X-Git-Url: https://git.saurik.com/redis.git/blobdiff_plain/ef37997608dfa5ad457961e0f3e427315ba4be83..e323635c2d9d9039442cd1014932e4dd314d2d06:/src/replication.c

diff --git a/src/replication.c b/src/replication.c
index 8eb36f83..543477b6 100644
--- a/src/replication.c
+++ b/src/replication.c
@@ -145,6 +145,46 @@ void syncCommand(redisClient *c) {
     return;
 }
 
+/* REPLCONF <option> <value> <option> <value> ...
+ * This command is used by a slave in order to configure the replication
+ * process before starting it with the SYNC command.
+ *
+ * Currently the only use of this command is to communicate to the master
+ * what is the listening port of the Slave redis instance, so that the
+ * master can accurately list slaves and their listening ports in
+ * the INFO output.
+ *
+ * In the future the same command can be used in order to configure
+ * the replication to initiate an incremental replication instead of a
+ * full resync. */
+void replconfCommand(redisClient *c) {
+    int j;
+
+    if ((c->argc % 2) == 0) {
+        /* Number of arguments must be odd to make sure that every
+         * option has a corresponding value. */
+        addReply(c,shared.syntaxerr);
+        return;
+    }
+
+    /* Process every option-value pair. */
+    for (j = 1; j < c->argc; j+=2) {
+        if (!strcasecmp(c->argv[j]->ptr,"listening-port")) {
+            long port;
+
+            if ((getLongFromObjectOrReply(c,c->argv[j+1],
+                    &port,NULL) != REDIS_OK))
+                return;
+            c->slave_listening_port = port;
+        } else {
+            addReplyErrorFormat(c,"Unrecognized REPLCONF option: %s",
+                (char*)c->argv[j]->ptr);
+            return;
+        }
+    }
+    addReply(c,shared.ok);
+}
+
 void sendBulkToSlave(aeEventLoop *el, int fd, void *privdata, int mask) {
     redisClient *slave = privdata;
     REDIS_NOTUSED(el);
@@ -271,16 +311,18 @@ void replicationAbortSyncTransfer(void) {
 }
 
 /* Asynchronously read the SYNC payload we receive from a master */
+#define REPL_MAX_WRITTEN_BEFORE_FSYNC (1024*1024*8) /* 8 MB */
 void readSyncBulkPayload(aeEventLoop *el, int fd, void *privdata, int mask) {
     char buf[4096];
     ssize_t nread, readlen;
+    off_t left;
     REDIS_NOTUSED(el);
     REDIS_NOTUSED(privdata);
     REDIS_NOTUSED(mask);
 
-    /* If repl_transfer_left == -1 we still have to read the bulk length
+    /* If repl_transfer_size == -1 we still have to read the bulk length
      * from the master reply. */
-    if (server.repl_transfer_left == -1) {
+    if (server.repl_transfer_size == -1) {
         if (syncReadLine(fd,buf,1024,server.repl_syncio_timeout*1000) == -1) {
             redisLog(REDIS_WARNING,
                 "I/O error reading bulk count from MASTER: %s",
@@ -303,16 +345,16 @@ void readSyncBulkPayload(aeEventLoop *el, int fd, void *privdata, int mask) {
             redisLog(REDIS_WARNING,"Bad protocol from MASTER, the first byte is not '$', are you sure the host and port are right?");
             goto error;
         }
-        server.repl_transfer_left = strtol(buf+1,NULL,10);
+        server.repl_transfer_size = strtol(buf+1,NULL,10);
         redisLog(REDIS_NOTICE,
             "MASTER <-> SLAVE sync: receiving %ld bytes from master",
-            server.repl_transfer_left);
+            server.repl_transfer_size);
         return;
     }
 
     /* Read bulk data */
-    readlen = (server.repl_transfer_left < (signed)sizeof(buf)) ?
-        server.repl_transfer_left : (signed)sizeof(buf);
+    left = server.repl_transfer_size - server.repl_transfer_read;
+    readlen = (left < (signed)sizeof(buf)) ? left : (signed)sizeof(buf);
     nread = read(fd,buf,readlen);
     if (nread <= 0) {
         redisLog(REDIS_WARNING,"I/O error trying to sync with MASTER: %s",
@@ -325,9 +367,23 @@ void readSyncBulkPayload(aeEventLoop *el, int fd, void *privdata, int mask) {
         redisLog(REDIS_WARNING,"Write error or short write writing to the DB dump file needed for MASTER <-> SLAVE synchronization: %s", strerror(errno));
         goto error;
     }
-    server.repl_transfer_left -= nread;
+    server.repl_transfer_read += nread;
+
+    /* Sync data on disk from time to time, otherwise at the end of the transfer
+     * we may suffer a big delay as the memory buffers are copied into the
+     * actual disk. */
+    if (server.repl_transfer_read >=
+        server.repl_transfer_last_fsync_off + REPL_MAX_WRITTEN_BEFORE_FSYNC)
+    {
+        off_t sync_size = server.repl_transfer_read -
+                          server.repl_transfer_last_fsync_off;
+        rdb_fsync_range(server.repl_transfer_fd,
+            server.repl_transfer_last_fsync_off, sync_size);
+        server.repl_transfer_last_fsync_off += sync_size;
+    }
+
     /* Check if the transfer is now complete */
-    if (server.repl_transfer_left == 0) {
+    if (server.repl_transfer_read == server.repl_transfer_size) {
         if (rename(server.repl_transfer_tmpfile,server.rdb_filename) == -1) {
             redisLog(REDIS_WARNING,"Failed trying to rename the temp DB into dump.rdb in MASTER <-> SLAVE synchronization: %s", strerror(errno));
             replicationAbortSyncTransfer();
@@ -378,8 +434,54 @@ error:
     return;
 }
 
+/* Send a synchronous command to the master. Used to send AUTH and
+ * REPLCONF commands before starting the replication with SYNC.
+ *
+ * On success NULL is returned.
+ * On error an sds string describing the error is returned.
+ */
+char *sendSynchronousCommand(int fd, ...) {
+    va_list ap;
+    sds cmd = sdsempty();
+    char *arg, buf[256];
+
+    /* Create the command to send to the master, we use simple inline
+     * protocol for simplicity as currently we only send simple strings. */
+    va_start(ap,fd);
+    while(1) {
+        arg = va_arg(ap, char*);
+        if (arg == NULL) break;
+
+        if (sdslen(cmd) != 0) cmd = sdscatlen(cmd," ",1);
+        cmd = sdscat(cmd,arg);
+    }
+    cmd = sdscatlen(cmd,"\r\n",2);
+
+    /* Transfer command to the server. */
+    if (syncWrite(fd,cmd,sdslen(cmd),server.repl_syncio_timeout*1000) == -1) {
+        sdsfree(cmd);
+        return sdscatprintf(sdsempty(),"Writing to master: %s",
+                strerror(errno));
+    }
+    sdsfree(cmd);
+
+    /* Read the reply from the server. */
+    if (syncReadLine(fd,buf,sizeof(buf),server.repl_syncio_timeout*1000) == -1)
+    {
+        return sdscatprintf(sdsempty(),"Reading from master: %s",
+                strerror(errno));
+    }
+
+    /* Check for errors from the server. */
+    if (buf[0] != '+') {
+        return sdscatprintf(sdsempty(),"Error from master: %s", buf);
+    }
+
+    return NULL; /* No errors. */
+}
+
 void syncWithMaster(aeEventLoop *el, int fd, void *privdata, int mask) {
-    char buf[1024], tmpfile[256];
+    char tmpfile[256], *err;
     int dfd, maxtries = 5;
     REDIS_NOTUSED(el);
     REDIS_NOTUSED(privdata);
@@ -400,24 +502,26 @@ void syncWithMaster(aeEventLoop *el, int fd, void *privdata, int mask) {
 
     /* AUTH with the master if required. */
     if(server.masterauth) {
-        char authcmd[1024];
-        size_t authlen;
-
-        authlen = snprintf(authcmd,sizeof(authcmd),"AUTH %s\r\n",server.masterauth);
-        if (syncWrite(fd,authcmd,authlen,server.repl_syncio_timeout*1000) == -1) {
-            redisLog(REDIS_WARNING,"Unable to AUTH to MASTER: %s",
-                strerror(errno));
-            goto error;
-        }
-        /* Read the AUTH result.  */
-        if (syncReadLine(fd,buf,1024,server.repl_syncio_timeout*1000) == -1) {
-            redisLog(REDIS_WARNING,"I/O error reading auth result from MASTER: %s",
-                strerror(errno));
+        err = sendSynchronousCommand(fd,"AUTH",server.masterauth,NULL);
+        if (err) {
+            redisLog(REDIS_WARNING,"Unable to AUTH to MASTER: %s",err);
+            sdsfree(err);
             goto error;
         }
-        if (buf[0] != '+') {
-            redisLog(REDIS_WARNING,"Cannot AUTH to MASTER, is the masterauth password correct?");
-            goto error;
+    }
+
+    /* Set the slave port, so that Master's INFO command can list the
+     * slave listening port correctly. */
+    {
+        sds port = sdsfromlonglong(server.port);
+        err = sendSynchronousCommand(fd,"REPLCONF","listening-port",port,
+                                         NULL);
+        sdsfree(port);
+        /* Ignore the error if any, not all the Redis versions support
+         * REPLCONF listening-port. */
+        if (err) {
+            redisLog(REDIS_NOTICE,"(non critical): Master does not understand REPLCONF listening-port: %s", err);
+            sdsfree(err);
         }
     }
 
@@ -450,7 +554,9 @@ void syncWithMaster(aeEventLoop *el, int fd, void *privdata, int mask) {
     }
 
     server.repl_state = REDIS_REPL_TRANSFER;
-    server.repl_transfer_left = -1;
+    server.repl_transfer_size = -1;
+    server.repl_transfer_read = 0;
+    server.repl_transfer_last_fsync_off = 0;
     server.repl_transfer_fd = dfd;
     server.repl_transfer_lastio = server.unixtime;
     server.repl_transfer_tmpfile = zstrdup(tmpfile);