Commit bfe85f7c authored by antirez's avatar antirez
Browse files

Merge branch 'unstable'

parents b8513c93 6b52ad87
...@@ -53,5 +53,6 @@ int anetWrite(int fd, char *buf, int count); ...@@ -53,5 +53,6 @@ int anetWrite(int fd, char *buf, int count);
int anetNonBlock(char *err, int fd); int anetNonBlock(char *err, int fd);
int anetTcpNoDelay(char *err, int fd); int anetTcpNoDelay(char *err, int fd);
int anetTcpKeepAlive(char *err, int fd); int anetTcpKeepAlive(char *err, int fd);
int anetPeerToString(int fd, char *ip, int *port);
#endif #endif
...@@ -284,9 +284,11 @@ int loadAppendOnlyFile(char *filename) { ...@@ -284,9 +284,11 @@ int loadAppendOnlyFile(char *filename) {
/* The fake client should not have a reply */ /* The fake client should not have a reply */
redisAssert(fakeClient->bufpos == 0 && listLength(fakeClient->reply) == 0); redisAssert(fakeClient->bufpos == 0 && listLength(fakeClient->reply) == 0);
/* Clean up, ready for the next command */ /* Clean up. Command code may have changed argv/argc so we use the
for (j = 0; j < argc; j++) decrRefCount(argv[j]); * argv/argc of the client instead of the local variables. */
zfree(argv); for (j = 0; j < fakeClient->argc; j++)
decrRefCount(fakeClient->argv[j]);
zfree(fakeClient->argv);
} }
/* This point can only be reached when EOF is reached without errors. /* This point can only be reached when EOF is reached without errors.
...@@ -346,7 +348,7 @@ int rewriteAppendOnlyFile(char *filename) { ...@@ -346,7 +348,7 @@ int rewriteAppendOnlyFile(char *filename) {
/* Iterate this DB writing every entry */ /* Iterate this DB writing every entry */
while((de = dictNext(di)) != NULL) { while((de = dictNext(di)) != NULL) {
sds keystr = dictGetEntryKey(de); sds keystr;
robj key, *o; robj key, *o;
time_t expiretime; time_t expiretime;
...@@ -429,12 +431,43 @@ int rewriteAppendOnlyFile(char *filename) { ...@@ -429,12 +431,43 @@ int rewriteAppendOnlyFile(char *filename) {
} }
} else if (o->type == REDIS_ZSET) { } else if (o->type == REDIS_ZSET) {
/* Emit the ZADDs needed to rebuild the sorted set */ /* Emit the ZADDs needed to rebuild the sorted set */
char cmd[]="*4\r\n$4\r\nZADD\r\n";
if (o->encoding == REDIS_ENCODING_ZIPLIST) {
unsigned char *zl = o->ptr;
unsigned char *eptr, *sptr;
unsigned char *vstr;
unsigned int vlen;
long long vll;
double score;
eptr = ziplistIndex(zl,0);
redisAssert(eptr != NULL);
sptr = ziplistNext(zl,eptr);
redisAssert(sptr != NULL);
while (eptr != NULL) {
redisAssert(ziplistGet(eptr,&vstr,&vlen,&vll));
score = zzlGetScore(sptr);
if (fwrite(cmd,sizeof(cmd)-1,1,fp) == 0) goto werr;
if (fwriteBulkObject(fp,&key) == 0) goto werr;
if (fwriteBulkDouble(fp,score) == 0) goto werr;
if (vstr != NULL) {
if (fwriteBulkString(fp,(char*)vstr,vlen) == 0)
goto werr;
} else {
if (fwriteBulkLongLong(fp,vll) == 0)
goto werr;
}
zzlNext(zl,&eptr,&sptr);
}
} else if (o->encoding == REDIS_ENCODING_SKIPLIST) {
zset *zs = o->ptr; zset *zs = o->ptr;
dictIterator *di = dictGetIterator(zs->dict); dictIterator *di = dictGetIterator(zs->dict);
dictEntry *de; dictEntry *de;
while((de = dictNext(di)) != NULL) { while((de = dictNext(di)) != NULL) {
char cmd[]="*4\r\n$4\r\nZADD\r\n";
robj *eleobj = dictGetEntryKey(de); robj *eleobj = dictGetEntryKey(de);
double *score = dictGetEntryVal(de); double *score = dictGetEntryVal(de);
...@@ -444,6 +477,9 @@ int rewriteAppendOnlyFile(char *filename) { ...@@ -444,6 +477,9 @@ int rewriteAppendOnlyFile(char *filename) {
if (fwriteBulkObject(fp,eleobj) == 0) goto werr; if (fwriteBulkObject(fp,eleobj) == 0) goto werr;
} }
dictReleaseIterator(di); dictReleaseIterator(di);
} else {
redisPanic("Unknown sorted set encoding");
}
} else if (o->type == REDIS_HASH) { } else if (o->type == REDIS_HASH) {
char cmd[]="*4\r\n$4\r\nHSET\r\n"; char cmd[]="*4\r\n$4\r\nHSET\r\n";
......
char *ascii_logo =
" _._ \n"
" _.-``__ ''-._ \n"
" _.-`` `. `_. ''-._ Redis %s (%s/%d) %s bit\n"
" .-`` .-```. ```\\/ _.,_ ''-._ \n"
" ( ' , .-` | `, ) Running in %s mode\n"
" |`-._`-...-` __...-.``-._|'` _.-'| Port: %d\n"
" | `-._ `._ / _.-' | PID: %ld\n"
" `-._ `-._ `-./ _.-' _.-' \n"
" |`-._`-._ `-.__.-' _.-'_.-'| \n"
" | `-._`-._ _.-'_.-' | http://redis.io \n"
" `-._ `-._`-.__.-'_.-' _.-' \n"
" |`-._`-._ `-.__.-' _.-'_.-'| \n"
" | `-._`-._ _.-'_.-' | \n"
" `-._ `-._`-.__.-'_.-' _.-' \n"
" `-._ `-.__.-' _.-' \n"
" `-._ _.-' \n"
" `-.__.-' \n\n";
#include "redis.h"
#include <arpa/inet.h>
#include <fcntl.h>
#include <unistd.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);
sds clusterGenNodesDescription(void);
clusterNode *clusterLookupNode(char *name);
int clusterNodeAddSlave(clusterNode *master, clusterNode *slave);
int clusterAddSlot(clusterNode *n, int slot);
/* -----------------------------------------------------------------------------
* Initialization
* -------------------------------------------------------------------------- */
void clusterGetRandomName(char *p) {
FILE *fp = fopen("/dev/urandom","r");
char *charset = "0123456789abcdef";
int j;
if (fp == NULL || fread(p,REDIS_CLUSTER_NAMELEN,1,fp) == 0) {
for (j = 0; j < REDIS_CLUSTER_NAMELEN; j++)
p[j] = rand();
}
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");
char *line;
int maxline, j;
if (fp == NULL) return REDIS_ERR;
/* Parse the file. Note that single liens of the cluster config file can
* be really long as they include all the hash slots of the node.
* This means in the worst possible case REDIS_CLUSTER_SLOTS/2 integers.
* To simplify we allocate 1024+REDIS_CLUSTER_SLOTS*16 bytes per line. */
maxline = 1024+REDIS_CLUSTER_SLOTS*16;
line = zmalloc(maxline);
while(fgets(line,maxline,fp) != NULL) {
int argc;
sds *argv = sdssplitargs(line,&argc);
clusterNode *n, *master;
char *p, *s;
/* Create this node if it does not exist */
n = clusterLookupNode(argv[0]);
if (!n) {
n = createClusterNode(argv[0],0);
clusterAddNode(n);
}
/* Address and port */
if ((p = strchr(argv[1],':')) == NULL) goto fmterr;
*p = '\0';
memcpy(n->ip,argv[1],strlen(argv[1])+1);
n->port = atoi(p+1);
/* Parse flags */
p = s = argv[2];
while(p) {
p = strchr(s,',');
if (p) *p = '\0';
if (!strcasecmp(s,"myself")) {
redisAssert(server.cluster.myself == NULL);
server.cluster.myself = n;
n->flags |= REDIS_NODE_MYSELF;
} else if (!strcasecmp(s,"master")) {
n->flags |= REDIS_NODE_MASTER;
} else if (!strcasecmp(s,"slave")) {
n->flags |= REDIS_NODE_SLAVE;
} else if (!strcasecmp(s,"fail?")) {
n->flags |= REDIS_NODE_PFAIL;
} else if (!strcasecmp(s,"fail")) {
n->flags |= REDIS_NODE_FAIL;
} else if (!strcasecmp(s,"handshake")) {
n->flags |= REDIS_NODE_HANDSHAKE;
} else if (!strcasecmp(s,"noaddr")) {
n->flags |= REDIS_NODE_NOADDR;
} else if (!strcasecmp(s,"noflags")) {
/* nothing to do */
} else {
redisPanic("Unknown flag in redis cluster config file");
}
if (p) s = p+1;
}
/* Get master if any. Set the master and populate master's
* slave list. */
if (argv[3][0] != '-') {
master = clusterLookupNode(argv[3]);
if (!master) {
master = createClusterNode(argv[3],0);
clusterAddNode(master);
}
n->slaveof = master;
clusterNodeAddSlave(master,n);
}
/* Set ping sent / pong received timestamps */
if (atoi(argv[4])) n->ping_sent = time(NULL);
if (atoi(argv[5])) n->pong_received = time(NULL);
/* Populate hash slots served by this instance. */
for (j = 7; j < argc; j++) {
int start, stop;
if (argv[j][0] == '[') {
/* Here we handle migrating / importing slots */
int slot;
char direction;
clusterNode *cn;
p = strchr(argv[j],'-');
redisAssert(p != NULL);
*p = '\0';
direction = p[1]; /* Either '>' or '<' */
slot = atoi(argv[j]+1);
p += 3;
cn = clusterLookupNode(p);
if (!cn) {
cn = createClusterNode(p,0);
clusterAddNode(cn);
}
if (direction == '>') {
server.cluster.migrating_slots_to[slot] = cn;
} else {
server.cluster.importing_slots_from[slot] = cn;
}
continue;
} else if ((p = strchr(argv[j],'-')) != NULL) {
*p = '\0';
start = atoi(argv[j]);
stop = atoi(p+1);
} else {
start = stop = atoi(argv[j]);
}
while(start <= stop) clusterAddSlot(n, start++);
}
sdssplitargs_free(argv,argc);
}
zfree(line);
fclose(fp);
/* Config sanity check */
redisAssert(server.cluster.myself != NULL);
redisLog(REDIS_NOTICE,"Node configuration loaded, I'm %.40s",
server.cluster.myself->name);
clusterUpdateState();
return REDIS_OK;
fmterr:
redisLog(REDIS_WARNING,"Unrecovarable error: corrupted cluster config file.");
fclose(fp);
exit(1);
}
/* Cluster node configuration is exactly the same as CLUSTER NODES output.
*
* This function writes the node config and returns 0, on error -1
* is returned. */
int clusterSaveConfig(void) {
sds ci = clusterGenNodesDescription();
int fd;
if ((fd = open(server.cluster.configfile,O_WRONLY|O_CREAT|O_TRUNC,0644))
== -1) goto err;
if (write(fd,ci,sdslen(ci)) != (ssize_t)sdslen(ci)) goto err;
close(fd);
sdsfree(ci);
return 0;
err:
sdsfree(ci);
return -1;
}
void clusterSaveConfigOrDie(void) {
if (clusterSaveConfig() == -1) {
redisLog(REDIS_WARNING,"Fatal: can't update cluster config file.");
exit(1);
}
}
void clusterInit(void) {
int saveconf = 0;
server.cluster.myself = NULL;
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(server.cluster.configfile) == REDIS_ERR) {
/* No configuration found. We will just use the random name provided
* by the createClusterNode() function. */
server.cluster.myself = createClusterNode(NULL,REDIS_NODE_MYSELF);
redisLog(REDIS_NOTICE,"No cluster configuration found, I'm %.40s",
server.cluster.myself->name);
clusterAddNode(server.cluster.myself);
saveconf = 1;
}
if (saveconf) clusterSaveConfigOrDie();
/* 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");
server.cluster.slots_to_keys = zslCreate();
}
/* -----------------------------------------------------------------------------
* 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();
clusterSaveConfigOrDie();
}
} 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) {
int update_config = 0;
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);
update_config = 1;
}
/* Get info from the gossip section */
clusterProcessGossipSection(hdr,link);
/* Anyway reply with a PONG */
clusterSendPing(link,CLUSTERMSG_TYPE_PONG);
/* Update config if needed */
if (update_config) clusterSaveConfigOrDie();
} else if (type == CLUSTERMSG_TYPE_PONG) {
int update_state = 0;
int update_config = 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;
update_config = 1;
} 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);
update_config = 1;
/* 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_state = update_config = 1;
}
}
}
}
}
/* Get info from the gossip section */
clusterProcessGossipSection(hdr,link);
/* Update the cluster state if needed */
if (update_state) clusterUpdateState();
if (update_config) clusterSaveConfigOrDie();
} else if (type == CLUSTERMSG_TYPE_FAIL && sender) {
clusterNode *failing;
failing = clusterLookupNode(hdr->data.fail.about.nodename);
if (failing && !(failing->flags & (REDIS_NODE_FAIL|REDIS_NODE_MYSELF)))
{
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();
clusterSaveConfigOrDie();
}
} 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", 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))
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 (delay < server.cluster.node_timeout) {
/* 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).
*
* The FAIL condition is also reversible if there are no slaves
* for this host, so no slave election should be in progress.
*
* TODO: consider all the implications of resurrecting a
* FAIL node. */
if (node->flags & REDIS_NODE_PFAIL) {
node->flags &= ~REDIS_NODE_PFAIL;
} else if (node->flags & REDIS_NODE_FAIL && !node->numslaves) {
node->flags &= ~REDIS_NODE_FAIL;
clusterUpdateState();
}
} else {
/* Timeout reached. Set the noad se possibly failing if it is
* not already in this state. */
if (!(node->flags & (REDIS_NODE_PFAIL|REDIS_NODE_FAIL))) {
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) {
if (clusterNodeSetSlotBit(n,slot) != 0)
return REDIS_ERR;
server.cluster.slots[slot] = n;
return REDIS_OK;
}
/* Delete the specified slot marking it as unassigned.
* Returns REDIS_OK if the slot was assigned, otherwise if the slot was
* already unassigned REDIS_ERR is returned. */
int clusterDelSlot(int slot) {
clusterNode *n = server.cluster.slots[slot];
if (!n) return REDIS_ERR;
redisAssert(clusterNodeClearSlotBit(n,slot) == 1);
server.cluster.slots[slot] = NULL;
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
* -------------------------------------------------------------------------- */
sds clusterGenNodesDescription(void) {
sds ci = sdsempty();
dictIterator *di;
dictEntry *de;
int j, start;
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",
(long) node->ping_sent,
(long) node->pong_received,
node->link ? "connected" : "disconnected");
/* Slots served by this instance */
start = -1;
for (j = 0; j < REDIS_CLUSTER_SLOTS; j++) {
int bit;
if ((bit = clusterNodeGetSlotBit(node,j)) != 0) {
if (start == -1) start = j;
}
if (start != -1 && (!bit || j == REDIS_CLUSTER_SLOTS-1)) {
if (j == REDIS_CLUSTER_SLOTS-1) j++;
if (start == j-1) {
ci = sdscatprintf(ci," %d",start);
} else {
ci = sdscatprintf(ci," %d-%d",start,j-1);
}
start = -1;
}
}
/* Just for MYSELF node we also dump info about slots that
* we are migrating to other instances or importing from other
* instances. */
if (node->flags & REDIS_NODE_MYSELF) {
for (j = 0; j < REDIS_CLUSTER_SLOTS; j++) {
if (server.cluster.migrating_slots_to[j]) {
ci = sdscatprintf(ci," [%d->-%.40s]",j,
server.cluster.migrating_slots_to[j]->name);
} else if (server.cluster.importing_slots_from[j]) {
ci = sdscatprintf(ci," [%d-<-%.40s]",j,
server.cluster.importing_slots_from[j]->name);
}
}
}
ci = sdscatlen(ci,"\n",1);
}
dictReleaseIterator(di);
return ci;
}
int getSlotOrReply(redisClient *c, robj *o) {
long long slot;
if (getLongLongFromObject(o,&slot) != REDIS_OK ||
slot < 0 || slot > REDIS_CLUSTER_SLOTS)
{
addReplyError(c,"Invalid or out of range slot");
return -1;
}
return (int) slot;
}
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) {
robj *o;
sds ci = clusterGenNodesDescription();
o = createObject(REDIS_STRING,ci);
addReplyBulk(c,o);
decrRefCount(o);
} else if ((!strcasecmp(c->argv[1]->ptr,"addslots") ||
!strcasecmp(c->argv[1]->ptr,"delslots")) && c->argc >= 3) {
int j, slot;
unsigned char *slots = zmalloc(REDIS_CLUSTER_SLOTS);
int del = !strcasecmp(c->argv[1]->ptr,"delslots");
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 ((slot = getSlotOrReply(c,c->argv[j])) == -1) {
zfree(slots);
return;
}
if (del && server.cluster.slots[slot] == NULL) {
addReplyErrorFormat(c,"Slot %d is already unassigned", slot);
zfree(slots);
return;
} else if (!del && server.cluster.slots[slot]) {
addReplyErrorFormat(c,"Slot %d 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;
/* If this slot was set as importing we can clear this
* state as now we are the real owner of the slot. */
if (server.cluster.importing_slots_from[j])
server.cluster.importing_slots_from[j] = NULL;
retval = del ? clusterDelSlot(j) :
clusterAddSlot(server.cluster.myself,j);
redisAssert(retval == REDIS_OK);
}
}
zfree(slots);
clusterUpdateState();
clusterSaveConfigOrDie();
addReply(c,shared.ok);
} else if (!strcasecmp(c->argv[1]->ptr,"setslot") && c->argc >= 4) {
/* SETSLOT 10 MIGRATING <instance ID> */
/* SETSLOT 10 IMPORTING <instance ID> */
/* SETSLOT 10 STABLE */
int slot;
clusterNode *n;
if ((slot = getSlotOrReply(c,c->argv[2])) == -1) return;
if (!strcasecmp(c->argv[3]->ptr,"migrating") && c->argc == 5) {
if (server.cluster.slots[slot] != server.cluster.myself) {
addReplyErrorFormat(c,"I'm not the owner of hash slot %u",slot);
return;
}
if ((n = clusterLookupNode(c->argv[4]->ptr)) == NULL) {
addReplyErrorFormat(c,"I don't know about node %s",
(char*)c->argv[4]->ptr);
return;
}
server.cluster.migrating_slots_to[slot] = n;
} else if (!strcasecmp(c->argv[3]->ptr,"importing") && c->argc == 5) {
if (server.cluster.slots[slot] == server.cluster.myself) {
addReplyErrorFormat(c,
"I'm already the owner of hash slot %u",slot);
return;
}
if ((n = clusterLookupNode(c->argv[4]->ptr)) == NULL) {
addReplyErrorFormat(c,"I don't know about node %s",
(char*)c->argv[3]->ptr);
return;
}
server.cluster.importing_slots_from[slot] = n;
} else if (!strcasecmp(c->argv[3]->ptr,"stable") && c->argc == 4) {
/* CLUSTER SETSLOT <SLOT> STABLE */
server.cluster.importing_slots_from[slot] = NULL;
server.cluster.migrating_slots_to[slot] = NULL;
} else if (!strcasecmp(c->argv[3]->ptr,"node") && c->argc == 4) {
/* CLUSTER SETSLOT <SLOT> NODE <NODE ID> */
clusterNode *n = clusterLookupNode(c->argv[4]->ptr);
if (!n) addReplyErrorFormat(c,"Unknown node %s",
(char*)c->argv[4]->ptr);
/* If this hash slot was served by 'myself' before to switch
* make sure there are no longer local keys for this hash slot. */
if (server.cluster.slots[slot] == server.cluster.myself &&
n != server.cluster.myself)
{
int numkeys;
robj **keys;
keys = zmalloc(sizeof(robj*)*1);
numkeys = GetKeysInSlot(slot, keys, 1);
zfree(keys);
if (numkeys == 0) {
addReplyErrorFormat(c, "Can't assign hashslot %d to a different node while I still hold keys for this hash slot.", slot);
return;
}
}
/* If this node was the slot owner and the slot was marked as
* migrating, assigning the slot to another node will clear
* the migratig status. */
if (server.cluster.slots[slot] == server.cluster.myself &&
server.cluster.migrating_slots_to[slot])
server.cluster.migrating_slots_to[slot] = NULL;
clusterDelSlot(slot);
clusterAddSlot(n,slot);
} else {
addReplyError(c,"Invalid CLUSTER SETSLOT action or number of arguments");
return;
}
clusterSaveConfigOrDie();
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"
"cluster_known_nodes:%lu\r\n"
, statestr[server.cluster.state],
slots_assigned,
slots_ok,
slots_pfail,
slots_fail,
dictSize(server.cluster.nodes)
);
addReplySds(c,sdscatprintf(sdsempty(),"$%lu\r\n",
(unsigned long)sdslen(info)));
addReplySds(c,info);
addReply(c,shared.crlf);
} else if (!strcasecmp(c->argv[1]->ptr,"keyslot") && c->argc == 3) {
sds key = c->argv[2]->ptr;
addReplyLongLong(c,keyHashSlot(key,sdslen(key)));
} else if (!strcasecmp(c->argv[1]->ptr,"getkeysinslot") && c->argc == 4) {
long long maxkeys, slot;
unsigned int numkeys, j;
robj **keys;
if (getLongLongFromObjectOrReply(c,c->argv[2],&slot,NULL) != REDIS_OK)
return;
if (getLongLongFromObjectOrReply(c,c->argv[3],&maxkeys,NULL) != REDIS_OK)
return;
if (slot < 0 || slot >= REDIS_CLUSTER_SLOTS || maxkeys < 0 ||
maxkeys > 1024*1024) {
addReplyError(c,"Invalid slot or number of keys");
return;
}
keys = zmalloc(sizeof(robj*)*maxkeys);
numkeys = GetKeysInSlot(slot, keys, maxkeys);
addReplyMultiBulkLen(c,numkeys);
for (j = 0; j < numkeys; j++) addReplyBulk(c,keys[j]);
zfree(keys);
} 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. */
if ((data[0] > 4 && data[0] < 9) ||
data[0] > 11 ||
(o = rdbLoadObject(data[0],fp)) == 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);
return;
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);
return;
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);
return;
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);
return;
}
/* DUMP keyname
* DUMP is actually not used by Redis Cluster but it is the obvious
* complement of RESTORE and can be useful for different applications. */
void dumpCommand(redisClient *c) {
char buf[64];
FILE *fp;
robj *o, *dumpobj;
sds dump = NULL;
off_t payload_len;
unsigned int type;
/* Check if the key is here. */
if ((o = lookupKeyRead(c->db,c->argv[1])) == NULL) {
addReply(c,shared.nullbulk);
return;
}
/* Create temp file */
snprintf(buf,sizeof(buf),"redis-dump-%d.tmp",getpid());
fp = fopen(buf,"w+");
if (!fp) {
redisLog(REDIS_WARNING,"Can't open tmp file for MIGRATE: %s",
strerror(errno));
addReplyErrorFormat(c,"DUMP failed, tmp file creation error: %s.",
strerror(errno));
return;
}
unlink(buf);
/* Dump the serailized object and read it back in memory.
* We prefix it with a one byte containing the type ID.
* This is the serialization format understood by RESTORE. */
if (rdbSaveObject(fp,o) == -1) goto file_wr_err;
payload_len = ftello(fp);
if (fseeko(fp,0,SEEK_SET) == -1) goto file_rd_err;
dump = sdsnewlen(NULL,payload_len+1);
if (payload_len && fread(dump+1,payload_len,1,fp) != 1) goto file_rd_err;
fclose(fp);
type = o->type;
if (type == REDIS_LIST && o->encoding == REDIS_ENCODING_ZIPLIST)
type = REDIS_LIST_ZIPLIST;
else if (type == REDIS_HASH && o->encoding == REDIS_ENCODING_ZIPMAP)
type = REDIS_HASH_ZIPMAP;
else if (type == REDIS_SET && o->encoding == REDIS_ENCODING_INTSET)
type = REDIS_SET_INTSET;
else
type = o->type;
dump[0] = type;
/* Transfer to the client */
dumpobj = createObject(REDIS_STRING,dump);
addReplyBulk(c,dumpobj);
decrRefCount(dumpobj);
return;
file_wr_err:
redisLog(REDIS_WARNING,"Can't write on tmp file for DUMP: %s",
strerror(errno));
addReplyErrorFormat(c,"DUMP failed, tmp file write error: %s.",
strerror(errno));
sdsfree(dump);
fclose(fp);
return;
file_rd_err:
redisLog(REDIS_WARNING,"Can't read from tmp file for DUMP: %s",
strerror(errno));
addReplyErrorFormat(c,"DUMP failed, tmp file read error: %s.",
strerror(errno));
sdsfree(dump);
fclose(fp);
return;
}
/* -----------------------------------------------------------------------------
* 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 the returned node should be used only for this request, the *ask
* integer is set to '1', otherwise to '0'. This is used in order to
* let the caller know if we should reply with -MOVED or with -ASK.
*
* If the request contains more than a single key NULL is returned,
* however a request with more then a key argument where the key is always
* the same is valid, like in: RPOPLPUSH mylist mylist.*/
clusterNode *getNodeByQuery(redisClient *c, struct redisCommand *cmd, robj **argv, int argc, int *hashslot, int *ask) {
clusterNode *n = NULL;
robj *firstkey = NULL;
multiState *ms, _ms;
multiCmd mc;
int i, slot = 0;
/* 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 {
/* In order to have a single codepath create a fake Multi State
* structure if the client is not in MULTI/EXEC state, this way
* we have a single codepath below. */
ms = &_ms;
_ms.commands = &mc;
_ms.count = 1;
mc.argv = argv;
mc.argc = argc;
mc.cmd = cmd;
}
/* Check that all the keys are the same key, and get the slot and
* node for this key. */
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_ALL);
for (j = 0; j < numkeys; j++) {
if (firstkey == NULL) {
/* This is the first key we see. Check what is the slot
* and node. */
firstkey = margv[keyindex[j]];
slot = keyHashSlot((char*)firstkey->ptr, sdslen(firstkey->ptr));
n = server.cluster.slots[slot];
redisAssert(n != NULL);
} else {
/* If it is not the first key, make sure it is exactly
* the same key as the first we saw. */
if (!equalStringObjects(firstkey,margv[keyindex[j]])) {
decrRefCount(firstkey);
getKeysFreeResult(keyindex);
return NULL;
}
}
}
getKeysFreeResult(keyindex);
}
if (ask) *ask = 0; /* This is the default. Set to 1 if needed later. */
/* No key at all in command? then we can serve the request
* without redirections. */
if (n == NULL) return server.cluster.myself;
if (hashslot) *hashslot = slot;
/* This request is about a slot we are migrating into another instance?
* Then we need to check if we have the key. If we have it we can reply.
* If instead is a new key, we pass the request to the node that is
* receiving the slot. */
if (n == server.cluster.myself &&
server.cluster.migrating_slots_to[slot] != NULL)
{
if (lookupKeyRead(&server.db[0],firstkey) == NULL) {
if (ask) *ask = 1;
return server.cluster.migrating_slots_to[slot];
}
}
/* Handle the case in which we are receiving this hash slot from
* another instance, so we'll accept the query even if in the table
* it is assigned to a different node. */
if (server.cluster.importing_slots_from[slot] != NULL)
return server.cluster.myself;
/* It's not a -ASK case. Base case: just return the right node. */
return n;
}
...@@ -261,6 +261,10 @@ void loadServerConfig(char *filename) { ...@@ -261,6 +261,10 @@ void loadServerConfig(char *filename) {
server.list_max_ziplist_value = memtoll(argv[1], NULL); server.list_max_ziplist_value = memtoll(argv[1], NULL);
} else if (!strcasecmp(argv[0],"set-max-intset-entries") && argc == 2) { } else if (!strcasecmp(argv[0],"set-max-intset-entries") && argc == 2) {
server.set_max_intset_entries = memtoll(argv[1], NULL); server.set_max_intset_entries = memtoll(argv[1], NULL);
} else if (!strcasecmp(argv[0],"zset-max-ziplist-entries") && argc == 2) {
server.zset_max_ziplist_entries = memtoll(argv[1], NULL);
} else if (!strcasecmp(argv[0],"zset-max-ziplist-value") && argc == 2) {
server.zset_max_ziplist_value = memtoll(argv[1], NULL);
} else if (!strcasecmp(argv[0],"rename-command") && argc == 3) { } else if (!strcasecmp(argv[0],"rename-command") && argc == 3) {
struct redisCommand *cmd = lookupCommand(argv[1]); struct redisCommand *cmd = lookupCommand(argv[1]);
int retval; int retval;
...@@ -285,6 +289,13 @@ void loadServerConfig(char *filename) { ...@@ -285,6 +289,13 @@ void loadServerConfig(char *filename) {
err = "Target command name already exists"; goto loaderr; 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 if (!strcasecmp(argv[0],"cluster-config-file") && argc == 2) {
zfree(server.cluster.configfile);
server.cluster.configfile = zstrdup(argv[1]);
} else { } else {
err = "Bad directive or wrong number of arguments"; goto loaderr; err = "Bad directive or wrong number of arguments"; goto loaderr;
} }
...@@ -443,6 +454,12 @@ void configSetCommand(redisClient *c) { ...@@ -443,6 +454,12 @@ void configSetCommand(redisClient *c) {
} else if (!strcasecmp(c->argv[2]->ptr,"set-max-intset-entries")) { } else if (!strcasecmp(c->argv[2]->ptr,"set-max-intset-entries")) {
if (getLongLongFromObject(o,&ll) == REDIS_ERR || ll < 0) goto badfmt; if (getLongLongFromObject(o,&ll) == REDIS_ERR || ll < 0) goto badfmt;
server.set_max_intset_entries = ll; server.set_max_intset_entries = ll;
} else if (!strcasecmp(c->argv[2]->ptr,"zset-max-ziplist-entries")) {
if (getLongLongFromObject(o,&ll) == REDIS_ERR || ll < 0) goto badfmt;
server.zset_max_ziplist_entries = ll;
} else if (!strcasecmp(c->argv[2]->ptr,"zset-max-ziplist-value")) {
if (getLongLongFromObject(o,&ll) == REDIS_ERR || ll < 0) goto badfmt;
server.zset_max_ziplist_value = ll;
} else { } else {
addReplyErrorFormat(c,"Unsupported CONFIG parameter: %s", addReplyErrorFormat(c,"Unsupported CONFIG parameter: %s",
(char*)c->argv[2]->ptr); (char*)c->argv[2]->ptr);
...@@ -594,6 +611,16 @@ void configGetCommand(redisClient *c) { ...@@ -594,6 +611,16 @@ void configGetCommand(redisClient *c) {
addReplyBulkLongLong(c,server.set_max_intset_entries); addReplyBulkLongLong(c,server.set_max_intset_entries);
matches++; matches++;
} }
if (stringmatch(pattern,"zset-max-ziplist-entries",0)) {
addReplyBulkCString(c,"zset-max-ziplist-entries");
addReplyBulkLongLong(c,server.zset_max_ziplist_entries);
matches++;
}
if (stringmatch(pattern,"zset-max-ziplist-value",0)) {
addReplyBulkCString(c,"zset-max-ziplist-value");
addReplyBulkLongLong(c,server.zset_max_ziplist_value);
matches++;
}
setDeferredMultiBulkLength(c,replylen,matches*2); setDeferredMultiBulkLength(c,replylen,matches*2);
} }
......
...@@ -10,18 +10,31 @@ ...@@ -10,18 +10,31 @@
* this expects a different allocation scheme. Therefore, *exclusively* use * this expects a different allocation scheme. Therefore, *exclusively* use
* either tcmalloc or OSX's malloc_size()! */ * either tcmalloc or OSX's malloc_size()! */
#if defined(USE_TCMALLOC) #if defined(USE_TCMALLOC)
#define REDIS_MALLOC "tcmalloc"
#include <google/tcmalloc.h> #include <google/tcmalloc.h>
#if TC_VERSION_MAJOR >= 1 && TC_VERSION_MINOR >= 6 #if TC_VERSION_MAJOR >= 1 && TC_VERSION_MINOR >= 6
#define HAVE_MALLOC_SIZE 1 #define HAVE_MALLOC_SIZE 1
#define redis_malloc_size(p) tc_malloc_size(p) #define redis_malloc_size(p) tc_malloc_size(p)
#endif #endif
#elif defined(USE_JEMALLOC)
#define REDIS_MALLOC "jemalloc"
#define JEMALLOC_MANGLE
#include <jemalloc/jemalloc.h>
#if JEMALLOC_VERSION_MAJOR >= 2 && JEMALLOC_VERSION_MINOR >= 1
#define HAVE_MALLOC_SIZE 1
#define redis_malloc_size(p) JEMALLOC_P(malloc_usable_size)(p)
#endif
#elif defined(__APPLE__) #elif defined(__APPLE__)
#include <malloc/malloc.h> #include <malloc/malloc.h>
#define HAVE_MALLOC_SIZE 1 #define HAVE_MALLOC_SIZE 1
#define redis_malloc_size(p) malloc_size(p) #define redis_malloc_size(p) malloc_size(p)
#endif #endif
/* Tefine redis_fstat to fstat or fstat64() */ #ifndef REDIS_MALLOC
#define REDIS_MALLOC "libc"
#endif
/* Define redis_fstat to fstat or fstat64() */
#if defined(__APPLE__) && !defined(MAC_OS_X_VERSION_10_6) #if defined(__APPLE__) && !defined(MAC_OS_X_VERSION_10_6)
#define redis_fstat fstat64 #define redis_fstat fstat64
#define redis_stat stat64 #define redis_stat stat64
......
#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;
}
...@@ -2,6 +2,9 @@ ...@@ -2,6 +2,9 @@
#include <signal.h> #include <signal.h>
void SlotToKeyAdd(robj *key);
void SlotToKeyDel(robj *key);
/*----------------------------------------------------------------------------- /*-----------------------------------------------------------------------------
* C-level DB API * C-level DB API
*----------------------------------------------------------------------------*/ *----------------------------------------------------------------------------*/
...@@ -131,6 +134,7 @@ int dbAdd(redisDb *db, robj *key, robj *val) { ...@@ -131,6 +134,7 @@ int dbAdd(redisDb *db, robj *key, robj *val) {
sds copy = sdsdup(key->ptr); sds copy = sdsdup(key->ptr);
dictAdd(db->dict, copy, val); dictAdd(db->dict, copy, val);
if (server.ds_enabled) cacheSetKeyMayExist(db,key); if (server.ds_enabled) cacheSetKeyMayExist(db,key);
if (server.cluster_enabled) SlotToKeyAdd(key);
return REDIS_OK; return REDIS_OK;
} }
} }
...@@ -146,6 +150,7 @@ int dbReplace(redisDb *db, robj *key, robj *val) { ...@@ -146,6 +150,7 @@ int dbReplace(redisDb *db, robj *key, robj *val) {
if ((oldval = dictFetchValue(db->dict,key->ptr)) == NULL) { if ((oldval = dictFetchValue(db->dict,key->ptr)) == NULL) {
sds copy = sdsdup(key->ptr); sds copy = sdsdup(key->ptr);
dictAdd(db->dict, copy, val); dictAdd(db->dict, copy, val);
if (server.cluster_enabled) SlotToKeyAdd(key);
retval = 1; retval = 1;
} else { } else {
dictReplace(db->dict, key->ptr, val); dictReplace(db->dict, key->ptr, val);
...@@ -198,7 +203,12 @@ int dbDelete(redisDb *db, robj *key) { ...@@ -198,7 +203,12 @@ int dbDelete(redisDb *db, robj *key) {
/* Deleting an entry from the expires dict will not free the sds of /* Deleting an entry from the expires dict will not free the sds of
* the key, because it is shared with the main dictionary. */ * the key, because it is shared with the main dictionary. */
if (dictSize(db->expires) > 0) dictDelete(db->expires,key->ptr); if (dictSize(db->expires) > 0) dictDelete(db->expires,key->ptr);
return dictDelete(db->dict,key->ptr) == DICT_OK; if (dictDelete(db->dict,key->ptr) == DICT_OK) {
if (server.cluster_enabled) SlotToKeyDel(key);
return 1;
} else {
return 0;
}
} }
/* Empty the whole database. /* Empty the whole database.
...@@ -307,6 +317,10 @@ void existsCommand(redisClient *c) { ...@@ -307,6 +317,10 @@ void existsCommand(redisClient *c) {
void selectCommand(redisClient *c) { void selectCommand(redisClient *c) {
int id = atoi(c->argv[1]->ptr); 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) { if (selectDb(c,id) == REDIS_ERR) {
addReplyError(c,"invalid DB index"); addReplyError(c,"invalid DB index");
} else { } else {
...@@ -428,6 +442,11 @@ void moveCommand(redisClient *c) { ...@@ -428,6 +442,11 @@ void moveCommand(redisClient *c) {
redisDb *src, *dst; redisDb *src, *dst;
int srcid; int srcid;
if (server.cluster_enabled) {
addReplyError(c,"MOVE is not allowed in cluster mode");
return;
}
/* Obtain source and target DB pointers */ /* Obtain source and target DB pointers */
src = c->db; src = c->db;
srcid = c->db->id; srcid = c->db->id;
...@@ -616,3 +635,108 @@ void persistCommand(redisClient *c) { ...@@ -616,3 +635,108 @@ void persistCommand(redisClient *c) {
} }
} }
} }
/* -----------------------------------------------------------------------------
* API to get key arguments from commands
* ---------------------------------------------------------------------------*/
int *getKeysUsingCommandTable(struct redisCommand *cmd,robj **argv, int argc, int *numkeys) {
int j, i = 0, last, *keys;
REDIS_NOTUSED(argv);
if (cmd->firstkey == 0) {
*numkeys = 0;
return NULL;
}
last = cmd->lastkey;
if (last < 0) last = argc+last;
keys = zmalloc(sizeof(int)*((last - cmd->firstkey)+1));
for (j = cmd->firstkey; j <= last; j += cmd->keystep) {
redisAssert(j < argc);
keys[i++] = j;
}
*numkeys = i;
return keys;
}
int *getKeysFromCommand(struct redisCommand *cmd,robj **argv, int argc, int *numkeys, int flags) {
if (cmd->getkeys_proc) {
return cmd->getkeys_proc(cmd,argv,argc,numkeys,flags);
} else {
return getKeysUsingCommandTable(cmd,argv,argc,numkeys);
}
}
void getKeysFreeResult(int *result) {
zfree(result);
}
int *noPreloadGetKeys(struct redisCommand *cmd,robj **argv, int argc, int *numkeys, int flags) {
if (flags & REDIS_GETKEYS_PRELOAD) {
*numkeys = 0;
return NULL;
} else {
return getKeysUsingCommandTable(cmd,argv,argc,numkeys);
}
}
int *renameGetKeys(struct redisCommand *cmd,robj **argv, int argc, int *numkeys, int flags) {
if (flags & REDIS_GETKEYS_PRELOAD) {
int *keys = zmalloc(sizeof(int));
*numkeys = 1;
keys[0] = 1;
return keys;
} else {
return getKeysUsingCommandTable(cmd,argv,argc,numkeys);
}
}
int *zunionInterGetKeys(struct redisCommand *cmd,robj **argv, int argc, int *numkeys, int flags) {
int i, num, *keys;
REDIS_NOTUSED(cmd);
REDIS_NOTUSED(flags);
num = atoi(argv[2]->ptr);
/* Sanity check. Don't return any key if the command is going to
* reply with syntax error. */
if (num > (argc-3)) {
*numkeys = 0;
return NULL;
}
keys = zmalloc(sizeof(int)*num);
for (i = 0; i < num; i++) keys[i] = 3+i;
*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;
}
...@@ -100,7 +100,7 @@ void computeDatasetDigest(unsigned char *final) { ...@@ -100,7 +100,7 @@ void computeDatasetDigest(unsigned char *final) {
mixDigest(digest,key,sdslen(key)); mixDigest(digest,key,sdslen(key));
/* Make sure the key is loaded if VM is active */ /* Make sure the key is loaded if VM is active */
o = lookupKeyRead(db,keyobj); o = dictGetEntryVal(de);
aux = htonl(o->type); aux = htonl(o->type);
mixDigest(digest,&aux,sizeof(aux)); mixDigest(digest,&aux,sizeof(aux));
...@@ -127,6 +127,39 @@ void computeDatasetDigest(unsigned char *final) { ...@@ -127,6 +127,39 @@ void computeDatasetDigest(unsigned char *final) {
} }
setTypeReleaseIterator(si); setTypeReleaseIterator(si);
} else if (o->type == REDIS_ZSET) { } else if (o->type == REDIS_ZSET) {
unsigned char eledigest[20];
if (o->encoding == REDIS_ENCODING_ZIPLIST) {
unsigned char *zl = o->ptr;
unsigned char *eptr, *sptr;
unsigned char *vstr;
unsigned int vlen;
long long vll;
double score;
eptr = ziplistIndex(zl,0);
redisAssert(eptr != NULL);
sptr = ziplistNext(zl,eptr);
redisAssert(sptr != NULL);
while (eptr != NULL) {
redisAssert(ziplistGet(eptr,&vstr,&vlen,&vll));
score = zzlGetScore(sptr);
memset(eledigest,0,20);
if (vstr != NULL) {
mixDigest(eledigest,vstr,vlen);
} else {
ll2string(buf,sizeof(buf),vll);
mixDigest(eledigest,buf,strlen(buf));
}
snprintf(buf,sizeof(buf),"%.17g",score);
mixDigest(eledigest,buf,strlen(buf));
xorDigest(digest,eledigest,20);
zzlNext(zl,&eptr,&sptr);
}
} else if (o->encoding == REDIS_ENCODING_SKIPLIST) {
zset *zs = o->ptr; zset *zs = o->ptr;
dictIterator *di = dictGetIterator(zs->dict); dictIterator *di = dictGetIterator(zs->dict);
dictEntry *de; dictEntry *de;
...@@ -134,7 +167,6 @@ void computeDatasetDigest(unsigned char *final) { ...@@ -134,7 +167,6 @@ void computeDatasetDigest(unsigned char *final) {
while((de = dictNext(di)) != NULL) { while((de = dictNext(di)) != NULL) {
robj *eleobj = dictGetEntryKey(de); robj *eleobj = dictGetEntryKey(de);
double *score = dictGetEntryVal(de); double *score = dictGetEntryVal(de);
unsigned char eledigest[20];
snprintf(buf,sizeof(buf),"%.17g",*score); snprintf(buf,sizeof(buf),"%.17g",*score);
memset(eledigest,0,20); memset(eledigest,0,20);
...@@ -143,6 +175,9 @@ void computeDatasetDigest(unsigned char *final) { ...@@ -143,6 +175,9 @@ void computeDatasetDigest(unsigned char *final) {
xorDigest(digest,eledigest,20); xorDigest(digest,eledigest,20);
} }
dictReleaseIterator(di); dictReleaseIterator(di);
} else {
redisPanic("Unknown sorted set encoding");
}
} else if (o->type == REDIS_HASH) { } else if (o->type == REDIS_HASH) {
hashTypeIterator *hi; hashTypeIterator *hi;
robj *obj; robj *obj;
......
...@@ -245,9 +245,9 @@ int dictRehashMilliseconds(dict *d, int ms) { ...@@ -245,9 +245,9 @@ int dictRehashMilliseconds(dict *d, int ms) {
} }
/* This function performs just a step of rehashing, and only if there are /* This function performs just a step of rehashing, and only if there are
* not iterators bound to our hash table. When we have iterators in the middle * no safe iterators bound to our hash table. When we have iterators in the
* of a rehashing we can't mess with the two hash tables otherwise some element * middle of a rehashing we can't mess with the two hash tables otherwise
* can be missed or duplicated. * some element can be missed or duplicated.
* *
* This function is called by common lookup or update operations in the * This function is called by common lookup or update operations in the
* dictionary so that the hash table automatically migrates from H1 to H2 * dictionary so that the hash table automatically migrates from H1 to H2
...@@ -424,17 +424,26 @@ dictIterator *dictGetIterator(dict *d) ...@@ -424,17 +424,26 @@ dictIterator *dictGetIterator(dict *d)
iter->d = d; iter->d = d;
iter->table = 0; iter->table = 0;
iter->index = -1; iter->index = -1;
iter->safe = 0;
iter->entry = NULL; iter->entry = NULL;
iter->nextEntry = NULL; iter->nextEntry = NULL;
return iter; return iter;
} }
dictIterator *dictGetSafeIterator(dict *d) {
dictIterator *i = dictGetIterator(d);
i->safe = 1;
return i;
}
dictEntry *dictNext(dictIterator *iter) dictEntry *dictNext(dictIterator *iter)
{ {
while (1) { while (1) {
if (iter->entry == NULL) { if (iter->entry == NULL) {
dictht *ht = &iter->d->ht[iter->table]; dictht *ht = &iter->d->ht[iter->table];
if (iter->index == -1 && iter->table == 0) iter->d->iterators++; if (iter->safe && iter->index == -1 && iter->table == 0)
iter->d->iterators++;
iter->index++; iter->index++;
if (iter->index >= (signed) ht->size) { if (iter->index >= (signed) ht->size) {
if (dictIsRehashing(iter->d) && iter->table == 0) { if (dictIsRehashing(iter->d) && iter->table == 0) {
...@@ -461,7 +470,8 @@ dictEntry *dictNext(dictIterator *iter) ...@@ -461,7 +470,8 @@ dictEntry *dictNext(dictIterator *iter)
void dictReleaseIterator(dictIterator *iter) void dictReleaseIterator(dictIterator *iter)
{ {
if (!(iter->index == -1 && iter->table == 0)) iter->d->iterators--; if (iter->safe && !(iter->index == -1 && iter->table == 0))
iter->d->iterators--;
zfree(iter); zfree(iter);
} }
......
...@@ -74,10 +74,13 @@ typedef struct dict { ...@@ -74,10 +74,13 @@ typedef struct dict {
int iterators; /* number of iterators currently running */ int iterators; /* number of iterators currently running */
} dict; } dict;
/* If safe is set to 1 this is a safe iteartor, that means, you can call
* dictAdd, dictFind, and other functions against the dictionary even while
* iterating. Otherwise it is a non safe iterator, and only dictNext()
* should be called while iterating. */
typedef struct dictIterator { typedef struct dictIterator {
dict *d; dict *d;
int table; int table, index, safe;
int index;
dictEntry *entry, *nextEntry; dictEntry *entry, *nextEntry;
} dictIterator; } dictIterator;
...@@ -132,6 +135,7 @@ dictEntry * dictFind(dict *d, const void *key); ...@@ -132,6 +135,7 @@ dictEntry * dictFind(dict *d, const void *key);
void *dictFetchValue(dict *d, const void *key); void *dictFetchValue(dict *d, const void *key);
int dictResize(dict *d); int dictResize(dict *d);
dictIterator *dictGetIterator(dict *d); dictIterator *dictGetIterator(dict *d);
dictIterator *dictGetSafeIterator(dict *d);
dictEntry *dictNext(dictIterator *iter); dictEntry *dictNext(dictIterator *iter);
void dictReleaseIterator(dictIterator *iter); void dictReleaseIterator(dictIterator *iter);
dictEntry *dictGetRandomKey(dict *d); dictEntry *dictGetRandomKey(dict *d);
......
...@@ -903,59 +903,6 @@ int waitForSwappedKey(redisClient *c, robj *key) { ...@@ -903,59 +903,6 @@ int waitForSwappedKey(redisClient *c, robj *key) {
return 1; return 1;
} }
/* Preload keys for any command with first, last and step values for
* the command keys prototype, as defined in the command table. */
void waitForMultipleSwappedKeys(redisClient *c, struct redisCommand *cmd, int argc, robj **argv) {
int j, last;
if (cmd->vm_firstkey == 0) return;
last = cmd->vm_lastkey;
if (last < 0) last = argc+last;
for (j = cmd->vm_firstkey; j <= last; j += cmd->vm_keystep) {
redisAssert(j < argc);
waitForSwappedKey(c,argv[j]);
}
}
/* Preload keys needed for the ZUNIONSTORE and ZINTERSTORE commands.
* Note that the number of keys to preload is user-defined, so we need to
* apply a sanity check against argc. */
void zunionInterBlockClientOnSwappedKeys(redisClient *c, struct redisCommand *cmd, int argc, robj **argv) {
int i, num;
REDIS_NOTUSED(cmd);
num = atoi(argv[2]->ptr);
if (num > (argc-3)) return;
for (i = 0; i < num; i++) {
waitForSwappedKey(c,argv[3+i]);
}
}
/* Preload keys needed to execute the entire MULTI/EXEC block.
*
* This function is called by blockClientOnSwappedKeys when EXEC is issued,
* and will block the client when any command requires a swapped out value. */
void execBlockClientOnSwappedKeys(redisClient *c, struct redisCommand *cmd, int argc, robj **argv) {
int i, margc;
struct redisCommand *mcmd;
robj **margv;
REDIS_NOTUSED(cmd);
REDIS_NOTUSED(argc);
REDIS_NOTUSED(argv);
if (!(c->flags & REDIS_MULTI)) return;
for (i = 0; i < c->mstate.count; i++) {
mcmd = c->mstate.commands[i].cmd;
margc = c->mstate.commands[i].argc;
margv = c->mstate.commands[i].argv;
if (mcmd->vm_preload_proc != NULL) {
mcmd->vm_preload_proc(c,mcmd,margc,margv);
} else {
waitForMultipleSwappedKeys(c,mcmd,margc,margv);
}
}
}
/* Is this client attempting to run a command against swapped keys? /* Is this client attempting to run a command against swapped keys?
* If so, block it ASAP, load the keys in background, then resume it. * If so, block it ASAP, load the keys in background, then resume it.
* *
...@@ -967,10 +914,39 @@ void execBlockClientOnSwappedKeys(redisClient *c, struct redisCommand *cmd, int ...@@ -967,10 +914,39 @@ void execBlockClientOnSwappedKeys(redisClient *c, struct redisCommand *cmd, int
* Return 1 if the client is marked as blocked, 0 if the client can * Return 1 if the client is marked as blocked, 0 if the client can
* continue as the keys it is going to access appear to be in memory. */ * continue as the keys it is going to access appear to be in memory. */
int blockClientOnSwappedKeys(redisClient *c, struct redisCommand *cmd) { int blockClientOnSwappedKeys(redisClient *c, struct redisCommand *cmd) {
if (cmd->vm_preload_proc != NULL) { int *keyindex, numkeys, j, i;
cmd->vm_preload_proc(c,cmd,c->argc,c->argv);
/* EXEC is a special case, we need to preload all the commands
* queued into the transaction */
if (cmd->proc == execCommand) {
struct redisCommand *mcmd;
robj **margv;
int margc;
if (!(c->flags & REDIS_MULTI)) return 0;
for (i = 0; i < c->mstate.count; i++) {
mcmd = c->mstate.commands[i].cmd;
margc = c->mstate.commands[i].argc;
margv = c->mstate.commands[i].argv;
keyindex = getKeysFromCommand(mcmd,margv,margc,&numkeys,
REDIS_GETKEYS_PRELOAD);
for (j = 0; j < numkeys; j++) {
redisLog(REDIS_DEBUG,"Preloading %s",
(char*)margv[keyindex[j]]->ptr);
waitForSwappedKey(c,margv[keyindex[j]]);
}
getKeysFreeResult(keyindex);
}
} else { } else {
waitForMultipleSwappedKeys(c,cmd,c->argc,c->argv); keyindex = getKeysFromCommand(cmd,c->argv,c->argc,&numkeys,
REDIS_GETKEYS_PRELOAD);
for (j = 0; j < numkeys; j++) {
redisLog(REDIS_DEBUG,"Preloading %s",
(char*)c->argv[keyindex[j]]->ptr);
waitForSwappedKey(c,c->argv[keyindex[j]]);
}
getKeysFreeResult(keyindex);
} }
/* If the client was blocked for at least one key, mark it as blocked. */ /* If the client was blocked for at least one key, mark it as blocked. */
......
...@@ -16,7 +16,6 @@ redisClient *createClient(int fd) { ...@@ -16,7 +16,6 @@ redisClient *createClient(int fd) {
anetNonBlock(NULL,fd); anetNonBlock(NULL,fd);
anetTcpNoDelay(NULL,fd); anetTcpNoDelay(NULL,fd);
if (!c) return NULL;
if (aeCreateFileEvent(server.el,fd,AE_READABLE, if (aeCreateFileEvent(server.el,fd,AE_READABLE,
readQueryFromClient, c) == AE_ERR) readQueryFromClient, c) == AE_ERR)
{ {
...@@ -60,9 +59,6 @@ redisClient *createClient(int fd) { ...@@ -60,9 +59,6 @@ redisClient *createClient(int fd) {
/* Set the event loop to listen for write events on the client's socket. /* Set the event loop to listen for write events on the client's socket.
* Typically gets called every time a reply is built. */ * Typically gets called every time a reply is built. */
int _installWriteEvent(redisClient *c) { int _installWriteEvent(redisClient *c) {
/* When CLOSE_AFTER_REPLY is set, no more replies may be added! */
redisAssert(!(c->flags & REDIS_CLOSE_AFTER_REPLY));
if (c->fd <= 0) return REDIS_ERR; if (c->fd <= 0) return REDIS_ERR;
if (c->bufpos == 0 && listLength(c->reply) == 0 && if (c->bufpos == 0 && listLength(c->reply) == 0 &&
(c->replstate == REDIS_REPL_NONE || (c->replstate == REDIS_REPL_NONE ||
...@@ -88,9 +84,15 @@ robj *dupLastObjectIfNeeded(list *reply) { ...@@ -88,9 +84,15 @@ robj *dupLastObjectIfNeeded(list *reply) {
return listNodeValue(ln); return listNodeValue(ln);
} }
/* -----------------------------------------------------------------------------
* Low level functions to add more data to output buffers.
* -------------------------------------------------------------------------- */
int _addReplyToBuffer(redisClient *c, char *s, size_t len) { int _addReplyToBuffer(redisClient *c, char *s, size_t len) {
size_t available = sizeof(c->buf)-c->bufpos; size_t available = sizeof(c->buf)-c->bufpos;
if (c->flags & REDIS_CLOSE_AFTER_REPLY) return REDIS_OK;
/* If there already are entries in the reply list, we cannot /* If there already are entries in the reply list, we cannot
* add anything more to the static buffer. */ * add anything more to the static buffer. */
if (listLength(c->reply) > 0) return REDIS_ERR; if (listLength(c->reply) > 0) return REDIS_ERR;
...@@ -105,6 +107,9 @@ int _addReplyToBuffer(redisClient *c, char *s, size_t len) { ...@@ -105,6 +107,9 @@ int _addReplyToBuffer(redisClient *c, char *s, size_t len) {
void _addReplyObjectToList(redisClient *c, robj *o) { void _addReplyObjectToList(redisClient *c, robj *o) {
robj *tail; robj *tail;
if (c->flags & REDIS_CLOSE_AFTER_REPLY) return;
if (listLength(c->reply) == 0) { if (listLength(c->reply) == 0) {
incrRefCount(o); incrRefCount(o);
listAddNodeTail(c->reply,o); listAddNodeTail(c->reply,o);
...@@ -128,6 +133,12 @@ void _addReplyObjectToList(redisClient *c, robj *o) { ...@@ -128,6 +133,12 @@ void _addReplyObjectToList(redisClient *c, robj *o) {
* needed it will be free'd, otherwise it ends up in a robj. */ * needed it will be free'd, otherwise it ends up in a robj. */
void _addReplySdsToList(redisClient *c, sds s) { void _addReplySdsToList(redisClient *c, sds s) {
robj *tail; robj *tail;
if (c->flags & REDIS_CLOSE_AFTER_REPLY) {
sdsfree(s);
return;
}
if (listLength(c->reply) == 0) { if (listLength(c->reply) == 0) {
listAddNodeTail(c->reply,createObject(REDIS_STRING,s)); listAddNodeTail(c->reply,createObject(REDIS_STRING,s));
} else { } else {
...@@ -148,6 +159,9 @@ void _addReplySdsToList(redisClient *c, sds s) { ...@@ -148,6 +159,9 @@ void _addReplySdsToList(redisClient *c, sds s) {
void _addReplyStringToList(redisClient *c, char *s, size_t len) { void _addReplyStringToList(redisClient *c, char *s, size_t len) {
robj *tail; robj *tail;
if (c->flags & REDIS_CLOSE_AFTER_REPLY) return;
if (listLength(c->reply) == 0) { if (listLength(c->reply) == 0) {
listAddNodeTail(c->reply,createStringObject(s,len)); listAddNodeTail(c->reply,createStringObject(s,len));
} else { } else {
...@@ -165,6 +179,11 @@ void _addReplyStringToList(redisClient *c, char *s, size_t len) { ...@@ -165,6 +179,11 @@ void _addReplyStringToList(redisClient *c, char *s, size_t len) {
} }
} }
/* -----------------------------------------------------------------------------
* Higher level functions to queue data on the client output buffer.
* The following functions are the ones that commands implementations will call.
* -------------------------------------------------------------------------- */
void addReply(redisClient *c, robj *obj) { void addReply(redisClient *c, robj *obj) {
if (_installWriteEvent(c) != REDIS_OK) return; if (_installWriteEvent(c) != REDIS_OK) return;
...@@ -301,6 +320,11 @@ void _addReplyLongLong(redisClient *c, long long ll, char prefix) { ...@@ -301,6 +320,11 @@ void _addReplyLongLong(redisClient *c, long long ll, char prefix) {
} }
void addReplyLongLong(redisClient *c, long long ll) { void addReplyLongLong(redisClient *c, long long ll) {
if (ll == 0)
addReply(c,shared.czero);
else if (ll == 1)
addReply(c,shared.cone);
else
_addReplyLongLong(c,ll,':'); _addReplyLongLong(c,ll,':');
} }
...@@ -395,7 +419,7 @@ void acceptTcpHandler(aeEventLoop *el, int fd, void *privdata, int mask) { ...@@ -395,7 +419,7 @@ void acceptTcpHandler(aeEventLoop *el, int fd, void *privdata, int mask) {
cfd = anetTcpAccept(server.neterr, fd, cip, &cport); cfd = anetTcpAccept(server.neterr, fd, cip, &cport);
if (cfd == AE_ERR) { if (cfd == AE_ERR) {
redisLog(REDIS_VERBOSE,"Accepting client connection: %s", server.neterr); redisLog(REDIS_WARNING,"Accepting client connection: %s", server.neterr);
return; return;
} }
redisLog(REDIS_VERBOSE,"Accepted %s:%d", cip, cport); redisLog(REDIS_VERBOSE,"Accepted %s:%d", cip, cport);
...@@ -410,7 +434,7 @@ void acceptUnixHandler(aeEventLoop *el, int fd, void *privdata, int mask) { ...@@ -410,7 +434,7 @@ void acceptUnixHandler(aeEventLoop *el, int fd, void *privdata, int mask) {
cfd = anetUnixAccept(server.neterr, fd); cfd = anetUnixAccept(server.neterr, fd);
if (cfd == AE_ERR) { if (cfd == AE_ERR) {
redisLog(REDIS_VERBOSE,"Accepting client connection: %s", server.neterr); redisLog(REDIS_WARNING,"Accepting client connection: %s", server.neterr);
return; return;
} }
redisLog(REDIS_VERBOSE,"Accepted connection to %s", server.unixsocket); redisLog(REDIS_VERBOSE,"Accepted connection to %s", server.unixsocket);
...@@ -502,12 +526,18 @@ void freeClient(redisClient *c) { ...@@ -502,12 +526,18 @@ void freeClient(redisClient *c) {
* close the connection with all our slaves if we have any, so * close the connection with all our slaves if we have any, so
* when we'll resync with the master the other slaves will sync again * when we'll resync with the master the other slaves will sync again
* with us as well. Note that also when the slave is not connected * with us as well. Note that also when the slave is not connected
* to the master it will keep refusing connections by other slaves. */ * to the master it will keep refusing connections by other slaves.
*
* We do this only if server.masterhost != NULL. If it is NULL this
* means the user called SLAVEOF NO ONE and we are freeing our
* link with the master, so no need to close link with slaves. */
if (server.masterhost != NULL) {
while (listLength(server.slaves)) { while (listLength(server.slaves)) {
ln = listFirst(server.slaves); ln = listFirst(server.slaves);
freeClient((redisClient*)ln->value); freeClient((redisClient*)ln->value);
} }
} }
}
/* Release memory */ /* Release memory */
zfree(c->argv); zfree(c->argv);
freeClientMultiState(c); freeClientMultiState(c);
...@@ -670,47 +700,57 @@ static void setProtocolError(redisClient *c, int pos) { ...@@ -670,47 +700,57 @@ static void setProtocolError(redisClient *c, int pos) {
int processMultibulkBuffer(redisClient *c) { int processMultibulkBuffer(redisClient *c) {
char *newline = NULL; char *newline = NULL;
char *eptr; int pos = 0, ok;
int pos = 0, tolerr; long long ll;
long bulklen;
if (c->multibulklen == 0) { if (c->multibulklen == 0) {
/* The client should have been reset */ /* The client should have been reset */
redisAssert(c->argc == 0); redisAssert(c->argc == 0);
/* Multi bulk length cannot be read without a \r\n */ /* Multi bulk length cannot be read without a \r\n */
newline = strstr(c->querybuf,"\r\n"); newline = strchr(c->querybuf,'\r');
if (newline == NULL) if (newline == NULL)
return REDIS_ERR; return REDIS_ERR;
/* Buffer should also contain \n */
if (newline-(c->querybuf) > ((signed)sdslen(c->querybuf)-2))
return REDIS_ERR;
/* We know for sure there is a whole line since newline != NULL, /* We know for sure there is a whole line since newline != NULL,
* so go ahead and find out the multi bulk length. */ * so go ahead and find out the multi bulk length. */
redisAssert(c->querybuf[0] == '*'); redisAssert(c->querybuf[0] == '*');
c->multibulklen = strtol(c->querybuf+1,&eptr,10); ok = string2ll(c->querybuf+1,newline-(c->querybuf+1),&ll);
pos = (newline-c->querybuf)+2; if (!ok || ll > 1024*1024) {
if (c->multibulklen <= 0) {
c->querybuf = sdsrange(c->querybuf,pos,-1);
return REDIS_OK;
} else if (c->multibulklen > 1024*1024) {
addReplyError(c,"Protocol error: invalid multibulk length"); addReplyError(c,"Protocol error: invalid multibulk length");
setProtocolError(c,pos); setProtocolError(c,pos);
return REDIS_ERR; return REDIS_ERR;
} }
pos = (newline-c->querybuf)+2;
if (ll <= 0) {
c->querybuf = sdsrange(c->querybuf,pos,-1);
return REDIS_OK;
}
c->multibulklen = ll;
/* Setup argv array on client structure */ /* Setup argv array on client structure */
if (c->argv) zfree(c->argv); if (c->argv) zfree(c->argv);
c->argv = zmalloc(sizeof(robj*)*c->multibulklen); c->argv = zmalloc(sizeof(robj*)*c->multibulklen);
/* Search new newline */
newline = strstr(c->querybuf+pos,"\r\n");
} }
redisAssert(c->multibulklen > 0); redisAssert(c->multibulklen > 0);
while(c->multibulklen) { while(c->multibulklen) {
/* Read bulk length if unknown */ /* Read bulk length if unknown */
if (c->bulklen == -1) { if (c->bulklen == -1) {
newline = strstr(c->querybuf+pos,"\r\n"); newline = strchr(c->querybuf+pos,'\r');
if (newline != NULL) { if (newline == NULL)
break;
/* Buffer should also contain \n */
if (newline-(c->querybuf) > ((signed)sdslen(c->querybuf)-2))
break;
if (c->querybuf[pos] != '$') { if (c->querybuf[pos] != '$') {
addReplyErrorFormat(c, addReplyErrorFormat(c,
"Protocol error: expected '$', got '%c'", "Protocol error: expected '$', got '%c'",
...@@ -719,21 +759,15 @@ int processMultibulkBuffer(redisClient *c) { ...@@ -719,21 +759,15 @@ int processMultibulkBuffer(redisClient *c) {
return REDIS_ERR; return REDIS_ERR;
} }
bulklen = strtol(c->querybuf+pos+1,&eptr,10); ok = string2ll(c->querybuf+pos+1,newline-(c->querybuf+pos+1),&ll);
tolerr = (eptr[0] != '\r'); if (!ok || ll < 0 || ll > 512*1024*1024) {
if (tolerr || bulklen == LONG_MIN || bulklen == LONG_MAX ||
bulklen < 0 || bulklen > 512*1024*1024)
{
addReplyError(c,"Protocol error: invalid bulk length"); addReplyError(c,"Protocol error: invalid bulk length");
setProtocolError(c,pos); setProtocolError(c,pos);
return REDIS_ERR; return REDIS_ERR;
} }
pos += eptr-(c->querybuf+pos)+2;
c->bulklen = bulklen; pos += newline-(c->querybuf+pos)+2;
} else { c->bulklen = ll;
/* No newline in current buffer, so wait for more data */
break;
}
} }
/* Read bulk argument */ /* Read bulk argument */
...@@ -845,3 +879,70 @@ void getClientsMaxBuffers(unsigned long *longest_output_list, ...@@ -845,3 +879,70 @@ void getClientsMaxBuffers(unsigned long *longest_output_list,
*biggest_input_buffer = bib; *biggest_input_buffer = bib;
} }
void clientCommand(redisClient *c) {
listNode *ln;
listIter li;
redisClient *client;
if (!strcasecmp(c->argv[1]->ptr,"list") && c->argc == 2) {
sds o = sdsempty();
time_t now = time(NULL);
listRewind(server.clients,&li);
while ((ln = listNext(&li)) != NULL) {
char ip[32], flags[16], *p;
int port;
client = listNodeValue(ln);
if (anetPeerToString(client->fd,ip,&port) == -1) continue;
p = flags;
if (client->flags & REDIS_SLAVE) {
if (client->flags & REDIS_MONITOR)
*p++ = 'O';
else
*p++ = 'S';
}
if (client->flags & REDIS_MASTER) *p++ = 'M';
if (p == flags) *p++ = 'N';
if (client->flags & REDIS_MULTI) *p++ = 'x';
if (client->flags & REDIS_BLOCKED) *p++ = 'b';
if (client->flags & REDIS_IO_WAIT) *p++ = 'i';
if (client->flags & REDIS_DIRTY_CAS) *p++ = 'd';
if (client->flags & REDIS_CLOSE_AFTER_REPLY) *p++ = 'c';
if (client->flags & REDIS_UNBLOCKED) *p++ = 'u';
*p++ = '\0';
o = sdscatprintf(o,
"addr=%s:%d fd=%d idle=%ld flags=%s db=%d sub=%d psub=%d\n",
ip,port,client->fd,
(long)(now - client->lastinteraction),
flags,
client->db->id,
(int) dictSize(client->pubsub_channels),
(int) listLength(client->pubsub_patterns));
}
addReplyBulkCBuffer(c,o,sdslen(o));
sdsfree(o);
} else if (!strcasecmp(c->argv[1]->ptr,"kill") && c->argc == 3) {
listRewind(server.clients,&li);
while ((ln = listNext(&li)) != NULL) {
char ip[32], addr[64];
int port;
client = listNodeValue(ln);
if (anetPeerToString(client->fd,ip,&port) == -1) continue;
snprintf(addr,sizeof(addr),"%s:%d",ip,port);
if (strcmp(addr,c->argv[2]->ptr) == 0) {
addReply(c,shared.ok);
if (c == client) {
client->flags |= REDIS_CLOSE_AFTER_REPLY;
} else {
freeClient(client);
}
return;
}
}
addReplyError(c,"No such client");
} else {
addReplyError(c, "Syntax error, try CLIENT (LIST | KILL ip:port)");
}
}
...@@ -93,10 +93,20 @@ robj *createHashObject(void) { ...@@ -93,10 +93,20 @@ robj *createHashObject(void) {
robj *createZsetObject(void) { robj *createZsetObject(void) {
zset *zs = zmalloc(sizeof(*zs)); zset *zs = zmalloc(sizeof(*zs));
robj *o;
zs->dict = dictCreate(&zsetDictType,NULL); zs->dict = dictCreate(&zsetDictType,NULL);
zs->zsl = zslCreate(); zs->zsl = zslCreate();
return createObject(REDIS_ZSET,zs); o = createObject(REDIS_ZSET,zs);
o->encoding = REDIS_ENCODING_SKIPLIST;
return o;
}
robj *createZsetZiplistObject(void) {
unsigned char *zl = ziplistNew();
robj *o = createObject(REDIS_ZSET,zl);
o->encoding = REDIS_ENCODING_ZIPLIST;
return o;
} }
void freeStringObject(robj *o) { void freeStringObject(robj *o) {
...@@ -132,11 +142,20 @@ void freeSetObject(robj *o) { ...@@ -132,11 +142,20 @@ void freeSetObject(robj *o) {
} }
void freeZsetObject(robj *o) { void freeZsetObject(robj *o) {
zset *zs = o->ptr; zset *zs;
switch (o->encoding) {
case REDIS_ENCODING_SKIPLIST:
zs = o->ptr;
dictRelease(zs->dict); dictRelease(zs->dict);
zslFree(zs->zsl); zslFree(zs->zsl);
zfree(zs); zfree(zs);
break;
case REDIS_ENCODING_ZIPLIST:
zfree(o->ptr);
break;
default:
redisPanic("Unknown sorted set encoding");
}
} }
void freeHashObject(robj *o) { void freeHashObject(robj *o) {
...@@ -183,6 +202,16 @@ int checkType(redisClient *c, robj *o, int type) { ...@@ -183,6 +202,16 @@ int checkType(redisClient *c, robj *o, int type) {
return 0; return 0;
} }
int isObjectRepresentableAsLongLong(robj *o, long long *llval) {
redisAssert(o->type == REDIS_STRING);
if (o->encoding == REDIS_ENCODING_INT) {
if (llval) *llval = (long) o->ptr;
return REDIS_OK;
} else {
return string2ll(o->ptr,sdslen(o->ptr),llval) ? REDIS_OK : REDIS_ERR;
}
}
/* Try to encode a string object in order to save space */ /* Try to encode a string object in order to save space */
robj *tryObjectEncoding(robj *o) { robj *tryObjectEncoding(robj *o) {
long value; long value;
...@@ -200,7 +229,7 @@ robj *tryObjectEncoding(robj *o) { ...@@ -200,7 +229,7 @@ robj *tryObjectEncoding(robj *o) {
redisAssert(o->type == REDIS_STRING); redisAssert(o->type == REDIS_STRING);
/* Check if we can represent this string as a long integer */ /* Check if we can represent this string as a long integer */
if (isStringRepresentableAsLong(s,&value) == REDIS_ERR) return o; if (!string2l(s,sdslen(s),&value)) return o;
/* Ok, this object can be encoded... /* Ok, this object can be encoded...
* *
...@@ -402,6 +431,7 @@ char *strEncoding(int encoding) { ...@@ -402,6 +431,7 @@ char *strEncoding(int encoding) {
case REDIS_ENCODING_LINKEDLIST: return "linkedlist"; case REDIS_ENCODING_LINKEDLIST: return "linkedlist";
case REDIS_ENCODING_ZIPLIST: return "ziplist"; case REDIS_ENCODING_ZIPLIST: return "ziplist";
case REDIS_ENCODING_INTSET: return "intset"; case REDIS_ENCODING_INTSET: return "intset";
case REDIS_ENCODING_SKIPLIST: return "skiplist";
default: return "unknown"; default: return "unknown";
} }
} }
...@@ -416,3 +446,42 @@ unsigned long estimateObjectIdleTime(robj *o) { ...@@ -416,3 +446,42 @@ unsigned long estimateObjectIdleTime(robj *o) {
REDIS_LRU_CLOCK_RESOLUTION; REDIS_LRU_CLOCK_RESOLUTION;
} }
} }
/* This is an helper function for the DEBUG command. We need to lookup keys
* without any modification of LRU or other parameters. */
robj *objectCommandLookup(redisClient *c, robj *key) {
dictEntry *de;
if ((de = dictFind(c->db->dict,key->ptr)) == NULL) return NULL;
return (robj*) dictGetEntryVal(de);
}
robj *objectCommandLookupOrReply(redisClient *c, robj *key, robj *reply) {
robj *o = objectCommandLookup(c,key);
if (!o) addReply(c, reply);
return o;
}
/* Object command allows to inspect the internals of an Redis Object.
* Usage: OBJECT <verb> ... arguments ... */
void objectCommand(redisClient *c) {
robj *o;
if (!strcasecmp(c->argv[1]->ptr,"refcount") && c->argc == 3) {
if ((o = objectCommandLookupOrReply(c,c->argv[2],shared.nullbulk))
== NULL) return;
addReplyLongLong(c,o->refcount);
} else if (!strcasecmp(c->argv[1]->ptr,"encoding") && c->argc == 3) {
if ((o = objectCommandLookupOrReply(c,c->argv[2],shared.nullbulk))
== NULL) return;
addReplyBulkCString(c,strEncoding(o->encoding));
} else if (!strcasecmp(c->argv[1]->ptr,"idletime") && c->argc == 3) {
if ((o = objectCommandLookupOrReply(c,c->argv[2],shared.nullbulk))
== NULL) return;
addReplyLongLong(c,estimateObjectIdleTime(o));
} else {
addReplyError(c,"Syntax error. Try OBJECT (refcount|encoding|idletime)");
}
}
...@@ -245,7 +245,7 @@ int rdbSaveDoubleValue(FILE *fp, double val) { ...@@ -245,7 +245,7 @@ int rdbSaveDoubleValue(FILE *fp, double val) {
return rdbWriteRaw(fp,buf,len); 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 rdbSaveObject(FILE *fp, robj *o) {
int n, nwritten = 0; int n, nwritten = 0;
...@@ -302,7 +302,13 @@ int rdbSaveObject(FILE *fp, robj *o) { ...@@ -302,7 +302,13 @@ int rdbSaveObject(FILE *fp, robj *o) {
redisPanic("Unknown set encoding"); redisPanic("Unknown set encoding");
} }
} else if (o->type == REDIS_ZSET) { } else if (o->type == REDIS_ZSET) {
/* Save a set value */ /* Save a sorted set value */
if (o->encoding == REDIS_ENCODING_ZIPLIST) {
size_t l = ziplistBlobLen((unsigned char*)o->ptr);
if ((n = rdbSaveRawString(fp,o->ptr,l)) == -1) return -1;
nwritten += n;
} else if (o->encoding == REDIS_ENCODING_SKIPLIST) {
zset *zs = o->ptr; zset *zs = o->ptr;
dictIterator *di = dictGetIterator(zs->dict); dictIterator *di = dictGetIterator(zs->dict);
dictEntry *de; dictEntry *de;
...@@ -320,6 +326,9 @@ int rdbSaveObject(FILE *fp, robj *o) { ...@@ -320,6 +326,9 @@ int rdbSaveObject(FILE *fp, robj *o) {
nwritten += n; nwritten += n;
} }
dictReleaseIterator(di); dictReleaseIterator(di);
} else {
redisPanic("Unknown sorted set encoding");
}
} else if (o->type == REDIS_HASH) { } else if (o->type == REDIS_HASH) {
/* Save a hash value */ /* Save a hash value */
if (o->encoding == REDIS_ENCODING_ZIPMAP) { if (o->encoding == REDIS_ENCODING_ZIPMAP) {
...@@ -386,6 +395,8 @@ int rdbSaveKeyValuePair(FILE *fp, robj *key, robj *val, ...@@ -386,6 +395,8 @@ int rdbSaveKeyValuePair(FILE *fp, robj *key, robj *val,
vtype = REDIS_LIST_ZIPLIST; vtype = REDIS_LIST_ZIPLIST;
else if (vtype == REDIS_SET && val->encoding == REDIS_ENCODING_INTSET) else if (vtype == REDIS_SET && val->encoding == REDIS_ENCODING_INTSET)
vtype = REDIS_SET_INTSET; vtype = REDIS_SET_INTSET;
else if (vtype == REDIS_ZSET && val->encoding == REDIS_ENCODING_ZIPLIST)
vtype = REDIS_ZSET_ZIPLIST;
/* Save type, key, value */ /* Save type, key, value */
if (rdbSaveType(fp,vtype) == -1) return -1; if (rdbSaveType(fp,vtype) == -1) return -1;
if (rdbSaveStringObject(fp,key) == -1) return -1; if (rdbSaveStringObject(fp,key) == -1) return -1;
...@@ -414,7 +425,7 @@ int rdbSave(char *filename) { ...@@ -414,7 +425,7 @@ int rdbSave(char *filename) {
strerror(errno)); strerror(errno));
return REDIS_ERR; return REDIS_ERR;
} }
if (fwrite("REDIS0001",9,1,fp) == 0) goto werr; if (fwrite("REDIS0002",9,1,fp) == 0) goto werr;
for (j = 0; j < server.dbnum; j++) { for (j = 0; j < server.dbnum; j++) {
redisDb *db = server.db+j; redisDb *db = server.db+j;
dict *d = db->dict; dict *d = db->dict;
...@@ -745,11 +756,13 @@ robj *rdbLoadObject(int type, FILE *fp) { ...@@ -745,11 +756,13 @@ robj *rdbLoadObject(int type, FILE *fp) {
} else if (type == REDIS_ZSET) { } else if (type == REDIS_ZSET) {
/* Read list/set value */ /* Read list/set value */
size_t zsetlen; size_t zsetlen;
size_t maxelelen = 0;
zset *zs; zset *zs;
if ((zsetlen = rdbLoadLen(fp,NULL)) == REDIS_RDB_LENERR) return NULL; if ((zsetlen = rdbLoadLen(fp,NULL)) == REDIS_RDB_LENERR) return NULL;
o = createZsetObject(); o = createZsetObject();
zs = o->ptr; zs = o->ptr;
/* Load every single element of the list/set */ /* Load every single element of the list/set */
while(zsetlen--) { while(zsetlen--) {
robj *ele; robj *ele;
...@@ -759,10 +772,21 @@ robj *rdbLoadObject(int type, FILE *fp) { ...@@ -759,10 +772,21 @@ robj *rdbLoadObject(int type, FILE *fp) {
if ((ele = rdbLoadEncodedStringObject(fp)) == NULL) return NULL; if ((ele = rdbLoadEncodedStringObject(fp)) == NULL) return NULL;
ele = tryObjectEncoding(ele); ele = tryObjectEncoding(ele);
if (rdbLoadDoubleValue(fp,&score) == -1) return NULL; if (rdbLoadDoubleValue(fp,&score) == -1) return NULL;
/* Don't care about integer-encoded strings. */
if (ele->encoding == REDIS_ENCODING_RAW &&
sdslen(ele->ptr) > maxelelen)
maxelelen = sdslen(ele->ptr);
znode = zslInsert(zs->zsl,score,ele); znode = zslInsert(zs->zsl,score,ele);
dictAdd(zs->dict,ele,&znode->score); dictAdd(zs->dict,ele,&znode->score);
incrRefCount(ele); /* added to skiplist */ incrRefCount(ele); /* added to skiplist */
} }
/* Convert *after* loading, since sorted sets are not stored ordered. */
if (zsetLength(o) <= server.zset_max_ziplist_entries &&
maxelelen <= server.zset_max_ziplist_value)
zsetConvert(o,REDIS_ENCODING_ZIPLIST);
} else if (type == REDIS_HASH) { } else if (type == REDIS_HASH) {
size_t hashlen; size_t hashlen;
...@@ -811,7 +835,8 @@ robj *rdbLoadObject(int type, FILE *fp) { ...@@ -811,7 +835,8 @@ robj *rdbLoadObject(int type, FILE *fp) {
} }
} else if (type == REDIS_HASH_ZIPMAP || } else if (type == REDIS_HASH_ZIPMAP ||
type == REDIS_LIST_ZIPLIST || type == REDIS_LIST_ZIPLIST ||
type == REDIS_SET_INTSET) type == REDIS_SET_INTSET ||
type == REDIS_ZSET_ZIPLIST)
{ {
robj *aux = rdbLoadStringObject(fp); robj *aux = rdbLoadStringObject(fp);
...@@ -846,8 +871,14 @@ robj *rdbLoadObject(int type, FILE *fp) { ...@@ -846,8 +871,14 @@ robj *rdbLoadObject(int type, FILE *fp) {
if (intsetLen(o->ptr) > server.set_max_intset_entries) if (intsetLen(o->ptr) > server.set_max_intset_entries)
setTypeConvert(o,REDIS_ENCODING_HT); setTypeConvert(o,REDIS_ENCODING_HT);
break; break;
case REDIS_ZSET_ZIPLIST:
o->type = REDIS_ZSET;
o->encoding = REDIS_ENCODING_ZIPLIST;
if (zsetLength(o) > server.zset_max_ziplist_entries)
zsetConvert(o,REDIS_ENCODING_SKIPLIST);
break;
default: default:
redisPanic("Unknown enoding"); redisPanic("Unknown encoding");
break; break;
} }
} else { } else {
...@@ -900,7 +931,7 @@ int rdbLoad(char *filename) { ...@@ -900,7 +931,7 @@ int rdbLoad(char *filename) {
return REDIS_ERR; return REDIS_ERR;
} }
rdbver = atoi(buf+5); rdbver = atoi(buf+5);
if (rdbver != 1) { if (rdbver < 1 || rdbver > 2) {
fclose(fp); fclose(fp);
redisLog(REDIS_WARNING,"Can't handle RDB format version %d",rdbver); redisLog(REDIS_WARNING,"Can't handle RDB format version %d",rdbver);
return REDIS_ERR; return REDIS_ERR;
......
...@@ -396,6 +396,9 @@ static sds cliFormatReplyRaw(redisReply *r) { ...@@ -396,6 +396,9 @@ static sds cliFormatReplyRaw(redisReply *r) {
/* Nothing... */ /* Nothing... */
break; break;
case REDIS_REPLY_ERROR: case REDIS_REPLY_ERROR:
out = sdscatlen(out,r->str,r->len);
out = sdscatlen(out,"\n",1);
break;
case REDIS_REPLY_STATUS: case REDIS_REPLY_STATUS:
case REDIS_REPLY_STRING: case REDIS_REPLY_STRING:
out = sdscatlen(out,r->str,r->len); out = sdscatlen(out,r->str,r->len);
...@@ -464,7 +467,18 @@ static int cliSendCommand(int argc, char **argv, int repeat) { ...@@ -464,7 +467,18 @@ static int cliSendCommand(int argc, char **argv, int repeat) {
return REDIS_OK; 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"))) ||
(argc == 2 && !strcasecmp(command,"client") &&
!strcasecmp(argv[1],"list")))
{
output_raw = 1;
}
if (!strcasecmp(command,"help") || !strcasecmp(command,"?")) { if (!strcasecmp(command,"help") || !strcasecmp(command,"?")) {
cliOutputHelp(--argc, ++argv); cliOutputHelp(--argc, ++argv);
return REDIS_OK; return REDIS_OK;
......
#!/usr/bin/env ruby
require 'rubygems'
require 'redis'
ClusterHashSlots = 4096
def xputs(s)
printf s
STDOUT.flush
end
class ClusterNode
def initialize(addr)
s = addr.split(":")
if s.length != 2
puts "Invalid node name #{addr}"
exit 1
end
@r = nil
@host = s[0]
@port = s[1]
@slots = {}
@dirty = false
end
def to_s
"#{@host}:#{@port}"
end
def connect(o={})
xputs "Connecting to node #{self}: "
begin
@r = Redis.new(:host => @host, :port => @port)
@r.ping
rescue
puts "ERROR"
puts "Sorry, can't connect to node #{self}"
exit 1 if o[:abort]
@r = nil
end
puts "OK"
end
def assert_cluster
info = @r.info
if !info["cluster_enabled"] || info["cluster_enabled"].to_i == 0
puts "Error: Node #{self} is not configured as a cluster node."
exit 1
end
end
def assert_empty
if !(@r.cluster("info").split("\r\n").index("cluster_known_nodes:1")) ||
(@r.info['db0'])
puts "Error: Node #{self} is not empty. Either the node already knows other nodes (check with nodes-info) or contains some key in database 0."
exit 1
end
end
def add_slots(slots)
slots.each{|s|
@slots[s] = :new
}
@dirty = true
end
def flush_node_config
return if !@dirty
new = []
@slots.each{|s,val|
if val == :new
new << s
@slots[s] = true
end
}
@r.cluster("addslots",*new)
@dirty = false
end
def info_string
# We want to display the hash slots assigned to this node
# as ranges, like in: "1-5,8-9,20-25,30"
#
# Note: this could be easily written without side effects,
# we use 'slots' just to split the computation into steps.
# First step: we want an increasing array of integers
# for instance: [1,2,3,4,5,8,9,20,21,22,23,24,25,30]
slots = @slots.keys.sort
# As we want to aggregate adiacent slots we convert all the
# slot integers into ranges (with just one element)
# So we have something like [1..1,2..2, ... and so forth.
slots.map!{|x| x..x}
# Finally we group ranges with adiacent elements.
slots = slots.reduce([]) {|a,b|
if !a.empty? && b.first == (a[-1].last)+1
a[0..-2] + [(a[-1].first)..(b.last)]
else
a + [b]
end
}
# Now our task is easy, we just convert ranges with just one
# element into a number, and a real range into a start-end format.
# Finally we join the array using the comma as separator.
slots = slots.map{|x|
x.count == 1 ? x.first.to_s : "#{x.first}-#{x.last}"
}.join(",")
"#{self.to_s.ljust(25)} slots:#{slots}"
end
def info
{
:host => @host,
:port => @port,
:slots => @slots,
:dirty => @dirty
}
end
def is_dirty?
@dirty
end
def r
@r
end
end
class RedisTrib
def initialize
@nodes = []
end
def check_arity(req_args, num_args)
if ((req_args > 0 and num_args != req_args) ||
(req_args < 0 and num_args < req_args.abs))
puts "Wrong number of arguments for specified sub command"
exit 1
end
end
def add_node(node)
@nodes << node
end
def check_cluster
puts "Performing Cluster Check (using node #{@nodes[0]})"
show_nodes
end
def alloc_slots
slots_per_node = ClusterHashSlots/@nodes.length
i = 0
@nodes.each{|n|
first = i*slots_per_node
last = first+slots_per_node-1
last = ClusterHashSlots-1 if i == @nodes.length-1
n.add_slots first..last
i += 1
}
end
def flush_nodes_config
@nodes.each{|n|
n.flush_node_config
}
end
def show_nodes
@nodes.each{|n|
puts n.info_string
}
end
def join_cluster
# We use a brute force approach to make sure the node will meet
# each other, that is, sending CLUSTER MEET messages to all the nodes
# about the very same node.
# Thanks to gossip this information should propagate across all the
# cluster in a matter of seconds.
first = false
@nodes.each{|n|
if !first then first = n.info; next; end # Skip the first node
n.r.cluster("meet",first[:host],first[:port])
}
end
def yes_or_die(msg)
print "#{msg} (type 'yes' to accept): "
STDOUT.flush
if !(STDIN.gets.chomp.downcase == "yes")
puts "Aborting..."
exit 1
end
end
# redis-trib subcommands implementations
def check_cluster_cmd
node = ClusterNode.new(ARGV[1])
node.connect(:abort => true)
node.assert_cluster
add_node(node)
check_cluster
end
def create_cluster_cmd
puts "Creating cluster"
ARGV[1..-1].each{|n|
node = ClusterNode.new(n)
node.connect(:abort => true)
node.assert_cluster
node.assert_empty
add_node(node)
}
puts "Performing hash slots allocation on #{@nodes.length} nodes..."
alloc_slots
show_nodes
yes_or_die "Can I set the above configuration?"
flush_nodes_config
puts "** Nodes configuration updated"
puts "** Sending CLUSTER MEET messages to join the cluster"
join_cluster
check_cluster
end
end
COMMANDS={
"create" => ["create_cluster_cmd", -2, "host1:port host2:port ... hostN:port"],
"check" => ["check_cluster_cmd", 2, "host:port"]
}
# Sanity check
if ARGV.length == 0
puts "Usage: redis-trib <command> <arguments ...>"
puts
COMMANDS.each{|k,v|
puts " #{k.ljust(20)} #{v[2]}"
}
puts
exit 1
end
rt = RedisTrib.new
cmd_spec = COMMANDS[ARGV[0].downcase]
if !cmd_spec
puts "Unknown redis-trib subcommand '#{ARGV[0]}'"
exit 1
end
rt.check_arity(cmd_spec[1],ARGV.length)
# Dispatch
rt.send(cmd_spec[0])
...@@ -70,12 +70,12 @@ struct redisServer server; /* server global state */ ...@@ -70,12 +70,12 @@ struct redisServer server; /* server global state */
struct redisCommand *commandTable; struct redisCommand *commandTable;
struct redisCommand redisCommandTable[] = { struct redisCommand redisCommandTable[] = {
{"get",getCommand,2,0,NULL,1,1,1,0,0}, {"get",getCommand,2,0,NULL,1,1,1,0,0},
{"set",setCommand,3,REDIS_CMD_DENYOOM,NULL,0,0,0,0,0}, {"set",setCommand,3,REDIS_CMD_DENYOOM,noPreloadGetKeys,1,1,1,0,0},
{"setnx",setnxCommand,3,REDIS_CMD_DENYOOM,NULL,0,0,0,0,0}, {"setnx",setnxCommand,3,REDIS_CMD_DENYOOM,noPreloadGetKeys,1,1,1,0,0},
{"setex",setexCommand,4,REDIS_CMD_DENYOOM,NULL,0,0,0,0,0}, {"setex",setexCommand,4,REDIS_CMD_DENYOOM,noPreloadGetKeys,2,2,1,0,0},
{"append",appendCommand,3,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0}, {"append",appendCommand,3,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0},
{"strlen",strlenCommand,2,0,NULL,1,1,1,0,0}, {"strlen",strlenCommand,2,0,NULL,1,1,1,0,0},
{"del",delCommand,-2,0,NULL,0,0,0,0,0}, {"del",delCommand,-2,0,noPreloadGetKeys,1,-1,1,0,0},
{"exists",existsCommand,2,0,NULL,1,1,1,0,0}, {"exists",existsCommand,2,0,NULL,1,1,1,0,0},
{"setbit",setbitCommand,4,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0}, {"setbit",setbitCommand,4,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0},
{"getbit",getbitCommand,3,0,NULL,1,1,1,0,0}, {"getbit",getbitCommand,3,0,NULL,1,1,1,0,0},
...@@ -85,8 +85,8 @@ struct redisCommand redisCommandTable[] = { ...@@ -85,8 +85,8 @@ struct redisCommand redisCommandTable[] = {
{"incr",incrCommand,2,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0}, {"incr",incrCommand,2,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0},
{"decr",decrCommand,2,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0}, {"decr",decrCommand,2,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0},
{"mget",mgetCommand,-2,0,NULL,1,-1,1,0,0}, {"mget",mgetCommand,-2,0,NULL,1,-1,1,0,0},
{"rpush",rpushCommand,3,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0}, {"rpush",rpushCommand,-3,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0},
{"lpush",lpushCommand,3,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0}, {"lpush",lpushCommand,-3,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0},
{"rpushx",rpushxCommand,3,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0}, {"rpushx",rpushxCommand,3,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0},
{"lpushx",lpushxCommand,3,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0}, {"lpushx",lpushxCommand,3,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0},
{"linsert",linsertCommand,5,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0}, {"linsert",linsertCommand,5,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0},
...@@ -94,7 +94,7 @@ struct redisCommand redisCommandTable[] = { ...@@ -94,7 +94,7 @@ struct redisCommand redisCommandTable[] = {
{"lpop",lpopCommand,2,0,NULL,1,1,1,0,0}, {"lpop",lpopCommand,2,0,NULL,1,1,1,0,0},
{"brpop",brpopCommand,-3,0,NULL,1,1,1,0,0}, {"brpop",brpopCommand,-3,0,NULL,1,1,1,0,0},
{"brpoplpush",brpoplpushCommand,4,REDIS_CMD_DENYOOM,NULL,1,2,1,0,0}, {"brpoplpush",brpoplpushCommand,4,REDIS_CMD_DENYOOM,NULL,1,2,1,0,0},
{"blpop",blpopCommand,-3,0,NULL,1,1,1,0,0}, {"blpop",blpopCommand,-3,0,NULL,1,-2,1,0,0},
{"llen",llenCommand,2,0,NULL,1,1,1,0,0}, {"llen",llenCommand,2,0,NULL,1,1,1,0,0},
{"lindex",lindexCommand,3,0,NULL,1,1,1,0,0}, {"lindex",lindexCommand,3,0,NULL,1,1,1,0,0},
{"lset",lsetCommand,4,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0}, {"lset",lsetCommand,4,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0},
...@@ -102,8 +102,8 @@ struct redisCommand redisCommandTable[] = { ...@@ -102,8 +102,8 @@ struct redisCommand redisCommandTable[] = {
{"ltrim",ltrimCommand,4,0,NULL,1,1,1,0,0}, {"ltrim",ltrimCommand,4,0,NULL,1,1,1,0,0},
{"lrem",lremCommand,4,0,NULL,1,1,1,0,0}, {"lrem",lremCommand,4,0,NULL,1,1,1,0,0},
{"rpoplpush",rpoplpushCommand,3,REDIS_CMD_DENYOOM,NULL,1,2,1,0,0}, {"rpoplpush",rpoplpushCommand,3,REDIS_CMD_DENYOOM,NULL,1,2,1,0,0},
{"sadd",saddCommand,3,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0}, {"sadd",saddCommand,-3,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0},
{"srem",sremCommand,3,0,NULL,1,1,1,0,0}, {"srem",sremCommand,-3,0,NULL,1,1,1,0,0},
{"smove",smoveCommand,4,0,NULL,1,2,1,0,0}, {"smove",smoveCommand,4,0,NULL,1,2,1,0,0},
{"sismember",sismemberCommand,3,0,NULL,1,1,1,0,0}, {"sismember",sismemberCommand,3,0,NULL,1,1,1,0,0},
{"scard",scardCommand,2,0,NULL,1,1,1,0,0}, {"scard",scardCommand,2,0,NULL,1,1,1,0,0},
...@@ -121,8 +121,8 @@ struct redisCommand redisCommandTable[] = { ...@@ -121,8 +121,8 @@ struct redisCommand redisCommandTable[] = {
{"zrem",zremCommand,3,0,NULL,1,1,1,0,0}, {"zrem",zremCommand,3,0,NULL,1,1,1,0,0},
{"zremrangebyscore",zremrangebyscoreCommand,4,0,NULL,1,1,1,0,0}, {"zremrangebyscore",zremrangebyscoreCommand,4,0,NULL,1,1,1,0,0},
{"zremrangebyrank",zremrangebyrankCommand,4,0,NULL,1,1,1,0,0}, {"zremrangebyrank",zremrangebyrankCommand,4,0,NULL,1,1,1,0,0},
{"zunionstore",zunionstoreCommand,-4,REDIS_CMD_DENYOOM,zunionInterBlockClientOnSwappedKeys,0,0,0,0,0}, {"zunionstore",zunionstoreCommand,-4,REDIS_CMD_DENYOOM,zunionInterGetKeys,0,0,0,0,0},
{"zinterstore",zinterstoreCommand,-4,REDIS_CMD_DENYOOM,zunionInterBlockClientOnSwappedKeys,0,0,0,0,0}, {"zinterstore",zinterstoreCommand,-4,REDIS_CMD_DENYOOM,zunionInterGetKeys,0,0,0,0,0},
{"zrange",zrangeCommand,-4,0,NULL,1,1,1,0,0}, {"zrange",zrangeCommand,-4,0,NULL,1,1,1,0,0},
{"zrangebyscore",zrangebyscoreCommand,-4,0,NULL,1,1,1,0,0}, {"zrangebyscore",zrangebyscoreCommand,-4,0,NULL,1,1,1,0,0},
{"zrevrangebyscore",zrevrangebyscoreCommand,-4,0,NULL,1,1,1,0,0}, {"zrevrangebyscore",zrevrangebyscoreCommand,-4,0,NULL,1,1,1,0,0},
...@@ -138,7 +138,7 @@ struct redisCommand redisCommandTable[] = { ...@@ -138,7 +138,7 @@ struct redisCommand redisCommandTable[] = {
{"hmset",hmsetCommand,-4,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0}, {"hmset",hmsetCommand,-4,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0},
{"hmget",hmgetCommand,-3,0,NULL,1,1,1,0,0}, {"hmget",hmgetCommand,-3,0,NULL,1,1,1,0,0},
{"hincrby",hincrbyCommand,4,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0}, {"hincrby",hincrbyCommand,4,REDIS_CMD_DENYOOM,NULL,1,1,1,0,0},
{"hdel",hdelCommand,3,0,NULL,1,1,1,0,0}, {"hdel",hdelCommand,-3,0,NULL,1,1,1,0,0},
{"hlen",hlenCommand,2,0,NULL,1,1,1,0,0}, {"hlen",hlenCommand,2,0,NULL,1,1,1,0,0},
{"hkeys",hkeysCommand,2,0,NULL,1,1,1,0,0}, {"hkeys",hkeysCommand,2,0,NULL,1,1,1,0,0},
{"hvals",hvalsCommand,2,0,NULL,1,1,1,0,0}, {"hvals",hvalsCommand,2,0,NULL,1,1,1,0,0},
...@@ -152,10 +152,10 @@ struct redisCommand redisCommandTable[] = { ...@@ -152,10 +152,10 @@ struct redisCommand redisCommandTable[] = {
{"randomkey",randomkeyCommand,1,0,NULL,0,0,0,0,0}, {"randomkey",randomkeyCommand,1,0,NULL,0,0,0,0,0},
{"select",selectCommand,2,0,NULL,0,0,0,0,0}, {"select",selectCommand,2,0,NULL,0,0,0,0,0},
{"move",moveCommand,3,0,NULL,1,1,1,0,0}, {"move",moveCommand,3,0,NULL,1,1,1,0,0},
{"rename",renameCommand,3,0,NULL,1,1,1,0,0}, {"rename",renameCommand,3,0,renameGetKeys,1,2,1,0,0},
{"renamenx",renamenxCommand,3,0,NULL,1,1,1,0,0}, {"renamenx",renamenxCommand,3,0,renameGetKeys,1,2,1,0,0},
{"expire",expireCommand,3,0,NULL,0,0,0,0,0}, {"expire",expireCommand,3,0,NULL,1,1,1,0,0},
{"expireat",expireatCommand,3,0,NULL,0,0,0,0,0}, {"expireat",expireatCommand,3,0,NULL,1,1,1,0,0},
{"keys",keysCommand,2,0,NULL,0,0,0,0,0}, {"keys",keysCommand,2,0,NULL,0,0,0,0,0},
{"dbsize",dbsizeCommand,1,0,NULL,0,0,0,0,0}, {"dbsize",dbsizeCommand,1,0,NULL,0,0,0,0,0},
{"auth",authCommand,2,0,NULL,0,0,0,0,0}, {"auth",authCommand,2,0,NULL,0,0,0,0,0},
...@@ -168,7 +168,7 @@ struct redisCommand redisCommandTable[] = { ...@@ -168,7 +168,7 @@ struct redisCommand redisCommandTable[] = {
{"lastsave",lastsaveCommand,1,0,NULL,0,0,0,0,0}, {"lastsave",lastsaveCommand,1,0,NULL,0,0,0,0,0},
{"type",typeCommand,2,0,NULL,1,1,1,0,0}, {"type",typeCommand,2,0,NULL,1,1,1,0,0},
{"multi",multiCommand,1,0,NULL,0,0,0,0,0}, {"multi",multiCommand,1,0,NULL,0,0,0,0,0},
{"exec",execCommand,1,REDIS_CMD_DENYOOM,execBlockClientOnSwappedKeys,0,0,0,0,0}, {"exec",execCommand,1,REDIS_CMD_DENYOOM,NULL,0,0,0,0,0},
{"discard",discardCommand,1,0,NULL,0,0,0,0,0}, {"discard",discardCommand,1,0,NULL,0,0,0,0,0},
{"sync",syncCommand,1,0,NULL,0,0,0,0,0}, {"sync",syncCommand,1,0,NULL,0,0,0,0,0},
{"flushdb",flushdbCommand,1,0,NULL,0,0,0,0,0}, {"flushdb",flushdbCommand,1,0,NULL,0,0,0,0,0},
...@@ -186,8 +186,14 @@ struct redisCommand redisCommandTable[] = { ...@@ -186,8 +186,14 @@ struct redisCommand redisCommandTable[] = {
{"psubscribe",psubscribeCommand,-2,0,NULL,0,0,0,0,0}, {"psubscribe",psubscribeCommand,-2,0,NULL,0,0,0,0,0},
{"punsubscribe",punsubscribeCommand,-1,0,NULL,0,0,0,0,0}, {"punsubscribe",punsubscribeCommand,-1,0,NULL,0,0,0,0,0},
{"publish",publishCommand,3,REDIS_CMD_FORCE_REPLICATION,NULL,0,0,0,0,0}, {"publish",publishCommand,3,REDIS_CMD_FORCE_REPLICATION,NULL,0,0,0,0,0},
{"watch",watchCommand,-2,0,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,4,0,NULL,0,0,0,0,0},
{"migrate",migrateCommand,6,0,NULL,0,0,0,0,0},
{"dump",dumpCommand,2,0,NULL,0,0,0,0,0},
{"object",objectCommand,-2,0,NULL,0,0,0,0,0},
{"client",clientCommand,-2,0,NULL,0,0,0,0,0}
}; };
/*============================ Utility functions ============================ */ /*============================ Utility functions ============================ */
...@@ -200,14 +206,20 @@ void redisLogRaw(int level, const char *msg) { ...@@ -200,14 +206,20 @@ void redisLogRaw(int level, const char *msg) {
time_t now = time(NULL); time_t now = time(NULL);
FILE *fp; FILE *fp;
char buf[64]; char buf[64];
int rawmode = (level & REDIS_LOG_RAW);
level &= 0xff; /* clear flags */
if (level < server.verbosity) return; if (level < server.verbosity) return;
fp = (server.logfile == NULL) ? stdout : fopen(server.logfile,"a"); fp = (server.logfile == NULL) ? stdout : fopen(server.logfile,"a");
if (!fp) return; if (!fp) return;
if (rawmode) {
fprintf(fp,"%s",msg);
} else {
strftime(buf,sizeof(buf),"%d %b %H:%M:%S",localtime(&now)); strftime(buf,sizeof(buf),"%d %b %H:%M:%S",localtime(&now));
fprintf(fp,"[%d] %s %c %s\n",(int)getpid(),buf,c[level],msg); fprintf(fp,"[%d] %s %c %s\n",(int)getpid(),buf,c[level],msg);
}
fflush(fp); fflush(fp);
if (server.logfile) fclose(fp); if (server.logfile) fclose(fp);
...@@ -222,7 +234,7 @@ void redisLog(int level, const char *fmt, ...) { ...@@ -222,7 +234,7 @@ void redisLog(int level, const char *fmt, ...) {
va_list ap; va_list ap;
char msg[REDIS_MAX_LOGMSG_LEN]; char msg[REDIS_MAX_LOGMSG_LEN];
if (level < server.verbosity) return; if ((level&0xff) < server.verbosity) return;
va_start(ap, fmt); va_start(ap, fmt);
vsnprintf(msg, sizeof(msg), fmt, ap); vsnprintf(msg, sizeof(msg), fmt, ap);
...@@ -440,6 +452,17 @@ dictType keylistDictType = { ...@@ -440,6 +452,17 @@ dictType keylistDictType = {
dictListDestructor /* val destructor */ 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) { int htNeedsResize(dict *dict) {
long long size, used; long long size, used;
...@@ -563,6 +586,10 @@ int serverCron(struct aeEventLoop *eventLoop, long long id, void *clientData) { ...@@ -563,6 +586,10 @@ int serverCron(struct aeEventLoop *eventLoop, long long id, void *clientData) {
*/ */
updateLRUClock(); updateLRUClock();
/* Record the max memory used since the server was started. */
if (zmalloc_used_memory() > server.stat_peak_memory)
server.stat_peak_memory = zmalloc_used_memory();
/* We received a SIGTERM, shutting down here in a safe way, as it is /* We received a SIGTERM, shutting down here in a safe way, as it is
* not ok doing so inside the signal handler. */ * not ok doing so inside the signal handler. */
if (server.shutdown_asap) { if (server.shutdown_asap) {
...@@ -669,6 +696,9 @@ int serverCron(struct aeEventLoop *eventLoop, long long id, void *clientData) { ...@@ -669,6 +696,9 @@ int serverCron(struct aeEventLoop *eventLoop, long long id, void *clientData) {
* to detect transfer failures. */ * to detect transfer failures. */
if (!(loops % 10)) replicationCron(); if (!(loops % 10)) replicationCron();
/* Run other sub-systems specific cron jobs */
if (server.cluster_enabled && !(loops % 10)) clusterCron();
server.cronloops++; server.cronloops++;
return 100; return 100;
} }
...@@ -821,8 +851,12 @@ void initServerConfig() { ...@@ -821,8 +851,12 @@ void initServerConfig() {
server.list_max_ziplist_entries = REDIS_LIST_MAX_ZIPLIST_ENTRIES; server.list_max_ziplist_entries = REDIS_LIST_MAX_ZIPLIST_ENTRIES;
server.list_max_ziplist_value = REDIS_LIST_MAX_ZIPLIST_VALUE; server.list_max_ziplist_value = REDIS_LIST_MAX_ZIPLIST_VALUE;
server.set_max_intset_entries = REDIS_SET_MAX_INTSET_ENTRIES; server.set_max_intset_entries = REDIS_SET_MAX_INTSET_ENTRIES;
server.zset_max_ziplist_entries = REDIS_ZSET_MAX_ZIPLIST_ENTRIES;
server.zset_max_ziplist_value = REDIS_ZSET_MAX_ZIPLIST_VALUE;
server.shutdown_asap = 0; server.shutdown_asap = 0;
server.cache_flush_delay = 0; server.cache_flush_delay = 0;
server.cluster_enabled = 0;
server.cluster.configfile = zstrdup("nodes.conf");
updateLRUClock(); updateLRUClock();
resetServerSaveParams(); resetServerSaveParams();
...@@ -928,6 +962,7 @@ void initServer() { ...@@ -928,6 +962,7 @@ void initServer() {
server.stat_starttime = time(NULL); server.stat_starttime = time(NULL);
server.stat_keyspace_misses = 0; server.stat_keyspace_misses = 0;
server.stat_keyspace_hits = 0; server.stat_keyspace_hits = 0;
server.stat_peak_memory = 0;
server.unixtime = time(NULL); server.unixtime = time(NULL);
aeCreateTimeEvent(server.el, 1, serverCron, NULL, NULL); aeCreateTimeEvent(server.el, 1, serverCron, NULL, NULL);
if (server.ipfd > 0 && aeCreateFileEvent(server.el,server.ipfd,AE_READABLE, if (server.ipfd > 0 && aeCreateFileEvent(server.el,server.ipfd,AE_READABLE,
...@@ -945,6 +980,7 @@ void initServer() { ...@@ -945,6 +980,7 @@ void initServer() {
} }
if (server.ds_enabled) dsInit(); if (server.ds_enabled) dsInit();
if (server.cluster_enabled) clusterInit();
srand(time(NULL)^getpid()); srand(time(NULL)^getpid());
} }
...@@ -1051,6 +1087,29 @@ int processCommand(redisClient *c) { ...@@ -1051,6 +1087,29 @@ int processCommand(redisClient *c) {
return REDIS_OK; 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 {
int ask;
clusterNode *n = getNodeByQuery(c,cmd,c->argv,c->argc,&hashslot,&ask);
if (n == NULL) {
addReplyError(c,"Multi keys request invalid in cluster");
return REDIS_OK;
} else if (n != server.cluster.myself) {
addReplySds(c,sdscatprintf(sdsempty(),
"-%s %d %s:%d\r\n", ask ? "ASK" : "MOVED",
hashslot,n->ip,n->port));
return REDIS_OK;
}
}
}
/* Handle the maxmemory directive. /* Handle the maxmemory directive.
* *
* First we try to free some memory if possible (if there are volatile * First we try to free some memory if possible (if there are volatile
...@@ -1189,7 +1248,6 @@ sds genRedisInfoString(char *section) { ...@@ -1189,7 +1248,6 @@ sds genRedisInfoString(char *section) {
sds info = sdsempty(); sds info = sdsempty();
time_t uptime = time(NULL)-server.stat_starttime; time_t uptime = time(NULL)-server.stat_starttime;
int j, numcommands; int j, numcommands;
char hmem[64];
struct rusage self_ru, c_ru; struct rusage self_ru, c_ru;
unsigned long lol, bib; unsigned long lol, bib;
int allsections = 0, defsections = 0; int allsections = 0, defsections = 0;
...@@ -1203,7 +1261,6 @@ sds genRedisInfoString(char *section) { ...@@ -1203,7 +1261,6 @@ sds genRedisInfoString(char *section) {
getrusage(RUSAGE_SELF, &self_ru); getrusage(RUSAGE_SELF, &self_ru);
getrusage(RUSAGE_CHILDREN, &c_ru); getrusage(RUSAGE_CHILDREN, &c_ru);
getClientsMaxBuffers(&lol,&bib); getClientsMaxBuffers(&lol,&bib);
bytesToHuman(hmem,zmalloc_used_memory());
/* Server */ /* Server */
if (allsections || defsections || !strcasecmp(section,"server")) { if (allsections || defsections || !strcasecmp(section,"server")) {
...@@ -1248,23 +1305,28 @@ sds genRedisInfoString(char *section) { ...@@ -1248,23 +1305,28 @@ sds genRedisInfoString(char *section) {
/* Memory */ /* Memory */
if (allsections || defsections || !strcasecmp(section,"memory")) { if (allsections || defsections || !strcasecmp(section,"memory")) {
char hmem[64];
char peak_hmem[64];
bytesToHuman(hmem,zmalloc_used_memory());
bytesToHuman(peak_hmem,server.stat_peak_memory);
if (sections++) info = sdscat(info,"\r\n"); if (sections++) info = sdscat(info,"\r\n");
info = sdscatprintf(info, info = sdscatprintf(info,
"# Memory\r\n" "# Memory\r\n"
"used_memory:%zu\r\n" "used_memory:%zu\r\n"
"used_memory_human:%s\r\n" "used_memory_human:%s\r\n"
"used_memory_rss:%zu\r\n" "used_memory_rss:%zu\r\n"
"used_memory_peak:%zu\r\n"
"used_memory_peak_human:%s\r\n"
"mem_fragmentation_ratio:%.2f\r\n" "mem_fragmentation_ratio:%.2f\r\n"
"use_tcmalloc:%d\r\n", "mem_allocator:%s\r\n",
zmalloc_used_memory(), zmalloc_used_memory(),
hmem, hmem,
zmalloc_get_rss(), zmalloc_get_rss(),
server.stat_peak_memory,
peak_hmem,
zmalloc_get_fragmentation_ratio(), zmalloc_get_fragmentation_ratio(),
#ifdef USE_TCMALLOC REDIS_MALLOC
1
#else
0
#endif
); );
} }
...@@ -1455,6 +1517,15 @@ sds genRedisInfoString(char *section) { ...@@ -1455,6 +1517,15 @@ sds genRedisInfoString(char *section) {
} }
} }
/* Clusetr */
if (allsections || defsections || !strcasecmp(section,"cluster")) {
if (sections++) info = sdscat(info,"\r\n");
info = sdscatprintf(info,
"# Cluster\r\n"
"cluster_enabled:%d\r\n",
server.cluster_enabled);
}
/* Key space */ /* Key space */
if (allsections || defsections || !strcasecmp(section,"keyspace")) { if (allsections || defsections || !strcasecmp(section,"keyspace")) {
if (sections++) info = sdscat(info,"\r\n"); if (sections++) info = sdscat(info,"\r\n");
...@@ -1663,6 +1734,23 @@ void usage() { ...@@ -1663,6 +1734,23 @@ void usage() {
exit(1); exit(1);
} }
void redisAsciiArt(void) {
#include "asciilogo.h"
char *buf = zmalloc(1024*16);
snprintf(buf,1024*16,ascii_logo,
REDIS_VERSION,
redisGitSHA1(),
strtol(redisGitDirty(),NULL,10) > 0,
(sizeof(long) == 8) ? "64" : "32",
server.cluster_enabled ? "cluster" : "stand alone",
server.port,
(long) getpid()
);
redisLogRaw(REDIS_NOTICE|REDIS_LOG_RAW,buf);
zfree(buf);
}
int main(int argc, char **argv) { int main(int argc, char **argv) {
long long start; long long start;
...@@ -1681,6 +1769,7 @@ int main(int argc, char **argv) { ...@@ -1681,6 +1769,7 @@ int main(int argc, char **argv) {
if (server.daemonize) daemonize(); if (server.daemonize) daemonize();
initServer(); initServer();
if (server.daemonize) createPidFile(); if (server.daemonize) createPidFile();
redisAsciiArt();
redisLog(REDIS_NOTICE,"Server started, Redis version " REDIS_VERSION); redisLog(REDIS_NOTICE,"Server started, Redis version " REDIS_VERSION);
#ifdef __linux__ #ifdef __linux__
linuxOvercommitMemoryWarning(); linuxOvercommitMemoryWarning();
......
...@@ -18,6 +18,7 @@ ...@@ -18,6 +18,7 @@
#include <inttypes.h> #include <inttypes.h>
#include <pthread.h> #include <pthread.h>
#include <syslog.h> #include <syslog.h>
#include <netinet/in.h>
#include "ae.h" /* Event driven programming library */ #include "ae.h" /* Event driven programming library */
#include "sds.h" /* Dynamic safe strings */ #include "sds.h" /* Dynamic safe strings */
...@@ -29,6 +30,7 @@ ...@@ -29,6 +30,7 @@
#include "ziplist.h" /* Compact list data structure */ #include "ziplist.h" /* Compact list data structure */
#include "intset.h" /* Compact integer set structure */ #include "intset.h" /* Compact integer set structure */
#include "version.h" #include "version.h"
#include "util.h"
/* Error codes */ /* Error codes */
#define REDIS_OK 0 #define REDIS_OK 0
...@@ -70,10 +72,12 @@ ...@@ -70,10 +72,12 @@
#define REDIS_ZSET 3 #define REDIS_ZSET 3
#define REDIS_HASH 4 #define REDIS_HASH 4
#define REDIS_VMPOINTER 8 #define REDIS_VMPOINTER 8
/* Object types only used for persistence in .rdb files */ /* Object types only used for persistence in .rdb files */
#define REDIS_HASH_ZIPMAP 9 #define REDIS_HASH_ZIPMAP 9
#define REDIS_LIST_ZIPLIST 10 #define REDIS_LIST_ZIPLIST 10
#define REDIS_SET_INTSET 11 #define REDIS_SET_INTSET 11
#define REDIS_ZSET_ZIPLIST 12
/* Objects encoding. Some kind of objects like Strings and Hashes can be /* Objects encoding. Some kind of objects like Strings and Hashes can be
* internally represented in multiple ways. The 'encoding' field of the object * internally represented in multiple ways. The 'encoding' field of the object
...@@ -85,6 +89,7 @@ ...@@ -85,6 +89,7 @@
#define REDIS_ENCODING_LINKEDLIST 4 /* Encoded as regular linked list */ #define REDIS_ENCODING_LINKEDLIST 4 /* Encoded as regular linked list */
#define REDIS_ENCODING_ZIPLIST 5 /* Encoded as ziplist */ #define REDIS_ENCODING_ZIPLIST 5 /* Encoded as ziplist */
#define REDIS_ENCODING_INTSET 6 /* Encoded as intset */ #define REDIS_ENCODING_INTSET 6 /* Encoded as intset */
#define REDIS_ENCODING_SKIPLIST 7 /* Encoded as skiplist */
/* Object types only used for dumping to disk */ /* Object types only used for dumping to disk */
#define REDIS_EXPIRETIME 253 #define REDIS_EXPIRETIME 253
...@@ -177,6 +182,7 @@ ...@@ -177,6 +182,7 @@
#define REDIS_VERBOSE 1 #define REDIS_VERBOSE 1
#define REDIS_NOTICE 2 #define REDIS_NOTICE 2
#define REDIS_WARNING 3 #define REDIS_WARNING 3
#define REDIS_LOG_RAW (1<<10) /* Modifier to log without timestamp */
/* Anti-warning macro... */ /* Anti-warning macro... */
#define REDIS_NOTUSED(V) ((void) V) #define REDIS_NOTUSED(V) ((void) V)
...@@ -195,6 +201,8 @@ ...@@ -195,6 +201,8 @@
#define REDIS_LIST_MAX_ZIPLIST_ENTRIES 512 #define REDIS_LIST_MAX_ZIPLIST_ENTRIES 512
#define REDIS_LIST_MAX_ZIPLIST_VALUE 64 #define REDIS_LIST_MAX_ZIPLIST_VALUE 64
#define REDIS_SET_MAX_INTSET_ENTRIES 512 #define REDIS_SET_MAX_INTSET_ENTRIES 512
#define REDIS_ZSET_MAX_ZIPLIST_ENTRIES 128
#define REDIS_ZSET_MAX_ZIPLIST_VALUE 64
/* Sets operations codes */ /* Sets operations codes */
#define REDIS_OP_UNION 0 #define REDIS_OP_UNION 0
...@@ -360,7 +368,147 @@ struct sharedObjectsStruct { ...@@ -360,7 +368,147 @@ struct sharedObjectsStruct {
*integers[REDIS_SHARED_INTEGERS]; *integers[REDIS_SHARED_INTEGERS];
}; };
/* Global server state structure */ /* ZSETs use a specialized version of Skiplists */
typedef struct zskiplistNode {
robj *obj;
double score;
struct zskiplistNode *backward;
struct zskiplistLevel {
struct zskiplistNode *forward;
unsigned int span;
} level[];
} zskiplistNode;
typedef struct zskiplist {
struct zskiplistNode *header, *tail;
unsigned long length;
int level;
} zskiplist;
typedef struct zset {
dict *dict;
zskiplist *zsl;
} zset;
/*-----------------------------------------------------------------------------
* 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 {
char *configfile;
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];
zskiplist *slots_to_keys;
} 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 { struct redisServer {
/* General */ /* General */
pthread_t mainthread; pthread_t mainthread;
...@@ -373,6 +521,7 @@ struct redisServer { ...@@ -373,6 +521,7 @@ struct redisServer {
char *unixsocket; char *unixsocket;
int ipfd; int ipfd;
int sofd; int sofd;
int cfd;
list *clients; list *clients;
list *slaves, *monitors; list *slaves, *monitors;
char neterr[ANET_ERR_LEN]; char neterr[ANET_ERR_LEN];
...@@ -393,6 +542,7 @@ struct redisServer { ...@@ -393,6 +542,7 @@ struct redisServer {
long long stat_evictedkeys; /* number of evicted keys (maxmemory) */ long long stat_evictedkeys; /* number of evicted keys (maxmemory) */
long long stat_keyspace_hits; /* number of successful lookups of keys */ long long stat_keyspace_hits; /* number of successful lookups of keys */
long long stat_keyspace_misses; /* number of failed lookups of keys */ long long stat_keyspace_misses; /* number of failed lookups of keys */
size_t stat_peak_memory; /* max used memory record */
/* Configuration */ /* Configuration */
int verbosity; int verbosity;
int maxidletime; int maxidletime;
...@@ -468,6 +618,8 @@ struct redisServer { ...@@ -468,6 +618,8 @@ struct redisServer {
size_t list_max_ziplist_entries; size_t list_max_ziplist_entries;
size_t list_max_ziplist_value; size_t list_max_ziplist_value;
size_t set_max_intset_entries; size_t set_max_intset_entries;
size_t zset_max_ziplist_entries;
size_t zset_max_ziplist_value;
time_t unixtime; /* Unix time sampled every second. */ time_t unixtime; /* Unix time sampled every second. */
/* Virtual memory I/O threads stuff */ /* Virtual memory I/O threads stuff */
/* An I/O thread process an element taken from the io_jobs queue and /* An I/O thread process an element taken from the io_jobs queue and
...@@ -499,6 +651,9 @@ struct redisServer { ...@@ -499,6 +651,9 @@ struct redisServer {
/* Misc */ /* Misc */
unsigned lruclock:22; /* clock incrementing every minute, for LRU */ unsigned lruclock:22; /* clock incrementing every minute, for LRU */
unsigned lruclock_padding:10; unsigned lruclock_padding:10;
/* Cluster */
int cluster_enabled;
clusterState cluster;
}; };
typedef struct pubsubPattern { typedef struct pubsubPattern {
...@@ -507,20 +662,19 @@ typedef struct pubsubPattern { ...@@ -507,20 +662,19 @@ typedef struct pubsubPattern {
} pubsubPattern; } pubsubPattern;
typedef void redisCommandProc(redisClient *c); typedef void redisCommandProc(redisClient *c);
typedef void redisVmPreloadProc(redisClient *c, struct redisCommand *cmd, int argc, robj **argv); typedef int *redisGetKeysProc(struct redisCommand *cmd, robj **argv, int argc, int *numkeys, int flags);
struct redisCommand { struct redisCommand {
char *name; char *name;
redisCommandProc *proc; redisCommandProc *proc;
int arity; int arity;
int flags; int flags;
/* Use a function to determine which keys need to be loaded /* Use a function to determine keys arguments in a command line.
* in the background prior to executing this command. Takes precedence * Used both for diskstore preloading and Redis Cluster. */
* over vm_firstkey and others, ignored when NULL */ redisGetKeysProc *getkeys_proc;
redisVmPreloadProc *vm_preload_proc;
/* What keys should be loaded in background when calling this command? */ /* What keys should be loaded in background when calling this command? */
int vm_firstkey; /* The first argument that's a key (0 = no keys) */ int firstkey; /* The first argument that's a key (0 = no keys) */
int vm_lastkey; /* THe last argument that's a key */ int lastkey; /* THe last argument that's a key */
int vm_keystep; /* The step between first and last key */ int keystep; /* The step between first and last key */
long long microseconds, calls; long long microseconds, calls;
}; };
...@@ -542,28 +696,6 @@ typedef struct _redisSortOperation { ...@@ -542,28 +696,6 @@ typedef struct _redisSortOperation {
robj *pattern; robj *pattern;
} redisSortOperation; } redisSortOperation;
/* ZSETs use a specialized version of Skiplists */
typedef struct zskiplistNode {
robj *obj;
double score;
struct zskiplistNode *backward;
struct zskiplistLevel {
struct zskiplistNode *forward;
unsigned int span;
} level[];
} zskiplistNode;
typedef struct zskiplist {
struct zskiplistNode *header, *tail;
unsigned long length;
int level;
} zskiplist;
typedef struct zset {
dict *dict;
zskiplist *zsl;
} zset;
/* DIsk store threaded I/O request message */ /* DIsk store threaded I/O request message */
#define REDIS_IOJOB_LOAD 0 #define REDIS_IOJOB_LOAD 0
#define REDIS_IOJOB_SAVE 1 #define REDIS_IOJOB_SAVE 1
...@@ -634,6 +766,7 @@ extern struct redisServer server; ...@@ -634,6 +766,7 @@ extern struct redisServer server;
extern struct sharedObjectsStruct shared; extern struct sharedObjectsStruct shared;
extern dictType setDictType; extern dictType setDictType;
extern dictType zsetDictType; extern dictType zsetDictType;
extern dictType clusterNodesDictType;
extern double R_Zero, R_PosInf, R_NegInf, R_Nan; extern double R_Zero, R_PosInf, R_NegInf, R_Nan;
dictType hashDictType; dictType hashDictType;
...@@ -720,6 +853,7 @@ void freeHashObject(robj *o); ...@@ -720,6 +853,7 @@ void freeHashObject(robj *o);
robj *createObject(int type, void *ptr); robj *createObject(int type, void *ptr);
robj *createStringObject(char *ptr, size_t len); robj *createStringObject(char *ptr, size_t len);
robj *dupStringObject(robj *o); robj *dupStringObject(robj *o);
int isObjectRepresentableAsLongLong(robj *o, long long *llongval);
robj *tryObjectEncoding(robj *o); robj *tryObjectEncoding(robj *o);
robj *getDecodedObject(robj *o); robj *getDecodedObject(robj *o);
size_t stringObjectLen(robj *o); size_t stringObjectLen(robj *o);
...@@ -730,6 +864,7 @@ robj *createSetObject(void); ...@@ -730,6 +864,7 @@ robj *createSetObject(void);
robj *createIntsetObject(void); robj *createIntsetObject(void);
robj *createHashObject(void); robj *createHashObject(void);
robj *createZsetObject(void); robj *createZsetObject(void);
robj *createZsetZiplistObject(void);
int getLongFromObjectOrReply(redisClient *c, robj *o, long *target, const char *msg); int getLongFromObjectOrReply(redisClient *c, robj *o, long *target, const char *msg);
int checkType(redisClient *c, robj *o, int type); int checkType(redisClient *c, robj *o, int type);
int getLongLongFromObjectOrReply(redisClient *c, robj *o, long long *target, const char *msg); int getLongLongFromObjectOrReply(redisClient *c, robj *o, long long *target, const char *msg);
...@@ -748,6 +883,7 @@ int fwriteBulkString(FILE *fp, char *s, unsigned long len); ...@@ -748,6 +883,7 @@ int fwriteBulkString(FILE *fp, char *s, unsigned long len);
int fwriteBulkDouble(FILE *fp, double d); int fwriteBulkDouble(FILE *fp, double d);
int fwriteBulkLongLong(FILE *fp, long long l); int fwriteBulkLongLong(FILE *fp, long long l);
int fwriteBulkObject(FILE *fp, robj *obj); int fwriteBulkObject(FILE *fp, robj *obj);
int fwriteBulkCount(FILE *fp, char prefix, int count);
/* Replication */ /* Replication */
void replicationFeedSlaves(list *slaves, int dictid, robj **argv, int argc); void replicationFeedSlaves(list *slaves, int dictid, robj **argv, int argc);
...@@ -789,9 +925,24 @@ int startAppendOnly(void); ...@@ -789,9 +925,24 @@ int startAppendOnly(void);
void backgroundRewriteDoneHandler(int exitcode, int bysignal); void backgroundRewriteDoneHandler(int exitcode, int bysignal);
/* Sorted sets data type */ /* Sorted sets data type */
/* Struct to hold a inclusive/exclusive range spec. */
typedef struct {
double min, max;
int minex, maxex; /* are min or max exclusive? */
} zrangespec;
zskiplist *zslCreate(void); zskiplist *zslCreate(void);
void zslFree(zskiplist *zsl); void zslFree(zskiplist *zsl);
zskiplistNode *zslInsert(zskiplist *zsl, double score, robj *obj); zskiplistNode *zslInsert(zskiplist *zsl, double score, robj *obj);
unsigned char *zzlInsert(unsigned char *zl, robj *ele, double score);
int zslDelete(zskiplist *zsl, double score, robj *obj);
zskiplistNode *zslFirstInRange(zskiplist *zsl, zrangespec range);
double zzlGetScore(unsigned char *sptr);
void zzlNext(unsigned char *zl, unsigned char **eptr, unsigned char **sptr);
void zzlPrev(unsigned char *zl, unsigned char **eptr, unsigned char **sptr);
unsigned int zsetLength(robj *zobj);
void zsetConvert(robj *zobj, int encoding);
/* Core functions */ /* Core functions */
void freeMemoryIfNeeded(void); void freeMemoryIfNeeded(void);
...@@ -829,8 +980,6 @@ void freeIOJob(iojob *j); ...@@ -829,8 +980,6 @@ void freeIOJob(iojob *j);
void queueIOJob(iojob *j); void queueIOJob(iojob *j);
void waitEmptyIOJobsQueue(void); void waitEmptyIOJobsQueue(void);
void processAllPendingIOJobs(void); void processAllPendingIOJobs(void);
void zunionInterBlockClientOnSwappedKeys(redisClient *c, struct redisCommand *cmd, int argc, robj **argv);
void execBlockClientOnSwappedKeys(redisClient *c, struct redisCommand *cmd, int argc, robj **argv);
int blockClientOnSwappedKeys(redisClient *c, struct redisCommand *cmd); int blockClientOnSwappedKeys(redisClient *c, struct redisCommand *cmd);
int dontWaitForSwappedKey(redisClient *c, robj *key); int dontWaitForSwappedKey(redisClient *c, robj *key);
void handleClientsBlockedOnSwappedKey(redisDb *db, robj *key); void handleClientsBlockedOnSwappedKey(redisDb *db, robj *key);
...@@ -881,16 +1030,6 @@ int pubsubUnsubscribeAllPatterns(redisClient *c, int notify); ...@@ -881,16 +1030,6 @@ int pubsubUnsubscribeAllPatterns(redisClient *c, int notify);
void freePubsubPattern(void *p); void freePubsubPattern(void *p);
int listMatchPubsubPattern(void *a, void *b); int listMatchPubsubPattern(void *a, void *b);
/* Utility functions */
int stringmatchlen(const char *pattern, int patternLen,
const char *string, int stringLen, int nocase);
int stringmatch(const char *pattern, const char *string, int nocase);
long long memtoll(const char *p, int *err);
int ll2string(char *s, size_t len, long long value);
int isStringRepresentableAsLong(sds s, long *longval);
int isStringRepresentableAsLongLong(sds s, long long *longval);
int isObjectRepresentableAsLongLong(robj *o, long long *llongval);
/* Configuration */ /* Configuration */
void loadServerConfig(char *filename); void loadServerConfig(char *filename);
void appendServerSaveParams(time_t seconds, int changes); void appendServerSaveParams(time_t seconds, int changes);
...@@ -916,6 +1055,25 @@ long long emptyDb(); ...@@ -916,6 +1055,25 @@ long long emptyDb();
int selectDb(redisClient *c, int id); int selectDb(redisClient *c, int id);
void signalModifiedKey(redisDb *db, robj *key); void signalModifiedKey(redisDb *db, robj *key);
void signalFlushedDb(int dbid); void signalFlushedDb(int dbid);
unsigned int GetKeysInSlot(unsigned int hashslot, robj **keys, unsigned int count);
/* API to get key arguments from commands */
#define REDIS_GETKEYS_ALL 0
#define REDIS_GETKEYS_PRELOAD 1
int *getKeysFromCommand(struct redisCommand *cmd, robj **argv, int argc, int *numkeys, int flags);
void getKeysFreeResult(int *result);
int *noPreloadGetKeys(struct redisCommand *cmd,robj **argv, int argc, int *numkeys, int flags);
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, int *ask);
/* Git SHA1 */ /* Git SHA1 */
char *redisGitSHA1(void); char *redisGitSHA1(void);
...@@ -1039,6 +1197,12 @@ void punsubscribeCommand(redisClient *c); ...@@ -1039,6 +1197,12 @@ void punsubscribeCommand(redisClient *c);
void publishCommand(redisClient *c); void publishCommand(redisClient *c);
void watchCommand(redisClient *c); void watchCommand(redisClient *c);
void unwatchCommand(redisClient *c); void unwatchCommand(redisClient *c);
void clusterCommand(redisClient *c);
void restoreCommand(redisClient *c);
void migrateCommand(redisClient *c);
void dumpCommand(redisClient *c);
void objectCommand(redisClient *c);
void clientCommand(redisClient *c);
#if defined(__GNUC__) #if defined(__GNUC__)
void *calloc(size_t count, size_t size) __attribute__ ((deprecated)); void *calloc(size_t count, size_t size) __attribute__ ((deprecated));
......
...@@ -36,11 +36,11 @@ ...@@ -36,11 +36,11 @@
#define SDS_ABORT_ON_OOM #define SDS_ABORT_ON_OOM
#include "sds.h"
#include <stdio.h> #include <stdio.h>
#include <stdlib.h> #include <stdlib.h>
#include <string.h> #include <string.h>
#include <ctype.h> #include <ctype.h>
#include "sds.h"
#include "zmalloc.h" #include "zmalloc.h"
static void sdsOomAbort(void) { static void sdsOomAbort(void) {
...@@ -51,7 +51,11 @@ static void sdsOomAbort(void) { ...@@ -51,7 +51,11 @@ static void sdsOomAbort(void) {
sds sdsnewlen(const void *init, size_t initlen) { sds sdsnewlen(const void *init, size_t initlen) {
struct sdshdr *sh; struct sdshdr *sh;
if (init) {
sh = zmalloc(sizeof(struct sdshdr)+initlen+1); sh = zmalloc(sizeof(struct sdshdr)+initlen+1);
} else {
sh = zcalloc(sizeof(struct sdshdr)+initlen+1);
}
#ifdef SDS_ABORT_ON_OOM #ifdef SDS_ABORT_ON_OOM
if (sh == NULL) sdsOomAbort(); if (sh == NULL) sdsOomAbort();
#else #else
...@@ -59,10 +63,8 @@ sds sdsnewlen(const void *init, size_t initlen) { ...@@ -59,10 +63,8 @@ sds sdsnewlen(const void *init, size_t initlen) {
#endif #endif
sh->len = initlen; sh->len = initlen;
sh->free = 0; sh->free = 0;
if (initlen) { if (initlen && init)
if (init) memcpy(sh->buf, init, initlen); memcpy(sh->buf, init, initlen);
else memset(sh->buf,0,initlen);
}
sh->buf[initlen] = '\0'; sh->buf[initlen] = '\0';
return (char*)sh->buf; return (char*)sh->buf;
} }
...@@ -76,11 +78,6 @@ sds sdsnew(const char *init) { ...@@ -76,11 +78,6 @@ sds sdsnew(const char *init) {
return sdsnewlen(init, initlen); return sdsnewlen(init, initlen);
} }
size_t sdslen(const sds s) {
struct sdshdr *sh = (void*) (s-(sizeof(struct sdshdr)));
return sh->len;
}
sds sdsdup(const sds s) { sds sdsdup(const sds s) {
return sdsnewlen(s, sdslen(s)); return sdsnewlen(s, sdslen(s));
} }
...@@ -90,11 +87,6 @@ void sdsfree(sds s) { ...@@ -90,11 +87,6 @@ void sdsfree(sds s) {
zfree(s-sizeof(struct sdshdr)); zfree(s-sizeof(struct sdshdr));
} }
size_t sdsavail(sds s) {
struct sdshdr *sh = (void*) (s-(sizeof(struct sdshdr)));
return sh->free;
}
void sdsupdatelen(sds s) { void sdsupdatelen(sds s) {
struct sdshdr *sh = (void*) (s-(sizeof(struct sdshdr))); struct sdshdr *sh = (void*) (s-(sizeof(struct sdshdr)));
int reallen = strlen(s); int reallen = strlen(s);
...@@ -306,15 +298,17 @@ int sdscmp(sds s1, sds s2) { ...@@ -306,15 +298,17 @@ int sdscmp(sds s1, sds s2) {
*/ */
sds *sdssplitlen(char *s, int len, char *sep, int seplen, int *count) { sds *sdssplitlen(char *s, int len, char *sep, int seplen, int *count) {
int elements = 0, slots = 5, start = 0, j; int elements = 0, slots = 5, start = 0, j;
sds *tokens;
if (seplen < 1 || len < 0) return NULL;
sds *tokens = zmalloc(sizeof(sds)*slots); tokens = zmalloc(sizeof(sds)*slots);
#ifdef SDS_ABORT_ON_OOM #ifdef SDS_ABORT_ON_OOM
if (tokens == NULL) sdsOomAbort(); if (tokens == NULL) sdsOomAbort();
#else
if (tokens == NULL) return NULL;
#endif #endif
if (seplen < 1 || len < 0 || tokens == NULL) {
*count = 0;
return NULL;
}
if (len == 0) { if (len == 0) {
*count = 0; *count = 0;
return tokens; return tokens;
...@@ -552,6 +546,13 @@ err: ...@@ -552,6 +546,13 @@ err:
return NULL; return NULL;
} }
void sdssplitargs_free(sds *argv, int argc) {
int j;
for (j = 0 ;j < argc; j++) sdsfree(argv[j]);
zfree(argv);
}
#ifdef SDS_TEST_MAIN #ifdef SDS_TEST_MAIN
#include <stdio.h> #include <stdio.h>
#include "testhelp.h" #include "testhelp.h"
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment