* POSSIBILITY OF SUCH DAMAGE.
*/
-#define REDIS_VERSION "0.09"
+#define REDIS_VERSION "0.091"
#include "fmacros.h"
/* Static server configuration */
#define REDIS_SERVERPORT 6379 /* TCP port */
#define REDIS_MAXIDLETIME (60*5) /* default client timeout */
-#define REDIS_QUERYBUF_LEN 1024
+#define REDIS_IOBUF_LEN 1024
#define REDIS_LOADBUF_LEN 1024
#define REDIS_MAX_ARGS 16
#define REDIS_DEFAULT_DBNUM 16
#define REDIS_HASH 3
/* Object types only used for dumping to disk */
+#define REDIS_EXPIRETIME 253
#define REDIS_SELECTDB 254
#define REDIS_EOF 255
#define REDIS_MASTER 4 /* This client is a master server */
#define REDIS_MONITOR 8 /* This client is a slave monitor, see MONITOR */
-/* Server replication state */
+/* Slave replication state - slave side */
#define REDIS_REPL_NONE 0 /* No active replication */
#define REDIS_REPL_CONNECT 1 /* Must connect to master */
#define REDIS_REPL_CONNECTED 2 /* Connected to master */
+/* Slave replication state - from the point of view of master
+ * Note that in SEND_BULK and ONLINE state the slave receives new updates
+ * in its output queue. In the WAIT_BGSAVE state instead the server is waiting
+ * to start the next background saving in order to send updates to it. */
+#define REDIS_REPL_WAIT_BGSAVE_START 3 /* master waits bgsave to start feeding it */
+#define REDIS_REPL_WAIT_BGSAVE_END 4 /* master waits bgsave to start bulk DB transmission */
+#define REDIS_REPL_SEND_BULK 5 /* master is sending the bulk DB */
+#define REDIS_REPL_ONLINE 6 /* bulk DB already transmitted, receive updates */
+
/* List related stuff */
#define REDIS_HEAD 0
#define REDIS_TAIL 1
/* A redis object, that is a type able to hold a string / list / set */
typedef struct redisObject {
- int type;
void *ptr;
+ int type;
int refcount;
} robj;
sds querybuf;
robj *argv[REDIS_MAX_ARGS];
int argc;
- int bulklen; /* bulk read len. -1 if not in bulk read mode */
+ int bulklen; /* bulk read len. -1 if not in bulk read mode */
list *reply;
int sentlen;
time_t lastinteraction; /* time of the last interaction, used for timeout */
- int flags; /* REDIS_CLOSE | REDIS_SLAVE | REDIS_MONITOR */
- int slaveseldb; /* slave selected db, if this client is a slave */
- int authenticated; /* when requirepass is non-NULL */
+ int flags; /* REDIS_CLOSE | REDIS_SLAVE | REDIS_MONITOR */
+ int slaveseldb; /* slave selected db, if this client is a slave */
+ int authenticated; /* when requirepass is non-NULL */
+ int replstate; /* replication state if this is a slave */
+ int repldbfd; /* replication DB file descriptor */
+ long repldboff; /* replication DB file offset */
+ off_t repldbsize; /* replication DB file size */
} redisClient;
struct saveparam {
int cronloops; /* number of times the cron function run */
list *objfreelist; /* A list of freed objects to avoid malloc() */
time_t lastsave; /* Unix time of last save succeeede */
- int usedmemory; /* Used memory in megabytes */
+ size_t usedmemory; /* Used memory in megabytes */
/* Fields used only for stats */
time_t stat_starttime; /* server start time */
long long stat_numcommands; /* number of processed commands */
int isslave;
char *masterhost;
int masterport;
- redisClient *master;
+ redisClient *master; /* client that is master for this slave */
int replstate;
/* Sort parameters - qsort_r() is only available under BSD so we
* have to take this state global, in order to pass it to sortCompare() */
static int expireIfNeeded(redisDb *db, robj *key);
static int deleteIfVolatile(redisDb *db, robj *key);
static int deleteKey(redisDb *db, robj *key);
+static time_t getExpire(redisDb *db, robj *key);
+static int setExpire(redisDb *db, robj *key, time_t when);
+static void updateSalvesWaitingBgsave(int bgsaveerr);
static void authCommand(redisClient *c);
static void pingCommand(redisClient *c);
static void scardCommand(redisClient *c);
static void sinterCommand(redisClient *c);
static void sinterstoreCommand(redisClient *c);
+static void sunionCommand(redisClient *c);
+static void sunionstoreCommand(redisClient *c);
static void syncCommand(redisClient *c);
static void flushdbCommand(redisClient *c);
static void flushallCommand(redisClient *c);
{"scard",scardCommand,2,REDIS_CMD_INLINE},
{"sinter",sinterCommand,-2,REDIS_CMD_INLINE},
{"sinterstore",sinterstoreCommand,-3,REDIS_CMD_INLINE},
+ {"sunion",sunionCommand,-2,REDIS_CMD_INLINE},
+ {"sunionstore",sunionstoreCommand,-3,REDIS_CMD_INLINE},
{"smembers",sinterCommand,2,REDIS_CMD_INLINE},
{"incrby",incrbyCommand,3,REDIS_CMD_INLINE},
{"decrby",decrbyCommand,3,REDIS_CMD_INLINE},
/* ====================== Redis server networking stuff ===================== */
void closeTimedoutClients(void) {
redisClient *c;
- listIter *li;
listNode *ln;
time_t now = time(NULL);
- li = listGetIterator(server.clients,AL_START_HEAD);
- if (!li) return;
- while ((ln = listNextElement(li)) != NULL) {
+ listRewind(server.clients);
+ while ((ln = listYield(server.clients)) != NULL) {
c = listNodeValue(ln);
if (!(c->flags & REDIS_SLAVE) && /* no timeout for slaves */
(now - c->lastinteraction > server.maxidletime)) {
freeClient(c);
}
}
- listReleaseIterator(li);
}
int serverCron(struct aeEventLoop *eventLoop, long long id, void *clientData) {
/* Show information about connected clients */
if (!(loops % 5)) {
- redisLog(REDIS_DEBUG,"%d clients connected (%d slaves), %d bytes in use",
+ redisLog(REDIS_DEBUG,"%d clients connected (%d slaves), %zu bytes in use",
listLength(server.clients)-listLength(server.slaves),
listLength(server.slaves),
server.usedmemory,
/* Check if a background saving in progress terminated */
if (server.bgsaveinprogress) {
int statloc;
+ /* XXX: TODO handle the case of the saving child killed */
if (wait4(-1,&statloc,WNOHANG,NULL)) {
int exitcode = WEXITSTATUS(statloc);
if (exitcode == 0) {
"Background saving error");
}
server.bgsaveinprogress = 0;
+ updateSalvesWaitingBgsave(exitcode == 0 ? REDIS_OK : REDIS_ERR);
}
} else {
/* If there is not a background saving in progress check if
assert(ln != NULL);
listDelNode(server.clients,ln);
if (c->flags & REDIS_SLAVE) {
+ if (c->replstate == REDIS_REPL_SEND_BULK && c->repldbfd != -1)
+ close(c->repldbfd);
list *l = (c->flags & REDIS_MONITOR) ? server.monitors : server.slaves;
ln = listSearchKey(l,c);
assert(ln != NULL);
static void glueReplyBuffersIfNeeded(redisClient *c) {
int totlen = 0;
- listNode *ln = c->reply->head, *next;
+ listNode *ln;
robj *o;
- while(ln) {
+ listRewind(c->reply);
+ while((ln = listYield(c->reply))) {
o = ln->value;
totlen += sdslen(o->ptr);
- ln = ln->next;
/* This optimization makes more sense if we don't have to copy
* too much data */
if (totlen > 1024) return;
char buf[1024];
int copylen = 0;
- ln = c->reply->head;
- while(ln) {
- next = ln->next;
+ listRewind(c->reply);
+ while((ln = listYield(c->reply))) {
o = ln->value;
memcpy(buf+copylen,o->ptr,sdslen(o->ptr));
copylen += sdslen(o->ptr);
listDelNode(c->reply,ln);
- ln = next;
}
/* Now the output buffer is empty, add the new single element */
addReplySds(c,sdsnewlen(buf,totlen));
}
static void replicationFeedSlaves(list *slaves, struct redisCommand *cmd, int dictid, robj **argv, int argc) {
- listNode *ln = slaves->head;
+ listNode *ln;
robj *outv[REDIS_MAX_ARGS*4]; /* enough room for args, spaces, newlines */
int outc = 0, j;
}
outv[outc++] = shared.crlf;
- while(ln) {
+ /* Increment all the refcounts at start and decrement at end in order to
+ * be sure to free objects if there is no slave in a replication state
+ * able to be feed with commands */
+ for (j = 0; j < outc; j++) incrRefCount(outv[j]);
+ listRewind(slaves);
+ while((ln = listYield(slaves))) {
redisClient *slave = ln->value;
+
+ /* Don't feed slaves that are still waiting for BGSAVE to start */
+ if (slave->replstate == REDIS_REPL_WAIT_BGSAVE_START) continue;
+
+ /* Feed all the other slaves, MONITORs and so on */
if (slave->slaveseldb != dictid) {
robj *selectcmd;
slave->slaveseldb = dictid;
}
for (j = 0; j < outc; j++) addReply(slave,outv[j]);
- ln = ln->next;
}
+ for (j = 0; j < outc; j++) decrRefCount(outv[j]);
}
static void readQueryFromClient(aeEventLoop *el, int fd, void *privdata, int mask) {
redisClient *c = (redisClient*) privdata;
- char buf[REDIS_QUERYBUF_LEN];
+ char buf[REDIS_IOBUF_LEN];
int nread;
REDIS_NOTUSED(el);
REDIS_NOTUSED(mask);
- nread = read(fd, buf, REDIS_QUERYBUF_LEN);
+ nread = read(fd, buf, REDIS_IOBUF_LEN);
if (nread == -1) {
if (errno == EAGAIN) {
nread = 0;
return REDIS_OK;
}
+static void *dupClientReplyValue(void *o) {
+ incrRefCount((robj*)o);
+ return 0;
+}
+
static redisClient *createClient(int fd) {
redisClient *c = zmalloc(sizeof(*c));
c->flags = 0;
c->lastinteraction = time(NULL);
c->authenticated = 0;
+ c->replstate = REDIS_REPL_NONE;
if ((c->reply = listCreate()) == NULL) oom("listCreate");
listSetFreeMethod(c->reply,decrRefCount);
+ listSetDupMethod(c->reply,dupClientReplyValue);
if (aeCreateFileEvent(server.el, c->fd, AE_READABLE,
readQueryFromClient, c, NULL) == AE_ERR) {
freeClient(c);
static void addReply(redisClient *c, robj *obj) {
if (listLength(c->reply) == 0 &&
+ (c->replstate == REDIS_REPL_NONE ||
+ c->replstate == REDIS_REPL_ONLINE) &&
aeCreateFileEvent(server.el, c->fd, AE_WRITABLE,
sendReplyToClient, c, NULL) == AE_ERR) return;
if (!listAddNodeTail(c->reply,obj)) oom("listAddNodeTail");
return 0;
}
+static int rdbSaveTime(FILE *fp, time_t t) {
+ int32_t t32 = (int32_t) t;
+ if (fwrite(&t32,4,1,fp) == 0) return -1;
+ return 0;
+}
+
/* check rdbLoadLen() comments for more info */
static int rdbSaveLen(FILE *fp, uint32_t len) {
unsigned char buf[2];
FILE *fp;
char tmpfile[256];
int j;
+ time_t now = time(NULL);
snprintf(tmpfile,256,"temp-%d.%ld.rdb",(int)time(NULL),(long int)random());
fp = fopen(tmpfile,"w");
}
if (fwrite("REDIS0001",9,1,fp) == 0) goto werr;
for (j = 0; j < server.dbnum; j++) {
- dict *d = server.db[j].dict;
+ redisDb *db = server.db+j;
+ dict *d = db->dict;
if (dictSize(d) == 0) continue;
di = dictGetIterator(d);
if (!di) {
while((de = dictNext(di)) != NULL) {
robj *key = dictGetEntryKey(de);
robj *o = dictGetEntryVal(de);
-
+ time_t expiretime = getExpire(db,key);
+
+ /* Save the expire time */
+ if (expiretime != -1) {
+ /* If this key is already expired skip it */
+ if (expiretime < now) continue;
+ if (rdbSaveType(fp,REDIS_EXPIRETIME) == -1) goto werr;
+ if (rdbSaveTime(fp,expiretime) == -1) goto werr;
+ }
+ /* Save the key and associated value */
if (rdbSaveType(fp,o->type) == -1) goto werr;
if (rdbSaveStringObject(fp,key) == -1) goto werr;
if (o->type == REDIS_STRING) {
} else if (o->type == REDIS_LIST) {
/* Save a list value */
list *list = o->ptr;
- listNode *ln = list->head;
+ listNode *ln;
+ listRewind(list);
if (rdbSaveLen(fp,listLength(list)) == -1) goto werr;
- while(ln) {
+ while((ln = listYield(list))) {
robj *eleobj = listNodeValue(ln);
if (rdbSaveStringObject(fp,eleobj) == -1) goto werr;
- ln = ln->next;
}
} else if (o->type == REDIS_SET) {
/* Save a set value */
return type;
}
+static time_t rdbLoadTime(FILE *fp) {
+ int32_t t32;
+ if (fread(&t32,4,1,fp) == 0) return -1;
+ return (time_t) t32;
+}
+
/* Load an encoded length from the DB, see the REDIS_RDB_* defines on the top
* of this file for a description of how this are stored on disk.
*
FILE *fp;
robj *keyobj = NULL;
uint32_t dbid;
- int type;
- int retval;
+ int type, retval, rdbver;
dict *d = server.db[0].dict;
+ redisDb *db = server.db+0;
char buf[1024];
- int rdbver;
+ time_t expiretime = -1, now = time(NULL);
+
fp = fopen(filename,"r");
if (!fp) return REDIS_ERR;
if (fread(buf,9,1,fp) == 0) goto eoferr;
/* Read type. */
if ((type = rdbLoadType(fp)) == -1) goto eoferr;
+ if (type == REDIS_EXPIRETIME) {
+ if ((expiretime = rdbLoadTime(fp)) == -1) goto eoferr;
+ /* We read the time so we need to read the object type again */
+ if ((type = rdbLoadType(fp)) == -1) goto eoferr;
+ }
if (type == REDIS_EOF) break;
/* Handle SELECT DB opcode as a special case */
if (type == REDIS_SELECTDB) {
redisLog(REDIS_WARNING,"FATAL: Data file was created with a Redis server configured to handle more than %d databases. Exiting\n", server.dbnum);
exit(1);
}
- d = server.db[dbid].dict;
+ db = server.db+dbid;
+ d = db->dict;
continue;
}
/* Read key */
redisLog(REDIS_WARNING,"Loading DB, duplicated key (%s) found! Unrecoverable error, exiting now.", keyobj->ptr);
exit(1);
}
+ /* Set the expire time if needed */
+ if (expiretime != -1) {
+ setExpire(db,keyobj,expiretime);
+ /* Delete this key if already expired */
+ if (expiretime < now) deleteKey(db,keyobj);
+ expiretime = -1;
+ }
keyobj = o = NULL;
}
fclose(fp);
while(1) {
de = dictGetRandomKey(c->db->dict);
- if (expireIfNeeded(c->db,dictGetEntryKey(de)) == 0) break;
+ if (!de || expireIfNeeded(c->db,dictGetEntryKey(de)) == 0) break;
}
if (de == NULL) {
+ addReply(c,shared.plus);
addReply(c,shared.crlf);
} else {
addReply(c,shared.plus);
}
static void saveCommand(redisClient *c) {
+ if (server.bgsaveinprogress) {
+ addReplySds(c,sdsnew("-ERR background save in progress\r\n"));
+ return;
+ }
if (rdbSave(server.dbfilename) == REDIS_OK) {
addReply(c,shared.ok);
} else {
static void shutdownCommand(redisClient *c) {
redisLog(REDIS_WARNING,"User requested shutdown, saving DB...");
+ /* XXX: TODO kill the child if there is a bgsave in progress */
if (rdbSave(server.dbfilename) == REDIS_OK) {
if (server.daemonize) {
unlink(server.pidfile);
lookupKeyRead(c->db,setskeys[j]);
if (!setobj) {
zfree(dv);
- addReply(c,shared.nokeyerr);
+ if (dstkey) {
+ deleteKey(c->db,dstkey);
+ addReply(c,shared.ok);
+ } else {
+ addReply(c,shared.nullmultibulk);
+ }
return;
}
if (setobj->type != REDIS_SET) {
}
dictReleaseIterator(di);
- if (!dstkey)
+ if (!dstkey) {
lenobj->ptr = sdscatprintf(sdsempty(),"*%d\r\n",cardinality);
- else
+ } else {
addReply(c,shared.ok);
+ server.dirty++;
+ }
zfree(dv);
}
sinterGenericCommand(c,c->argv+2,c->argc-2,c->argv[1]);
}
+static void sunionGenericCommand(redisClient *c, robj **setskeys, int setsnum, robj *dstkey) {
+ dict **dv = zmalloc(sizeof(dict*)*setsnum);
+ dictIterator *di;
+ dictEntry *de;
+ robj *lenobj = NULL, *dstset = NULL;
+ int j, cardinality = 0;
+
+ if (!dv) oom("sunionCommand");
+ for (j = 0; j < setsnum; j++) {
+ robj *setobj;
+
+ setobj = dstkey ?
+ lookupKeyWrite(c->db,setskeys[j]) :
+ lookupKeyRead(c->db,setskeys[j]);
+ if (!setobj) {
+ dv[j] = NULL;
+ continue;
+ }
+ if (setobj->type != REDIS_SET) {
+ zfree(dv);
+ addReply(c,shared.wrongtypeerr);
+ return;
+ }
+ dv[j] = setobj->ptr;
+ }
+
+ /* We need a temp set object to store our union. If the dstkey
+ * is not NULL (that is, we are inside an SUNIONSTORE operation) then
+ * this set object will be the resulting object to set into the target key*/
+ dstset = createSetObject();
+
+ /* The first thing we should output is the total number of elements...
+ * since this is a multi-bulk write, but at this stage we don't know
+ * the intersection set size, so we use a trick, append an empty object
+ * to the output list and save the pointer to later modify it with the
+ * right length */
+ if (!dstkey) {
+ lenobj = createObject(REDIS_STRING,NULL);
+ addReply(c,lenobj);
+ decrRefCount(lenobj);
+ } else {
+ /* If we have a target key where to store the resulting set
+ * create this key with an empty set inside */
+ deleteKey(c->db,dstkey);
+ dictAdd(c->db->dict,dstkey,dstset);
+ incrRefCount(dstkey);
+ server.dirty++;
+ }
+
+ /* Iterate all the elements of all the sets, add every element a single
+ * time to the result set */
+ for (j = 0; j < setsnum; j++) {
+ if (!dv[j]) continue; /* non existing keys are like empty sets */
+
+ di = dictGetIterator(dv[j]);
+ if (!di) oom("dictGetIterator");
+
+ while((de = dictNext(di)) != NULL) {
+ robj *ele;
+
+ /* dictAdd will not add the same element multiple times */
+ ele = dictGetEntryKey(de);
+ if (dictAdd(dstset->ptr,ele,NULL) == DICT_OK) {
+ incrRefCount(ele);
+ if (!dstkey) {
+ addReplySds(c,sdscatprintf(sdsempty(),
+ "$%d\r\n",sdslen(ele->ptr)));
+ addReply(c,ele);
+ addReply(c,shared.crlf);
+ cardinality++;
+ }
+ }
+ }
+ dictReleaseIterator(di);
+ }
+
+ if (!dstkey) {
+ lenobj->ptr = sdscatprintf(sdsempty(),"*%d\r\n",cardinality);
+ decrRefCount(dstset);
+ } else {
+ addReply(c,shared.ok);
+ server.dirty++;
+ }
+ zfree(dv);
+}
+
+static void sunionCommand(redisClient *c) {
+ sunionGenericCommand(c,c->argv+1,c->argc-1,NULL);
+}
+
+static void sunionstoreCommand(redisClient *c) {
+ sunionGenericCommand(c,c->argv+2,c->argc-2,c->argv[1]);
+}
+
static void flushdbCommand(redisClient *c) {
dictEmpty(c->db->dict);
dictEmpty(c->db->expires);
+ server.dirty++;
addReply(c,shared.ok);
rdbSave(server.dbfilename);
}
static void flushallCommand(redisClient *c) {
emptyDb();
+ server.dirty++;
addReply(c,shared.ok);
rdbSave(server.dbfilename);
}
j = 0;
if (sortval->type == REDIS_LIST) {
list *list = sortval->ptr;
- listNode *ln = list->head;
- while(ln) {
+ listNode *ln;
+
+ listRewind(list);
+ while((ln = listYield(list))) {
robj *ele = ln->value;
vector[j].obj = ele;
vector[j].u.score = 0;
vector[j].u.cmpobj = NULL;
- ln = ln->next;
j++;
}
} else {
outputlen = getop ? getop*(end-start+1) : end-start+1;
addReplySds(c,sdscatprintf(sdsempty(),"*%d\r\n",outputlen));
for (j = start; j <= end; j++) {
- listNode *ln = operations->head;
+ listNode *ln;
if (!getop) {
addReplySds(c,sdscatprintf(sdsempty(),"$%d\r\n",
sdslen(vector[j].obj->ptr)));
addReply(c,vector[j].obj);
addReply(c,shared.crlf);
}
- while(ln) {
+ listRewind(operations);
+ while((ln = listYield(operations))) {
redisSortOperation *sop = ln->value;
robj *val = lookupKeyByPattern(c->db,sop->pattern,
vector[j].obj);
} else if (sop->type == REDIS_SORT_DEL) {
/* TODO */
}
- ln = ln->next;
}
}
"redis_version:%s\r\n"
"connected_clients:%d\r\n"
"connected_slaves:%d\r\n"
- "used_memory:%d\r\n"
+ "used_memory:%zu\r\n"
"changes_since_last_save:%lld\r\n"
"last_save_time:%d\r\n"
"total_connections_received:%lld\r\n"
}
}
+/* Return the expire time of the specified key, or -1 if no expire
+ * is associated with this key (i.e. the key is non volatile) */
+static time_t getExpire(redisDb *db, robj *key) {
+ dictEntry *de;
+
+ /* No expire? return ASAP */
+ if (dictSize(db->expires) == 0 ||
+ (de = dictFind(db->expires,key)) == NULL) return -1;
+
+ return (time_t) dictGetEntryVal(de);
+}
+
static int expireIfNeeded(redisDb *db, robj *key) {
time_t when;
dictEntry *de;
(de = dictFind(db->expires,key)) == NULL) return 0;
/* Delete the key */
+ server.dirty++;
dictDelete(db->expires,key);
return dictDelete(db->dict,key) == DICT_OK;
}
/* =============================== Replication ============================= */
-/* Send the whole output buffer syncronously to the slave. This a general operation in theory, but it is actually useful only for replication. */
-static int flushClientOutput(redisClient *c) {
- int retval;
- time_t start = time(NULL);
-
- while(listLength(c->reply)) {
- if (time(NULL)-start > 5) return REDIS_ERR; /* 5 seconds timeout */
- retval = aeWait(c->fd,AE_WRITABLE,1000);
- if (retval == -1) {
- return REDIS_ERR;
- } else if (retval & AE_WRITABLE) {
- sendReplyToClient(NULL, c->fd, c, AE_WRITABLE);
- }
- }
- return REDIS_OK;
-}
-
static int syncWrite(int fd, char *ptr, ssize_t size, int timeout) {
ssize_t nwritten, ret = size;
time_t start = time(NULL);
}
static void syncCommand(redisClient *c) {
- struct stat sb;
- int fd = -1, len;
- time_t start = time(NULL);
- char sizebuf[32];
-
/* ignore SYNC if aleady slave or in monitor mode */
if (c->flags & REDIS_SLAVE) return;
- redisLog(REDIS_NOTICE,"Slave ask for syncronization");
- if (flushClientOutput(c) == REDIS_ERR ||
- rdbSave(server.dbfilename) != REDIS_OK)
- goto closeconn;
-
- fd = open(server.dbfilename, O_RDONLY);
- if (fd == -1 || fstat(fd,&sb) == -1) goto closeconn;
- len = sb.st_size;
-
- snprintf(sizebuf,32,"$%d\r\n",len);
- if (syncWrite(c->fd,sizebuf,strlen(sizebuf),5) == -1) goto closeconn;
- while(len) {
- char buf[1024];
- int nread;
+ /* SYNC can't be issued when the server has pending data to send to
+ * the client about already issued commands. We need a fresh reply
+ * buffer registering the differences between the BGSAVE and the current
+ * dataset, so that we can copy to other slaves if needed. */
+ if (listLength(c->reply) != 0) {
+ addReplySds(c,sdsnew("-ERR SYNC is invalid with pending input\r\n"));
+ return;
+ }
- if (time(NULL)-start > REDIS_MAX_SYNC_TIME) goto closeconn;
- nread = read(fd,buf,1024);
- if (nread == -1) goto closeconn;
- len -= nread;
- if (syncWrite(c->fd,buf,nread,5) == -1) goto closeconn;
+ redisLog(REDIS_NOTICE,"Slave ask for synchronization");
+ /* Here we need to check if there is a background saving operation
+ * in progress, or if it is required to start one */
+ if (server.bgsaveinprogress) {
+ /* Ok a background save is in progress. Let's check if it is a good
+ * one for replication, i.e. if there is another slave that is
+ * registering differences since the server forked to save */
+ redisClient *slave;
+ listNode *ln;
+
+ listRewind(server.slaves);
+ while((ln = listYield(server.slaves))) {
+ slave = ln->value;
+ if (slave->replstate == REDIS_REPL_WAIT_BGSAVE_END) break;
+ }
+ if (ln) {
+ /* Perfect, the server is already registering differences for
+ * another slave. Set the right state, and copy the buffer. */
+ listRelease(c->reply);
+ c->reply = listDup(slave->reply);
+ if (!c->reply) oom("listDup copying slave reply list");
+ c->replstate = REDIS_REPL_WAIT_BGSAVE_END;
+ redisLog(REDIS_NOTICE,"Waiting for end of BGSAVE for SYNC");
+ } else {
+ /* No way, we need to wait for the next BGSAVE in order to
+ * register differences */
+ c->replstate = REDIS_REPL_WAIT_BGSAVE_START;
+ redisLog(REDIS_NOTICE,"Waiting for next BGSAVE for SYNC");
+ }
+ } else {
+ /* Ok we don't have a BGSAVE in progress, let's start one */
+ redisLog(REDIS_NOTICE,"Starting BGSAVE for SYNC");
+ if (rdbSaveBackground(server.dbfilename) != REDIS_OK) {
+ redisLog(REDIS_NOTICE,"Replication failed, can't BGSAVE");
+ addReplySds(c,sdsnew("-ERR Unalbe to perform background save\r\n"));
+ return;
+ }
+ c->replstate = REDIS_REPL_WAIT_BGSAVE_END;
}
- if (syncWrite(c->fd,"\r\n",2,5) == -1) goto closeconn;
- close(fd);
+ c->repldbfd = -1;
c->flags |= REDIS_SLAVE;
c->slaveseldb = 0;
if (!listAddNodeTail(server.slaves,c)) oom("listAddNodeTail");
- redisLog(REDIS_NOTICE,"Syncronization with slave succeeded");
return;
+}
-closeconn:
- if (fd != -1) close(fd);
- c->flags |= REDIS_CLOSE;
- redisLog(REDIS_WARNING,"Syncronization with slave failed");
- return;
+static void sendBulkToSlave(aeEventLoop *el, int fd, void *privdata, int mask) {
+ redisClient *slave = privdata;
+ REDIS_NOTUSED(el);
+ REDIS_NOTUSED(mask);
+ char buf[REDIS_IOBUF_LEN];
+ ssize_t nwritten, buflen;
+
+ if (slave->repldboff == 0) {
+ /* Write the bulk write count before to transfer the DB. In theory here
+ * we don't know how much room there is in the output buffer of the
+ * socket, but in pratice SO_SNDLOWAT (the minimum count for output
+ * operations) will never be smaller than the few bytes we need. */
+ sds bulkcount;
+
+ bulkcount = sdscatprintf(sdsempty(),"$%lld\r\n",(unsigned long long)
+ slave->repldbsize);
+ if (write(fd,bulkcount,sdslen(bulkcount)) != (signed)sdslen(bulkcount))
+ {
+ sdsfree(bulkcount);
+ freeClient(slave);
+ return;
+ }
+ sdsfree(bulkcount);
+ }
+ lseek(slave->repldbfd,slave->repldboff,SEEK_SET);
+ buflen = read(slave->repldbfd,buf,REDIS_IOBUF_LEN);
+ if (buflen <= 0) {
+ redisLog(REDIS_WARNING,"Read error sending DB to slave: %s",
+ (buflen == 0) ? "premature EOF" : strerror(errno));
+ freeClient(slave);
+ return;
+ }
+ if ((nwritten = write(fd,buf,buflen)) == -1) {
+ redisLog(REDIS_DEBUG,"Write error sending DB to slave: %s",
+ strerror(errno));
+ freeClient(slave);
+ return;
+ }
+ slave->repldboff += nwritten;
+ if (slave->repldboff == slave->repldbsize) {
+ close(slave->repldbfd);
+ slave->repldbfd = -1;
+ aeDeleteFileEvent(server.el,slave->fd,AE_WRITABLE);
+ slave->replstate = REDIS_REPL_ONLINE;
+ if (aeCreateFileEvent(server.el, slave->fd, AE_WRITABLE,
+ sendReplyToClient, slave, NULL) == AE_ERR) {
+ freeClient(slave);
+ return;
+ }
+ addReplySds(slave,sdsempty());
+ redisLog(REDIS_NOTICE,"Synchronization with slave succeeded");
+ }
+}
+
+static void updateSalvesWaitingBgsave(int bgsaveerr) {
+ listNode *ln;
+ int startbgsave = 0;
+
+ listRewind(server.slaves);
+ while((ln = listYield(server.slaves))) {
+ redisClient *slave = ln->value;
+
+ if (slave->replstate == REDIS_REPL_WAIT_BGSAVE_START) {
+ startbgsave = 1;
+ slave->replstate = REDIS_REPL_WAIT_BGSAVE_END;
+ } else if (slave->replstate == REDIS_REPL_WAIT_BGSAVE_END) {
+ struct stat buf;
+
+ if (bgsaveerr != REDIS_OK) {
+ freeClient(slave);
+ redisLog(REDIS_WARNING,"SYNC failed. BGSAVE child returned an error");
+ continue;
+ }
+ if ((slave->repldbfd = open(server.dbfilename,O_RDONLY)) == -1 ||
+ fstat(slave->repldbfd,&buf) == -1) {
+ freeClient(slave);
+ redisLog(REDIS_WARNING,"SYNC failed. Can't open/stat DB after BGSAVE: %s", strerror(errno));
+ continue;
+ }
+ slave->repldboff = 0;
+ slave->repldbsize = buf.st_size;
+ slave->replstate = REDIS_REPL_SEND_BULK;
+ aeDeleteFileEvent(server.el,slave->fd,AE_WRITABLE);
+ if (aeCreateFileEvent(server.el, slave->fd, AE_WRITABLE, sendBulkToSlave, slave, NULL) == AE_ERR) {
+ freeClient(slave);
+ continue;
+ }
+ }
+ }
+ if (startbgsave) {
+ if (rdbSaveBackground(server.dbfilename) != REDIS_OK) {
+ listRewind(server.slaves);
+ redisLog(REDIS_WARNING,"SYNC failed. BGSAVE failed");
+ while((ln = listYield(server.slaves))) {
+ redisClient *slave = ln->value;
+
+ if (slave->replstate == REDIS_REPL_WAIT_BGSAVE_START)
+ freeClient(slave);
+ }
+ }
+ }
}
static int syncWithMaster(void) {
projects / redis.git / blobdiff