.nodes);
+    while((de = dictNext(di)) != NULL) {
+        clusterNode *node = dictGetEntryVal(de);
+        int delay;
+
+        if (node->flags &
+            (REDIS_NODE_MYSELF|REDIS_NODE_NOADDR|REDIS_NODE_HANDSHAKE|
+             REDIS_NODE_FAIL)) continue;
+        /* Check only if we already sent a ping and did not received
+         * a reply yet. */
+        if (node->ping_sent == 0 ||
+            node->ping_sent <= node->pong_received) continue;
+
+        delay = time(NULL) - node->pong_received;
+        if (node->flags & REDIS_NODE_PFAIL) {
+            /* The PFAIL condition can be reversed without external
+             * help if it is not transitive (that is, if it does not
+             * turn into a FAIL state). */
+            if (delay < server.cluster.node_timeout)
+                node->flags &= ~REDIS_NODE_PFAIL;
+        } else {
+            if (delay >= server.cluster.node_timeout) {
+                redisLog(REDIS_DEBUG,"*** NODE %.40s possibly failing",
+                    node->name);
+                node->flags |= REDIS_NODE_PFAIL;
+            }
+        }
+    }
+    dictReleaseIterator(di);
+}
+
+/* -----------------------------------------------------------------------------
+ * Slots management
+ * -------------------------------------------------------------------------- */
+
+/* Set the slot bit and return the old value. */
+int clusterNodeSetSlotBit(clusterNode *n, int slot) {
+    off_t byte = slot/8;
+    int bit = slot&7;
+    int old = (n->slots[byte] & (1<<bit)) != 0;
+    n->slots[byte] |= 1<<bit;
+    return old;
+}
+
+/* Clear the slot bit and return the old value. */
+int clusterNodeClearSlotBit(clusterNode *n, int slot) {
+    off_t byte = slot/8;
+    int bit = slot&7;
+    int old = (n->slots[byte] & (1<<bit)) != 0;
+    n->slots[byte] &= ~(1<<bit);
+    return old;
+}
+
+/* Return the slot bit from the cluster node structure. */
+int clusterNodeGetSlotBit(clusterNode *n, int slot) {
+    off_t byte = slot/8;
+    int bit = slot&7;
+    return (n->slots[byte] & (1<<bit)) != 0;
+}
+
+/* Add the specified slot to the list of slots that node 'n' will
+ * serve. Return REDIS_OK if the operation ended with success.
+ * If the slot is already assigned to another instance this is considered
+ * an error and REDIS_ERR is returned. */
+int clusterAddSlot(clusterNode *n, int slot) {
+    redisAssert(clusterNodeSetSlotBit(n,slot) == 0);
+    server.cluster.slots[slot] = server.cluster.myself;
+    printf("SLOT %d added to %.40s\n", slot, n->name);
+    return REDIS_OK;
+}
+
+/* -----------------------------------------------------------------------------
+ * Cluster state evaluation function
+ * -------------------------------------------------------------------------- */
+void clusterUpdateState(void) {
+    int ok = 1;
+    int j;
+
+    for (j = 0; j < REDIS_CLUSTER_SLOTS; j++) {
+        if (server.cluster.slots[j] == NULL ||
+            server.cluster.slots[j]->flags & (REDIS_NODE_FAIL))
+        {
+            ok = 0;
+            break;
+        }
+    }
+    if (ok) {
+        if (server.cluster.state == REDIS_CLUSTER_NEEDHELP) {
+            server.cluster.state = REDIS_CLUSTER_NEEDHELP;
+        } else {
+            server.cluster.state = REDIS_CLUSTER_OK;
+        }
+    } else {
+        server.cluster.state = REDIS_CLUSTER_FAIL;
+    }
+}
+
+/* -----------------------------------------------------------------------------
+ * CLUSTER command
+ * -------------------------------------------------------------------------- */
+
+void clusterCommand(redisClient *c) {
+    if (server.cluster_enabled == 0) {
+        addReplyError(c,"This instance has cluster support disabled");
+        return;
+    }
+
+    if (!strcasecmp(c->argv[1]->ptr,"meet") && c->argc == 4) {
+        clusterNode *n;
+        struct sockaddr_in sa;
+        long port;
+
+        /* Perform sanity checks on IP/port */
+        if (inet_aton(c->argv[2]->ptr,&sa.sin_addr) == 0) {
+            addReplyError(c,"Invalid IP address in MEET");
+            return;
+        }
+        if (getLongFromObjectOrReply(c, c->argv[3], &port, NULL) != REDIS_OK ||
+                    port < 0 || port > (65535-REDIS_CLUSTER_PORT_INCR))
+        {
+            addReplyError(c,"Invalid TCP port specified");
+            return;
+        }
+
+        /* Finally add the node to the cluster with a random name, this 
+         * will get fixed in the first handshake (ping/pong). */
+        n = createClusterNode(NULL,REDIS_NODE_HANDSHAKE|REDIS_NODE_MEET);
+        strncpy(n->ip,inet_ntoa(sa.sin_addr),sizeof(n->ip));
+        n->port = port;
+        clusterAddNode(n);
+        addReply(c,shared.ok);
+    } else if (!strcasecmp(c->argv[1]->ptr,"nodes") && c->argc == 2) {
+        sds ci = sdsempty();
+        dictIterator *di;
+        dictEntry *de;
+        robj *o;
+
+        di = dictGetIterator(server.cluster.nodes);
+        while((de = dictNext(di)) != NULL) {
+            clusterNode *node = dictGetEntryVal(de);
+
+            /* Node coordinates */
+            ci = sdscatprintf(ci,"%.40s %s:%d ",
+                node->name,
+                node->ip,
+                node->port);
+
+            /* Flags */
+            if (node->flags == 0) ci = sdscat(ci,"noflags,");
+            if (node->flags & REDIS_NODE_MYSELF) ci = sdscat(ci,"myself,");
+            if (node->flags & REDIS_NODE_MASTER) ci = sdscat(ci,"master,");
+            if (node->flags & REDIS_NODE_SLAVE) ci = sdscat(ci,"slave,");
+            if (node->flags & REDIS_NODE_PFAIL) ci = sdscat(ci,"fail?,");
+            if (node->flags & REDIS_NODE_FAIL) ci = sdscat(ci,"fail,");
+            if (node->flags & REDIS_NODE_HANDSHAKE) ci =sdscat(ci,"handshake,");
+            if (node->flags & REDIS_NODE_NOADDR) ci = sdscat(ci,"noaddr,");
+            if (ci[sdslen(ci)-1] == ',') ci[sdslen(ci)-1] = ' ';
+
+            /* Slave of... or just "-" */
+            if (node->slaveof)
+                ci = sdscatprintf(ci,"%.40s ",node->slaveof->name);
+            else
+                ci = sdscatprintf(ci,"- ");
+
+            /* Latency from the POV of this node, link status */
+            ci = sdscatprintf(ci,"%ld %ld %s\n",
+                (long) node->ping_sent,
+                (long) node->pong_received,
+                node->link ? "connected" : "disconnected");
+        }
+        dictReleaseIterator(di);
+        o = createObject(REDIS_STRING,ci);
+        addReplyBulk(c,o);
+        decrRefCount(o);
+    } else if (!strcasecmp(c->argv[1]->ptr,"addslots") && c->argc >= 3) {
+        int j;
+        long long slot;
+        unsigned char *slots = zmalloc(REDIS_CLUSTER_SLOTS);
+
+        memset(slots,0,REDIS_CLUSTER_SLOTS);
+        /* Check that all the arguments are parsable and that all the
+         * slots are not already busy. */
+        for (j = 2; j < c->argc; j++) {
+            if (getLongLongFromObject(c->argv[j],&slot) != REDIS_OK ||
+                slot < 0 || slot > REDIS_CLUSTER_SLOTS)
+            {
+                addReplyError(c,"Invalid or out of range slot index");
+                zfree(slots);
+                return;
+            }
+            if (server.cluster.slots[slot]) {
+                addReplyErrorFormat(c,"Slot %lld is already busy", slot);
+                zfree(slots);
+                return;
+            }
+            if (slots[slot]++ == 1) {
+                addReplyErrorFormat(c,"Slot %d specified multiple times",
+                    (int)slot);
+                zfree(slots);
+                return;
+            }
+        }
+        for (j = 0; j < REDIS_CLUSTER_SLOTS; j++) {
+            if (slots[j]) {
+                int retval = clusterAddSlot(server.cluster.myself,j);
+                
+                redisAssert(retval == REDIS_OK);
+            }
+        }
+        zfree(slots);
+        clusterUpdateState();
+        addReply(c,shared.ok);
+    } else if (!strcasecmp(c->argv[1]->ptr,"info") && c->argc == 2) {
+        char *statestr[] = {"ok","fail","needhelp"};
+        int slots_assigned = 0, slots_ok = 0, slots_pfail = 0, slots_fail = 0;
+        int j;
+
+        for (j = 0; j < REDIS_CLUSTER_SLOTS; j++) {
+            clusterNode *n = server.cluster.slots[j];
+
+            if (n == NULL) continue;
+            slots_assigned++;
+            if (n->flags & REDIS_NODE_FAIL) {
+                slots_fail++;
+            } else if (n->flags & REDIS_NODE_PFAIL) {
+                slots_pfail++;
+            } else {
+                slots_ok++;
+            }
+        }
+
+        sds info = sdscatprintf(sdsempty(),
+            "cluster_state:%s\r\n"
+            "cluster_slots_assigned:%d\r\n"
+            "cluster_slots_ok:%d\r\n"
+            "cluster_slots_pfail:%d\r\n"
+            "cluster_slots_fail:%d\r\n"
+            , statestr[server.cluster.state],
+            slots_assigned,
+            slots_ok,
+            slots_pfail,
+            slots_fail
+        );
+        addReplySds(c,sdscatprintf(sdsempty(),"$%lu\r\n",
+            (unsigned long)sdslen(info)));
+        addReplySds(c,info);
+        addReply(c,shared.crlf);
+    } else {
+        addReplyError(c,"Wrong CLUSTER subcommand or number of arguments");
+    }
+}
+
+/* -----------------------------------------------------------------------------
+ * RESTORE and MIGRATE commands
+ * -------------------------------------------------------------------------- */
+
+/* RESTORE key ttl serialized-value */
+void restoreCommand(redisClient *c) {
+    FILE *fp;
+    char buf[64];
+    robj *o;
+    unsigned char *data;
+    long ttl;
+
+    /* Make sure this key does not already exist here... */
+    if (dbExists(c->db,c->argv[1])) {
+        addReplyError(c,"Target key name is busy.");
+        return;
+    }
+
+    /* Check if the TTL value makes sense */
+    if (getLongFromObjectOrReply(c,c->argv[2],&ttl,NULL) != REDIS_OK) {
+        return;
+    } else if (ttl < 0) {
+        addReplyError(c,"Invalid TTL value, must be >= 0");
+        return;
+    }
+
+    /* rdbLoadObject() only works against file descriptors so we need to
+     * dump the serialized object into a file and reload. */
+    snprintf(buf,sizeof(buf),"redis-restore-%d.tmp",getpid());
+    fp = fopen(buf,"w+");
+    if (!fp) {
+        redisLog(REDIS_WARNING,"Can't open tmp file for RESTORE: %s",
+            strerror(errno));
+        addReplyErrorFormat(c,"RESTORE failed, tmp file creation error: %s",
+            strerror(errno));
+        return;
+    }
+    unlink(buf);
+
+    /* Write the actual data and rewind the file */
+    data = (unsigned char*) c->argv[3]->ptr;
+    if (fwrite(data+1,sdslen((sds)data)-1,1,fp) != 1) {
+        redisLog(REDIS_WARNING,"Can't write against tmp file for RESTORE: %s",
+            strerror(errno));
+        addReplyError(c,"RESTORE failed, tmp file I/O error.");
+        fclose(fp);
+        return;
+    }
+    rewind(fp);
+
+    /* Finally create the object from the serialized dump and
+     * store it at the specified key. */
+    o = rdbLoadObject(data[0],fp);
+    if (o == NULL) {
+        addReplyError(c,"Bad data format.");
+        fclose(fp);
+        return;
+    }
+    fclose(fp);
+
+    /* Create the key and set the TTL if any */
+    dbAdd(c->db,c->argv[1],o);
+    if (ttl) setExpire(c->db,c->argv[1],time(NULL)+ttl);
+    addReply(c,shared.ok);
+}
+
+/* MIGRATE host port key dbid timeout */
+void migrateCommand(redisClient *c) {
+    int fd;
+    long timeout;
+    long dbid;
+    char buf[64];
+    FILE *fp;
+    time_t ttl;
+    robj *o;
+    unsigned char type;
+    off_t payload_len;
+
+    /* Sanity check */
+    if (getLongFromObjectOrReply(c,c->argv[5],&timeout,NULL) != REDIS_OK)
+        return;
+    if (getLongFromObjectOrReply(c,c->argv[4],&dbid,NULL) != REDIS_OK)
+        return;
+    if (timeout <= 0) timeout = 1;
+
+    /* Check if the key is here. If not we reply with success as there is
+     * nothing to migrate (for instance the key expired in the meantime), but
+     * we include such information in the reply string. */
+    if ((o = lookupKeyRead(c->db,c->argv[3])) == NULL) {
+        addReplySds(c,sdsnew("+NOKEY"));
+        return;
+    }
+    
+    /* Connect */
+    fd = anetTcpNonBlockConnect(server.neterr,c->argv[1]->ptr,
+                atoi(c->argv[2]->ptr));
+    if (fd == -1) {
+        addReplyErrorFormat(c,"Can't connect to target node: %s",
+            server.neterr);
+        return;
+    }
+    if ((aeWait(fd,AE_WRITABLE,timeout*1000) & AE_WRITABLE) == 0) {
+        addReplyError(c,"Timeout connecting to the client");
+        return;
+    }
+
+    /* Create temp file */
+    snprintf(buf,sizeof(buf),"redis-migrate-%d.tmp",getpid());
+    fp = fopen(buf,"w+");
+    if (!fp) {
+        redisLog(REDIS_WARNING,"Can't open tmp file for MIGRATE: %s",
+            strerror(errno));
+        addReplyErrorFormat(c,"MIGRATE failed, tmp file creation error: %s.",
+            strerror(errno));
+        return;
+    }
+    unlink(buf);
+
+    /* Build the SELECT + RESTORE query writing it in our temp file. */
+    if (fwriteBulkCount(fp,'*',2) == 0) goto file_wr_err;
+    if (fwriteBulkString(fp,"SELECT",6) == 0) goto file_wr_err;
+    if (fwriteBulkLongLong(fp,dbid) == 0) goto file_wr_err;
+
+    ttl = getExpire(c->db,c->argv[3]);
+    type = o->type;
+    if (fwriteBulkCount(fp,'*',4) == 0) goto file_wr_err;
+    if (fwriteBulkString(fp,"RESTORE",7) == 0) goto file_wr_err;
+    if (fwriteBulkObject(fp,c->argv[3]) == 0) goto file_wr_err;
+    if (fwriteBulkLongLong(fp, (ttl == -1) ? 0 : ttl) == 0) goto file_wr_err;
+
+    /* Finally the last argument that is the serailized object payload
+     * in the form: <type><rdb-serailized-object>. */
+    payload_len = rdbSavedObjectLen(o);
+    if (fwriteBulkCount(fp,'$',payload_len+1) == 0) goto file_wr_err;
+    if (fwrite(&type,1,1,fp) == 0) goto file_wr_err;
+    if (rdbSaveObject(fp,o) == -1) goto file_wr_err;
+    if (fwrite("\r\n",2,1,fp) == 0) goto file_wr_err;
+
+    /* Tranfer the query to the other node */
+    rewind(fp);
+    {
+        char buf[4096];
+        size_t nread;
+
+        while ((nread = fread(buf,1,sizeof(buf),fp)) != 0) {
+            int nwritten;
+
+            nwritten = syncWrite(fd,buf,nread,timeout);
+            if (nwritten != (signed)nread) goto socket_wr_err;
+        }
+        if (ferror(fp)) goto file_rd_err;
+    }
+
+    /* Read back the reply */
+    {
+        char buf1[1024];
+        char buf2[1024];
+
+        /* Read the two replies */
+        if (syncReadLine(fd, buf1, sizeof(buf1), timeout) <= 0)
+            goto socket_rd_err;
+        if (syncReadLine(fd, buf2, sizeof(buf2), timeout) <= 0)
+                goto socket_rd_err;
+        if (buf1[0] == '-' || buf2[0] == '-') {
+            addReplyErrorFormat(c,"Target instance replied with error: %s",
+                (buf1[0] == '-') ? buf1+1 : buf2+1);
+        } else {
+            dbDelete(c->db,c->argv[3]);
+            addReply(c,shared.ok);
+        }
+    }
+    fclose(fp);
+    close(fd);
+    return;
+
+file_wr_err:
+    redisLog(REDIS_WARNING,"Can't write on tmp file for MIGRATE: %s",
+        strerror(errno));
+    addReplyErrorFormat(c,"MIGRATE failed, tmp file write error: %s.",
+        strerror(errno));
+    fclose(fp);
+    close(fd);
+
+file_rd_err:
+    redisLog(REDIS_WARNING,"Can't read from tmp file for MIGRATE: %s",
+        strerror(errno));
+    addReplyErrorFormat(c,"MIGRATE failed, tmp file read error: %s.",
+        strerror(errno));
+    fclose(fp);
+    close(fd);
+
+socket_wr_err:
+    redisLog(REDIS_NOTICE,"Can't write to target node for MIGRATE: %s",
+        strerror(errno));
+    addReplyErrorFormat(c,"MIGRATE failed, writing to target node: %s.",
+        strerror(errno));
+    fclose(fp);
+    close(fd);
+
+socket_rd_err:
+    redisLog(REDIS_NOTICE,"Can't read from target node for MIGRATE: %s",
+        strerror(errno));
+    addReplyErrorFormat(c,"MIGRATE failed, reading from target node: %s.",
+        strerror(errno));
+    fclose(fp);
+    close(fd);
+}
+
+/* -----------------------------------------------------------------------------
+ * Cluster functions related to serving / redirecting clients
+ * -------------------------------------------------------------------------- */
+
+/* Return the pointer to the cluster node that is able to serve the query
+ * as all the keys belong to hash slots for which the node is in charge.
+ *
+ * If keys in query spawn multiple nodes NULL is returned. */
+clusterNode *getNodeByQuery(redisClient *c, struct redisCommand *cmd, robj **argv, int argc, int *hashslot) {
+    clusterNode *n = NULL;
+    multiState *ms, _ms;
+    multiCmd mc;
+    int i;
+
+    /* We handle all the cases as if they were EXEC commands, so we have
+     * a common code path for everything */
+    if (cmd->proc == execCommand) {
+        /* If REDIS_MULTI flag is not set EXEC is just going to return an
+         * error. */
+        if (!(c->flags & REDIS_MULTI)) return server.cluster.myself;
+        ms = &c->mstate;
+    } else {
+        /* Create a fake Multi State structure, with just one command */
+        ms = &_ms;
+        _ms.commands = &mc;
+        _ms.count = 1;
+        mc.argv = argv;
+        mc.argc = argc;
+        mc.cmd = cmd;
+    }
+
+    for (i = 0; i < ms->count; i++) {
+        struct redisCommand *mcmd;
+        robj **margv;
+        int margc, *keyindex, numkeys, j;
+
+        mcmd = ms->commands[i].cmd;
+        margc = ms->commands[i].argc;
+        margv = ms->commands[i].argv;
+
+        keyindex = getKeysFromCommand(mcmd,margv,margc,&numkeys,
+                                      REDIS_GETKEYS_PRELOAD);
+        for (j = 0; j < numkeys; j++) {
+            int slot = keyHashSlot((char*)margv[keyindex[j]]->ptr,
+                sdslen(margv[keyindex[j]]->ptr));
+            struct clusterNode *slotnode;
+
+            slotnode = server.cluster.slots[slot];
+            if (hashslot) *hashslot = slot;
+            /* Node not assigned? (Should never happen actually
+             * if we reached this function).
+             * Different node than the previous one?
+             * Return NULL, the cluster can't serve multi-node requests */
+            if (slotnode == NULL || (n && slotnode != n)) {
+                getKeysFreeResult(keyindex);
+                return NULL;
+            } else {
+                n = slotnode;
+            }
+        }
+        getKeysFreeResult(keyindex);
+    }
+    return (n == NULL) ? server.cluster.myself : n;
+}
diff --git a/src/config.c b/src/config.c
index e0bf1574a..e40fdedad 100644
--- a/src/config.c
+++ b/src/config.c
@@ -285,6 +285,10 @@ void loadServerConfig(char *filename) {
                     err = "Target command name already exists"; goto loaderr;
                 }
             }
+        } else if (!strcasecmp(argv[0],"cluster-enabled") && argc == 2) {
+            if ((server.cluster_enabled = yesnotoi(argv[1])) == -1) {
+                err = "argument must be 'yes' or 'no'"; goto loaderr;
+            }
         } else {
             err = "Bad directive or wrong number of arguments"; goto loaderr;
         }
diff --git a/src/crc16.c b/src/crc16.c
new file mode 100644
index 000000000..37fb4ca77
--- /dev/null
+++ b/src/crc16.c
@@ -0,0 +1,74 @@
+#include "redis.h"
+
+/*      
+ * Copyright 2001-2010 Georges Menie (www.menie.org)
+ * Copyright 2010 Salvatore Sanfilippo (adapted to Redis coding style)
+ * All rights reserved.
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ *     * Redistributions of source code must retain the above copyright
+ *       notice, this list of conditions and the following disclaimer.
+ *     * Redistributions in binary form must reproduce the above copyright
+ *       notice, this list of conditions and the following disclaimer in the
+ *       documentation and/or other materials provided with the distribution.
+ *     * Neither the name of the University of California, Berkeley nor the
+ *       names of its contributors may be used to endorse or promote products
+ *       derived from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND ANY
+ * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED. IN NO EVENT SHALL THE REGENTS AND CONTRIBUTORS BE LIABLE FOR ANY
+ * DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+/* CRC16 implementation acording to CCITT standards */
+
+static const uint16_t crc16tab[256]= {
+    0x0000,0x1021,0x2042,0x3063,0x4084,0x50a5,0x60c6,0x70e7,
+    0x8108,0x9129,0xa14a,0xb16b,0xc18c,0xd1ad,0xe1ce,0xf1ef,
+    0x1231,0x0210,0x3273,0x2252,0x52b5,0x4294,0x72f7,0x62d6,
+    0x9339,0x8318,0xb37b,0xa35a,0xd3bd,0xc39c,0xf3ff,0xe3de,
+    0x2462,0x3443,0x0420,0x1401,0x64e6,0x74c7,0x44a4,0x5485,
+    0xa56a,0xb54b,0x8528,0x9509,0xe5ee,0xf5cf,0xc5ac,0xd58d,
+    0x3653,0x2672,0x1611,0x0630,0x76d7,0x66f6,0x5695,0x46b4,
+    0xb75b,0xa77a,0x9719,0x8738,0xf7df,0xe7fe,0xd79d,0xc7bc,
+    0x48c4,0x58e5,0x6886,0x78a7,0x0840,0x1861,0x2802,0x3823,
+    0xc9cc,0xd9ed,0xe98e,0xf9af,0x8948,0x9969,0xa90a,0xb92b,
+    0x5af5,0x4ad4,0x7ab7,0x6a96,0x1a71,0x0a50,0x3a33,0x2a12,
+    0xdbfd,0xcbdc,0xfbbf,0xeb9e,0x9b79,0x8b58,0xbb3b,0xab1a,
+    0x6ca6,0x7c87,0x4ce4,0x5cc5,0x2c22,0x3c03,0x0c60,0x1c41,
+    0xedae,0xfd8f,0xcdec,0xddcd,0xad2a,0xbd0b,0x8d68,0x9d49,
+    0x7e97,0x6eb6,0x5ed5,0x4ef4,0x3e13,0x2e32,0x1e51,0x0e70,
+    0xff9f,0xefbe,0xdfdd,0xcffc,0xbf1b,0xaf3a,0x9f59,0x8f78,
+    0x9188,0x81a9,0xb1ca,0xa1eb,0xd10c,0xc12d,0xf14e,0xe16f,
+    0x1080,0x00a1,0x30c2,0x20e3,0x5004,0x4025,0x7046,0x6067,
+    0x83b9,0x9398,0xa3fb,0xb3da,0xc33d,0xd31c,0xe37f,0xf35e,
+    0x02b1,0x1290,0x22f3,0x32d2,0x4235,0x5214,0x6277,0x7256,
+    0xb5ea,0xa5cb,0x95a8,0x8589,0xf56e,0xe54f,0xd52c,0xc50d,
+    0x34e2,0x24c3,0x14a0,0x0481,0x7466,0x6447,0x5424,0x4405,
+    0xa7db,0xb7fa,0x8799,0x97b8,0xe75f,0xf77e,0xc71d,0xd73c,
+    0x26d3,0x36f2,0x0691,0x16b0,0x6657,0x7676,0x4615,0x5634,
+    0xd94c,0xc96d,0xf90e,0xe92f,0x99c8,0x89e9,0xb98a,0xa9ab,
+    0x5844,0x4865,0x7806,0x6827,0x18c0,0x08e1,0x3882,0x28a3,
+    0xcb7d,0xdb5c,0xeb3f,0xfb1e,0x8bf9,0x9bd8,0xabbb,0xbb9a,
+    0x4a75,0x5a54,0x6a37,0x7a16,0x0af1,0x1ad0,0x2ab3,0x3a92,
+    0xfd2e,0xed0f,0xdd6c,0xcd4d,0xbdaa,0xad8b,0x9de8,0x8dc9,
+    0x7c26,0x6c07,0x5c64,0x4c45,0x3ca2,0x2c83,0x1ce0,0x0cc1,
+    0xef1f,0xff3e,0xcf5d,0xdf7c,0xaf9b,0xbfba,0x8fd9,0x9ff8,
+    0x6e17,0x7e36,0x4e55,0x5e74,0x2e93,0x3eb2,0x0ed1,0x1ef0
+};
+  
+uint16_t crc16(const char *buf, int len) {
+    int counter;
+    uint16_t crc = 0;
+    for (counter = 0; counter < len; counter++)
+            crc = (crc<<8) ^ crc16tab[((crc>>8) ^ *buf++)&0x00FF];
+    return crc;
+}
diff --git a/src/db.c b/src/db.c
index fea2f12ec..85a220626 100644
--- a/src/db.c
+++ b/src/db.c
@@ -307,6 +307,10 @@ void existsCommand(redisClient *c) {
 void selectCommand(redisClient *c) {
     int id = atoi(c->argv[1]->ptr);
 
+    if (server.cluster_enabled) {
+        addReplyError(c,"SELECT is not allowed in cluster mode");
+        return;
+    }
     if (selectDb(c,id) == REDIS_ERR) {
         addReplyError(c,"invalid DB index");
     } else {
@@ -428,6 +432,11 @@ void moveCommand(redisClient *c) {
     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;
diff --git a/src/rdb.c b/src/rdb.c
index c9fb3e830..f14467d12 100644
--- a/src/rdb.c
+++ b/src/rdb.c
@@ -245,7 +245,7 @@ int rdbSaveDoubleValue(FILE *fp, double val) {
     return rdbWriteRaw(fp,buf,len);
 }
 
-/* Save a Redis object. */
+/* Save a Redis object. Returns -1 on error, 0 on success. */
 int rdbSaveObject(FILE *fp, robj *o) {
     int n, nwritten = 0;
 
diff --git a/src/redis-cli.c b/src/redis-cli.c
index 123231d47..1b23c0b1f 100644
--- a/src/redis-cli.c
+++ b/src/redis-cli.c
@@ -394,15 +394,18 @@ static sds cliFormatReplyRaw(redisReply *r) {
     switch (r->type) {
     case REDIS_REPLY_NIL:
         /* Nothing... */
-    break;
+        break;
     case REDIS_REPLY_ERROR:
+        out = sdscatlen(out,r->str,r->len);
+        out = sdscatlen(out,"\n",1);
+        break;
     case REDIS_REPLY_STATUS:
     case REDIS_REPLY_STRING:
         out = sdscatlen(out,r->str,r->len);
-    break;
+        break;
     case REDIS_REPLY_INTEGER:
         out = sdscatprintf(out,"%lld",r->integer);
-    break;
+        break;
     case REDIS_REPLY_ARRAY:
         for (i = 0; i < r->elements; i++) {
             if (i > 0) out = sdscat(out,config.mb_delim);
@@ -410,7 +413,7 @@ static sds cliFormatReplyRaw(redisReply *r) {
             out = sdscatlen(out,tmp,sdslen(tmp));
             sdsfree(tmp);
         }
-    break;
+        break;
     default:
         fprintf(stderr,"Unknown reply type: %d\n", r->type);
         exit(1);
@@ -464,7 +467,15 @@ static int cliSendCommand(int argc, char **argv, int repeat) {
         return REDIS_OK;
     }
 
-    output_raw = !strcasecmp(command,"info");
+    output_raw = 0;
+    if (!strcasecmp(command,"info") ||
+        (argc == 2 && !strcasecmp(command,"cluster") &&
+                      (!strcasecmp(argv[1],"nodes") ||
+                       !strcasecmp(argv[1],"info"))))
+    {
+        output_raw = 1;
+    }
+
     if (!strcasecmp(command,"help") || !strcasecmp(command,"?")) {
         cliOutputHelp(--argc, ++argv);
         return REDIS_OK;
diff --git a/src/redis.c b/src/redis.c
index 19fd912cb..25205752e 100644
--- a/src/redis.c
+++ b/src/redis.c
@@ -187,7 +187,10 @@ struct redisCommand redisCommandTable[] = {
     {"punsubscribe",punsubscribeCommand,-1,0,NULL,0,0,0,0,0},
     {"publish",publishCommand,3,REDIS_CMD_FORCE_REPLICATION,NULL,0,0,0,0,0},
     {"watch",watchCommand,-2,0,noPreloadGetKeys,1,-1,1,0,0},
-    {"unwatch",unwatchCommand,1,0,NULL,0,0,0,0,0}
+    {"unwatch",unwatchCommand,1,0,NULL,0,0,0,0,0},
+    {"cluster",clusterCommand,-2,0,NULL,0,0,0,0,0},
+    {"restore",restoreCommand,3,0,NULL,0,0,0,0,0},
+    {"migrate",migrateCommand,6,0,NULL,0,0,0,0,0}
 };
 
 /*============================ Utility functions ============================ */
@@ -440,6 +443,17 @@ dictType keylistDictType = {
     dictListDestructor          /* val destructor */
 };
 
+/* Cluster nodes hash table, mapping nodes addresses 1.2.3.4:6379 to
+ * clusterNode structures. */
+dictType clusterNodesDictType = {
+    dictSdsHash,                /* hash function */
+    NULL,                       /* key dup */
+    NULL,                       /* val dup */
+    dictSdsKeyCompare,          /* key compare */
+    dictSdsDestructor,          /* key destructor */
+    NULL                        /* val destructor */
+};
+
 int htNeedsResize(dict *dict) {
     long long size, used;
 
@@ -669,6 +683,9 @@ int serverCron(struct aeEventLoop *eventLoop, long long id, void *clientData) {
      * to detect transfer failures. */
     if (!(loops % 10)) replicationCron();
 
+    /* Run other sub-systems specific cron jobs */
+    if (server.cluster_enabled && !(loops % 10)) clusterCron();
+
     server.cronloops++;
     return 100;
 }
@@ -823,6 +840,7 @@ void initServerConfig() {
     server.set_max_intset_entries = REDIS_SET_MAX_INTSET_ENTRIES;
     server.shutdown_asap = 0;
     server.cache_flush_delay = 0;
+    server.cluster_enabled = 0;
 
     updateLRUClock();
     resetServerSaveParams();
@@ -945,6 +963,7 @@ void initServer() {
     }
 
     if (server.ds_enabled) dsInit();
+    if (server.cluster_enabled) clusterInit();
     srand(time(NULL)^getpid());
 }
 
@@ -1051,6 +1070,27 @@ int processCommand(redisClient *c) {
         return REDIS_OK;
     }
 
+    /* If cluster is enabled, redirect here */
+    if (server.cluster_enabled &&
+                !(cmd->getkeys_proc == NULL && cmd->firstkey == 0)) {
+        int hashslot;
+
+        if (server.cluster.state != REDIS_CLUSTER_OK) {
+            addReplyError(c,"The cluster is down. Check with CLUSTER INFO for more information");
+            return REDIS_OK;
+        } else {
+            clusterNode *n = getNodeByQuery(c,cmd,c->argv,c->argc,&hashslot);
+            if (n == NULL) {
+                addReplyError(c,"Invalid cross-node request");
+                return REDIS_OK;
+            } else if (n != server.cluster.myself) {
+                addReplySds(c,sdscatprintf(sdsempty(),
+                    "-MOVED %d %s:%d\r\n",hashslot,n->ip,n->port));
+                return REDIS_OK;
+            }
+        }
+    }
+
     /* Handle the maxmemory directive.
      *
      * First we try to free some memory if possible (if there are volatile
diff --git a/src/redis.h b/src/redis.h
index b51883309..0ddd4730f 100644
--- a/src/redis.h
+++ b/src/redis.h
@@ -18,6 +18,7 @@
 #include <inttypes.h>
 #include <pthread.h>
 #include <syslog.h>
+#include <netinet/in.h>
 
 #include "ae.h"     /* Event driven programming library */
 #include "sds.h"    /* Dynamic safe strings */
@@ -360,7 +361,123 @@ struct sharedObjectsStruct {
     *integers[REDIS_SHARED_INTEGERS];
 };
 
-/* Global server state structure */
+/*-----------------------------------------------------------------------------
+ * Redis cluster data structures
+ *----------------------------------------------------------------------------*/
+
+#define REDIS_CLUSTER_SLOTS 4096
+#define REDIS_CLUSTER_OK 0          /* Everything looks ok */
+#define REDIS_CLUSTER_FAIL 1        /* The cluster can't work */
+#define REDIS_CLUSTER_NEEDHELP 2    /* The cluster works, but needs some help */
+#define REDIS_CLUSTER_NAMELEN 40    /* sha1 hex length */
+#define REDIS_CLUSTER_PORT_INCR 10000 /* Cluster port = baseport + PORT_INCR */
+
+struct clusterNode;
+
+/* clusterLink encapsulates everything needed to talk with a remote node. */
+typedef struct clusterLink {
+    int fd;                     /* TCP socket file descriptor */
+    sds sndbuf;                 /* Packet send buffer */
+    sds rcvbuf;                 /* Packet reception buffer */
+    struct clusterNode *node;   /* Node related to this link if any, or NULL */
+} clusterLink;
+
+/* Node flags */
+#define REDIS_NODE_MASTER 1     /* The node is a master */
+#define REDIS_NODE_SLAVE 2      /* The node is a slave */
+#define REDIS_NODE_PFAIL 4      /* Failure? Need acknowledge */
+#define REDIS_NODE_FAIL 8       /* The node is believed to be malfunctioning */
+#define REDIS_NODE_MYSELF 16    /* This node is myself */
+#define REDIS_NODE_HANDSHAKE 32 /* We have still to exchange the first ping */
+#define REDIS_NODE_NOADDR   64  /* We don't know the address of this node */
+#define REDIS_NODE_MEET 128     /* Send a MEET message to this node */
+#define REDIS_NODE_NULL_NAME "\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000\000"
+
+struct clusterNode {
+    char name[REDIS_CLUSTER_NAMELEN]; /* Node name, hex string, sha1-size */
+    int flags;      /* REDIS_NODE_... */
+    unsigned char slots[REDIS_CLUSTER_SLOTS/8]; /* slots handled by this node */
+    int numslaves;  /* Number of slave nodes, if this is a master */
+    struct clusterNode **slaves; /* pointers to slave nodes */
+    struct clusterNode *slaveof; /* pointer to the master node */
+    time_t ping_sent;       /* Unix time we sent latest ping */
+    time_t pong_received;   /* Unix time we received the pong */
+    char *configdigest;         /* Configuration digest of this node */
+    time_t configdigest_ts;     /* Configuration digest timestamp */
+    char ip[16];                /* Latest known IP address of this node */
+    int port;                   /* Latest known port of this node */
+    clusterLink *link;          /* TCP/IP link with this node */
+};
+typedef struct clusterNode clusterNode;
+
+typedef struct {
+    clusterNode *myself;  /* This node */
+    int state;            /* REDIS_CLUSTER_OK, REDIS_CLUSTER_FAIL, ... */
+    int node_timeout;
+    dict *nodes;          /* Hash table of name -> clusterNode structures */
+    clusterNode *migrating_slots_to[REDIS_CLUSTER_SLOTS];
+    clusterNode *importing_slots_from[REDIS_CLUSTER_SLOTS];
+    clusterNode *slots[REDIS_CLUSTER_SLOTS];
+} clusterState;
+
+/* Redis cluster messages header */
+
+/* Note that the PING, PONG and MEET messages are actually the same exact
+ * kind of packet. PONG is the reply to ping, in the extact format as a PING,
+ * while MEET is a special PING that forces the receiver to add the sender
+ * as a node (if it is not already in the list). */
+#define CLUSTERMSG_TYPE_PING 0          /* Ping */
+#define CLUSTERMSG_TYPE_PONG 1          /* Pong (reply to Ping) */
+#define CLUSTERMSG_TYPE_MEET 2          /* Meet "let's join" message */
+#define CLUSTERMSG_TYPE_FAIL 3          /* Mark node xxx as failing */
+
+/* Initially we don't know our "name", but we'll find it once we connect
+ * to the first node, using the getsockname() function. Then we'll use this
+ * address for all the next messages. */
+typedef struct {
+    char nodename[REDIS_CLUSTER_NAMELEN];
+    uint32_t ping_sent;
+    uint32_t pong_received;
+    char ip[16];    /* IP address last time it was seen */
+    uint16_t port;  /* port last time it was seen */
+    uint16_t flags;
+    uint32_t notused; /* for 64 bit alignment */
+} clusterMsgDataGossip;
+
+typedef struct {
+    char nodename[REDIS_CLUSTER_NAMELEN];
+} clusterMsgDataFail;
+
+union clusterMsgData {
+    /* PING, MEET and PONG */
+    struct {
+        /* Array of N clusterMsgDataGossip structures */
+        clusterMsgDataGossip gossip[1];
+    } ping;
+    /* FAIL */
+    struct {
+        clusterMsgDataFail about;
+    } fail;
+};
+
+typedef struct {
+    uint32_t totlen;    /* Total length of this message */
+    uint16_t type;      /* Message type */
+    uint16_t count;     /* Only used for some kind of messages. */
+    char sender[REDIS_CLUSTER_NAMELEN]; /* Name of the sender node */
+    unsigned char myslots[REDIS_CLUSTER_SLOTS/8];
+    char slaveof[REDIS_CLUSTER_NAMELEN];
+    char configdigest[32];
+    uint16_t port;      /* Sender TCP base port */
+    unsigned char state; /* Cluster state from the POV of the sender */
+    unsigned char notused[5]; /* Reserved for future use. For alignment. */
+    union clusterMsgData data;
+} clusterMsg;
+
+/*-----------------------------------------------------------------------------
+ * Global server state
+ *----------------------------------------------------------------------------*/
+
 struct redisServer {
     /* General */
     pthread_t mainthread;
@@ -373,6 +490,7 @@ struct redisServer {
     char *unixsocket;
     int ipfd;
     int sofd;
+    int cfd;
     list *clients;
     list *slaves, *monitors;
     char neterr[ANET_ERR_LEN];
@@ -499,6 +617,8 @@ struct redisServer {
     /* Misc */
     unsigned lruclock:22;        /* clock incrementing every minute, for LRU */
     unsigned lruclock_padding:10;
+    int cluster_enabled;
+    clusterState cluster;
 };
 
 typedef struct pubsubPattern {
@@ -633,6 +753,7 @@ extern struct redisServer server;
 extern struct sharedObjectsStruct shared;
 extern dictType setDictType;
 extern dictType zsetDictType;
+extern dictType clusterNodesDictType;
 extern double R_Zero, R_PosInf, R_NegInf, R_Nan;
 dictType hashDictType;
 
@@ -747,6 +868,7 @@ int fwriteBulkString(FILE *fp, char *s, unsigned long len);
 int fwriteBulkDouble(FILE *fp, double d);
 int fwriteBulkLongLong(FILE *fp, long long l);
 int fwriteBulkObject(FILE *fp, robj *obj);
+int fwriteBulkCount(FILE *fp, char prefix, int count);
 
 /* Replication */
 void replicationFeedSlaves(list *slaves, int dictid, robj **argv, int argc);
@@ -923,6 +1045,15 @@ int *noPreloadGetKeys(struct redisCommand *cmd,robj **argv, int argc, int *numke
 int *renameGetKeys(struct redisCommand *cmd,robj **argv, int argc, int *numkeys, int flags);
 int *zunionInterGetKeys(struct redisCommand *cmd,robj **argv, int argc, int *numkeys, int flags);
 
+/* Cluster */
+void clusterInit(void);
+unsigned short crc16(const char *buf, int len);
+unsigned int keyHashSlot(char *key, int keylen);
+clusterNode *createClusterNode(char *nodename, int flags);
+int clusterAddNode(clusterNode *node);
+void clusterCron(void);
+clusterNode *getNodeByQuery(redisClient *c, struct redisCommand *cmd, robj **argv, int argc, int *hashslot);
+
 /* Git SHA1 */
 char *redisGitSHA1(void);
 char *redisGitDirty(void);
@@ -1045,6 +1176,9 @@ void punsubscribeCommand(redisClient *c);
 void publishCommand(redisClient *c);
 void watchCommand(redisClient *c);
 void unwatchCommand(redisClient *c);
+void clusterCommand(redisClient *c);
+void restoreCommand(redisClient *c);
+void migrateCommand(redisClient *c);
 
 #if defined(__GNUC__)
 void *calloc(size_t count, size_t size) __attribute__ ((deprecated));
diff --git a/src/syncio.c b/src/syncio.c
index 28ac18114..3d0e0451b 100644
--- a/src/syncio.c
+++ b/src/syncio.c
@@ -107,6 +107,7 @@ int syncReadLine(int fd, char *ptr, ssize_t size, int timeout) {
 int fwriteBulkString(FILE *fp, char *s, unsigned long len) {
     char cbuf[128];
     int clen;
+
     cbuf[0] = '$';
     clen = 1+ll2string(cbuf+1,sizeof(cbuf)-1,len);
     cbuf[clen++] = '\r';
@@ -117,6 +118,19 @@ int fwriteBulkString(FILE *fp, char *s, unsigned long len) {
     return 1;
 }
 
+/* Write a multi bulk count in the form "*<count>\r\n" */
+int fwriteBulkCount(FILE *fp, char prefix, int count) {
+    char cbuf[128];
+    int clen;
+
+    cbuf[0] = prefix;
+    clen = 1+ll2string(cbuf+1,sizeof(cbuf)-1,count);
+    cbuf[clen++] = '\r';
+    cbuf[clen++] = '\n';
+    if (fwrite(cbuf,clen,1,fp) == 0) return 0;
+    return 1;
+}
+
 /* Write a double value in bulk format $<count>\r\n<payload>\r\n */
 int fwriteBulkDouble(FILE *fp, double d) {
     char buf[128], dbuf[128];