X-Git-Url: https://git.saurik.com/redis.git/blobdiff_plain/acc75bfd4f1607625876d74d6447efaaf505db59..52d46855d90bfddb3e03d957c48101d123e4dfe7:/src/replication.c

diff --git a/src/replication.c b/src/replication.c
index 363ce54a..712af110 100644
--- a/src/replication.c
+++ b/src/replication.c
@@ -5,41 +5,13 @@
 #include <fcntl.h>
 #include <sys/stat.h>
 
+/* ---------------------------------- MASTER -------------------------------- */
+
 void replicationFeedSlaves(list *slaves, int dictid, robj **argv, int argc) {
     listNode *ln;
     listIter li;
-    int outc = 0, j;
-    robj **outv;
-    /* We need 1+(ARGS*3) objects since commands are using the new protocol
-     * and we one 1 object for the first "*<count>\r\n" multibulk count, then
-     * for every additional object we have "$<count>\r\n" + object + "\r\n". */
-    robj *static_outv[REDIS_STATIC_ARGS*3+1];
-    robj *lenobj;
-
-    if (argc <= REDIS_STATIC_ARGS) {
-        outv = static_outv;
-    } else {
-        outv = zmalloc(sizeof(robj*)*(argc*3+1));
-    }
-
-    lenobj = createObject(REDIS_STRING,
-            sdscatprintf(sdsempty(), "*%d\r\n", argc));
-    lenobj->refcount = 0;
-    outv[outc++] = lenobj;
-    for (j = 0; j < argc; j++) {
-        lenobj = createObject(REDIS_STRING,
-            sdscatprintf(sdsempty(),"$%lu\r\n",
-                (unsigned long) stringObjectLen(argv[j])));
-        lenobj->refcount = 0;
-        outv[outc++] = lenobj;
-        outv[outc++] = argv[j];
-        outv[outc++] = shared.crlf;
-    }
+    int j;
 
-    /* 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,&li);
     while((ln = listNext(&li))) {
         redisClient *slave = ln->value;
@@ -47,7 +19,9 @@ void replicationFeedSlaves(list *slaves, int dictid, robj **argv, int argc) {
         /* 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 */
+        /* Feed slaves that are waiting for the initial SYNC (so these commands
+         * are queued in the output buffer until the intial SYNC completes),
+         * or are already in sync with the master. */
         if (slave->slaveseldb != dictid) {
             robj *selectcmd;
 
@@ -71,10 +45,9 @@ void replicationFeedSlaves(list *slaves, int dictid, robj **argv, int argc) {
             addReply(slave,selectcmd);
             slave->slaveseldb = dictid;
         }
-        for (j = 0; j < outc; j++) addReply(slave,outv[j]);
+        addReplyMultiBulkLen(slave,argc);
+        for (j = 0; j < argc; j++) addReplyBulk(slave,argv[j]);
     }
-    for (j = 0; j < outc; j++) decrRefCount(outv[j]);
-    if (outv != static_outv) zfree(outv);
 }
 
 void replicationFeedMonitors(list *monitors, int dictid, robj **argv, int argc) {
@@ -86,7 +59,7 @@ void replicationFeedMonitors(list *monitors, int dictid, robj **argv, int argc)
     struct timeval tv;
 
     gettimeofday(&tv,NULL);
-    cmdrepr = sdscatprintf(cmdrepr,"%ld.%ld ",(long)tv.tv_sec,(long)tv.tv_usec);
+    cmdrepr = sdscatprintf(cmdrepr,"%ld.%06ld ",(long)tv.tv_sec,(long)tv.tv_usec);
     if (dictid != 0) cmdrepr = sdscatprintf(cmdrepr,"(db %d) ", dictid);
 
     for (j = 0; j < argc; j++) {
@@ -110,68 +83,6 @@ void replicationFeedMonitors(list *monitors, int dictid, robj **argv, int argc)
     decrRefCount(cmdobj);
 }
 
-int syncWrite(int fd, char *ptr, ssize_t size, int timeout) {
-    ssize_t nwritten, ret = size;
-    time_t start = time(NULL);
-
-    timeout++;
-    while(size) {
-        if (aeWait(fd,AE_WRITABLE,1000) & AE_WRITABLE) {
-            nwritten = write(fd,ptr,size);
-            if (nwritten == -1) return -1;
-            ptr += nwritten;
-            size -= nwritten;
-        }
-        if ((time(NULL)-start) > timeout) {
-            errno = ETIMEDOUT;
-            return -1;
-        }
-    }
-    return ret;
-}
-
-int syncRead(int fd, char *ptr, ssize_t size, int timeout) {
-    ssize_t nread, totread = 0;
-    time_t start = time(NULL);
-
-    timeout++;
-    while(size) {
-        if (aeWait(fd,AE_READABLE,1000) & AE_READABLE) {
-            nread = read(fd,ptr,size);
-            if (nread == -1) return -1;
-            ptr += nread;
-            size -= nread;
-            totread += nread;
-        }
-        if ((time(NULL)-start) > timeout) {
-            errno = ETIMEDOUT;
-            return -1;
-        }
-    }
-    return totread;
-}
-
-int syncReadLine(int fd, char *ptr, ssize_t size, int timeout) {
-    ssize_t nread = 0;
-
-    size--;
-    while(size) {
-        char c;
-
-        if (syncRead(fd,&c,1,timeout) == -1) return -1;
-        if (c == '\n') {
-            *ptr = '\0';
-            if (nread && *(ptr-1) == '\r') *(ptr-1) = '\0';
-            return nread;
-        } else {
-            *ptr++ = c;
-            *ptr = '\0';
-            nread++;
-        }
-    }
-    return nread;
-}
-
 void syncCommand(redisClient *c) {
     /* ignore SYNC if aleady slave or in monitor mode */
     if (c->flags & REDIS_SLAVE) return;
@@ -179,7 +90,7 @@ void syncCommand(redisClient *c) {
     /* Refuse SYNC requests if we are a slave but the link with our master
      * is not ok... */
     if (server.masterhost && server.replstate != REDIS_REPL_CONNECTED) {
-        addReplySds(c,sdsnew("-ERR Can't SYNC while not connected with my master\r\n"));
+        addReplyError(c,"Can't SYNC while not connected with my master");
         return;
     }
 
@@ -188,7 +99,7 @@ void syncCommand(redisClient *c) {
      * 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"));
+        addReplyError(c,"SYNC is invalid with pending input");
         return;
     }
 
@@ -226,7 +137,7 @@ void syncCommand(redisClient *c) {
         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"));
+            addReplyError(c,"Unable to perform background save");
             return;
         }
         c->replstate = REDIS_REPL_WAIT_BGSAVE_END;
@@ -350,68 +261,165 @@ void updateSlavesWaitingBgsave(int bgsaveerr) {
     }
 }
 
-int syncWithMaster(void) {
-    char buf[1024], tmpfile[256], authcmd[1024];
-    long dumpsize;
-    int fd = anetTcpConnect(NULL,server.masterhost,server.masterport);
+/* ----------------------------------- SLAVE -------------------------------- */
+
+/* Abort the async download of the bulk dataset while SYNC-ing with master */
+void replicationAbortSyncTransfer(void) {
+    redisAssert(server.replstate == REDIS_REPL_TRANSFER);
+
+    aeDeleteFileEvent(server.el,server.repl_transfer_s,AE_READABLE);
+    close(server.repl_transfer_s);
+    close(server.repl_transfer_fd);
+    unlink(server.repl_transfer_tmpfile);
+    zfree(server.repl_transfer_tmpfile);
+    server.replstate = REDIS_REPL_CONNECT;
+}
+
+/* Asynchronously read the SYNC payload we receive from a master */
+void readSyncBulkPayload(aeEventLoop *el, int fd, void *privdata, int mask) {
+    char buf[4096];
+    ssize_t nread, readlen;
+    REDIS_NOTUSED(el);
+    REDIS_NOTUSED(privdata);
+    REDIS_NOTUSED(mask);
+
+    /* If repl_transfer_left == -1 we still have to read the bulk length
+     * from the master reply. */
+    if (server.repl_transfer_left == -1) {
+        if (syncReadLine(fd,buf,1024,server.repl_syncio_timeout) == -1) {
+            redisLog(REDIS_WARNING,
+                "I/O error reading bulk count from MASTER: %s",
+                strerror(errno));
+            goto error;
+        }
+
+        if (buf[0] == '-') {
+            redisLog(REDIS_WARNING,
+                "MASTER aborted replication with an error: %s",
+                buf+1);
+            goto error;
+        } else if (buf[0] == '\0') {
+            /* At this stage just a newline works as a PING in order to take
+             * the connection live. So we refresh our last interaction
+             * timestamp. */
+            server.repl_transfer_lastio = time(NULL);
+            return;
+        } else if (buf[0] != '$') {
+            redisLog(REDIS_WARNING,"Bad protocol from MASTER, the first byte is not '$', are you sure the host and port are right?");
+            goto error;
+        }
+        server.repl_transfer_left = strtol(buf+1,NULL,10);
+        redisLog(REDIS_NOTICE,
+            "MASTER <-> SLAVE sync: receiving %ld bytes from master",
+            server.repl_transfer_left);
+        return;
+    }
+
+    /* Read bulk data */
+    readlen = (server.repl_transfer_left < (signed)sizeof(buf)) ?
+        server.repl_transfer_left : (signed)sizeof(buf);
+    nread = read(fd,buf,readlen);
+    if (nread <= 0) {
+        redisLog(REDIS_WARNING,"I/O error trying to sync with MASTER: %s",
+            (nread == -1) ? strerror(errno) : "connection lost");
+        replicationAbortSyncTransfer();
+        return;
+    }
+    server.repl_transfer_lastio = time(NULL);
+    if (write(server.repl_transfer_fd,buf,nread) != nread) {
+        redisLog(REDIS_WARNING,"Write error or short write writing to the DB dump file needed for MASTER <-> SLAVE synchrnonization: %s", strerror(errno));
+        goto error;
+    }
+    server.repl_transfer_left -= nread;
+    /* Check if the transfer is now complete */
+    if (server.repl_transfer_left == 0) {
+        if (rename(server.repl_transfer_tmpfile,server.dbfilename) == -1) {
+            redisLog(REDIS_WARNING,"Failed trying to rename the temp DB into dump.rdb in MASTER <-> SLAVE synchronization: %s", strerror(errno));
+            replicationAbortSyncTransfer();
+            return;
+        }
+        redisLog(REDIS_NOTICE, "MASTER <-> SLAVE sync: Loading DB in memory");
+        emptyDb();
+        /* Before loading the DB into memory we need to delete the readable
+         * handler, otherwise it will get called recursively since
+         * rdbLoad() will call the event loop to process events from time to
+         * time for non blocking loading. */
+        aeDeleteFileEvent(server.el,server.repl_transfer_s,AE_READABLE);
+        if (rdbLoad(server.dbfilename) != REDIS_OK) {
+            redisLog(REDIS_WARNING,"Failed trying to load the MASTER synchronization DB from disk");
+            replicationAbortSyncTransfer();
+            return;
+        }
+        /* Final setup of the connected slave <- master link */
+        zfree(server.repl_transfer_tmpfile);
+        close(server.repl_transfer_fd);
+        server.master = createClient(server.repl_transfer_s);
+        server.master->flags |= REDIS_MASTER;
+        server.master->authenticated = 1;
+        server.replstate = REDIS_REPL_CONNECTED;
+        redisLog(REDIS_NOTICE, "MASTER <-> SLAVE sync: Finished with success");
+        /* Rewrite the AOF file now that the dataset changed. */
+        if (server.appendonly) rewriteAppendOnlyFileBackground();
+    }
+
+    return;
+
+error:
+    replicationAbortSyncTransfer();
+    return;
+}
+
+void syncWithMaster(aeEventLoop *el, int fd, void *privdata, int mask) {
+    char buf[1024], tmpfile[256];
     int dfd, maxtries = 5;
+    REDIS_NOTUSED(el);
+    REDIS_NOTUSED(privdata);
+    REDIS_NOTUSED(mask);
 
-    if (fd == -1) {
-        redisLog(REDIS_WARNING,"Unable to connect to MASTER: %s",
-            strerror(errno));
-        return REDIS_ERR;
+    /* If this event fired after the user turned the instance into a master
+     * with SLAVEOF NO ONE we must just return ASAP. */
+    if (server.replstate == REDIS_REPL_NONE) {
+        close(fd);
+        return;
     }
 
+    redisLog(REDIS_NOTICE,"Non blocking connect for SYNC fired the event.");
+    /* This event should only be triggered once since it is used to have a
+     * non-blocking connect(2) to the master. It has been triggered when this
+     * function is called, so we can delete it. */
+    aeDeleteFileEvent(server.el,fd,AE_READABLE|AE_WRITABLE);
+
     /* AUTH with the master if required. */
     if(server.masterauth) {
-    	snprintf(authcmd, 1024, "AUTH %s\r\n", server.masterauth);
-    	if (syncWrite(fd, authcmd, strlen(server.masterauth)+7, 5) == -1) {
-            close(fd);
+        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) == -1) {
             redisLog(REDIS_WARNING,"Unable to AUTH to MASTER: %s",
                 strerror(errno));
-            return REDIS_ERR;
-    	}
+            goto error;
+        }
         /* Read the AUTH result.  */
-        if (syncReadLine(fd,buf,1024,3600) == -1) {
-            close(fd);
+        if (syncReadLine(fd,buf,1024,server.repl_syncio_timeout) == -1) {
             redisLog(REDIS_WARNING,"I/O error reading auth result from MASTER: %s",
                 strerror(errno));
-            return REDIS_ERR;
+            goto error;
         }
         if (buf[0] != '+') {
-            close(fd);
             redisLog(REDIS_WARNING,"Cannot AUTH to MASTER, is the masterauth password correct?");
-            return REDIS_ERR;
+            goto error;
         }
     }
 
     /* Issue the SYNC command */
-    if (syncWrite(fd,"SYNC \r\n",7,5) == -1) {
-        close(fd);
+    if (syncWrite(fd,"SYNC \r\n",7,server.repl_syncio_timeout) == -1) {
         redisLog(REDIS_WARNING,"I/O error writing to MASTER: %s",
             strerror(errno));
-        return REDIS_ERR;
-    }
-    /* Read the bulk write count */
-    if (syncReadLine(fd,buf,1024,3600) == -1) {
-        close(fd);
-        redisLog(REDIS_WARNING,"I/O error reading bulk count from MASTER: %s",
-            strerror(errno));
-        return REDIS_ERR;
+        goto error;
     }
-    if (buf[0] == '-') {
-        close(fd);
-        redisLog(REDIS_WARNING,"MASTER aborted replication with an error: %s",
-            buf+1);
-        return REDIS_ERR;
-    } else if (buf[0] != '$') {
-        close(fd);
-        redisLog(REDIS_WARNING,"Bad protocol from MASTER, the first byte is not '$', are you sure the host and port are right?");
-        return REDIS_ERR;
-    }
-    dumpsize = strtol(buf+1,NULL,10);
-    redisLog(REDIS_NOTICE,"Receiving %ld bytes data dump from MASTER",dumpsize);
-    /* Read the bulk write data on a temp file */
+
+    /* Prepare a suitable temp file for bulk transfer */
     while(maxtries--) {
         snprintf(tmpfile,256,
             "temp-%d.%ld.rdb",(int)time(NULL),(long int)getpid());
@@ -420,47 +428,51 @@ int syncWithMaster(void) {
         sleep(1);
     }
     if (dfd == -1) {
-        close(fd);
         redisLog(REDIS_WARNING,"Opening the temp file needed for MASTER <-> SLAVE synchronization: %s",strerror(errno));
-        return REDIS_ERR;
+        goto error;
     }
-    while(dumpsize) {
-        int nread, nwritten;
-
-        nread = read(fd,buf,(dumpsize < 1024)?dumpsize:1024);
-        if (nread <= 0) {
-            redisLog(REDIS_WARNING,"I/O error trying to sync with MASTER: %s",
-                (nread == -1) ? strerror(errno) : "connection lost");
-            close(fd);
-            close(dfd);
-            return REDIS_ERR;
-        }
-        nwritten = write(dfd,buf,nread);
-        if (nwritten == -1) {
-            redisLog(REDIS_WARNING,"Write error writing to the DB dump file needed for MASTER <-> SLAVE synchrnonization: %s", strerror(errno));
-            close(fd);
-            close(dfd);
-            return REDIS_ERR;
-        }
-        dumpsize -= nread;
+
+    /* Setup the non blocking download of the bulk file. */
+    if (aeCreateFileEvent(server.el,fd, AE_READABLE,readSyncBulkPayload,NULL)
+            == AE_ERR)
+    {
+        redisLog(REDIS_WARNING,"Can't create readable event for SYNC");
+        goto error;
     }
-    close(dfd);
-    if (rename(tmpfile,server.dbfilename) == -1) {
-        redisLog(REDIS_WARNING,"Failed trying to rename the temp DB into dump.rdb in MASTER <-> SLAVE synchronization: %s", strerror(errno));
-        unlink(tmpfile);
-        close(fd);
+
+    server.replstate = REDIS_REPL_TRANSFER;
+    server.repl_transfer_left = -1;
+    server.repl_transfer_fd = dfd;
+    server.repl_transfer_lastio = time(NULL);
+    server.repl_transfer_tmpfile = zstrdup(tmpfile);
+    return;
+
+error:
+    server.replstate = REDIS_REPL_CONNECT;
+    close(fd);
+    return;
+}
+
+int connectWithMaster(void) {
+    int fd;
+
+    fd = anetTcpNonBlockConnect(NULL,server.masterhost,server.masterport);
+    if (fd == -1) {
+        redisLog(REDIS_WARNING,"Unable to connect to MASTER: %s",
+            strerror(errno));
         return REDIS_ERR;
     }
-    emptyDb();
-    if (rdbLoad(server.dbfilename) != REDIS_OK) {
-        redisLog(REDIS_WARNING,"Failed trying to load the MASTER synchronization DB from disk");
+
+    if (aeCreateFileEvent(server.el,fd,AE_READABLE|AE_WRITABLE,syncWithMaster,NULL) ==
+            AE_ERR)
+    {
         close(fd);
+        redisLog(REDIS_WARNING,"Can't create readable event for SYNC");
         return REDIS_ERR;
     }
-    server.master = createClient(fd);
-    server.master->flags |= REDIS_MASTER;
-    server.master->authenticated = 1;
-    server.replstate = REDIS_REPL_CONNECTED;
+
+    server.repl_transfer_s = fd;
+    server.replstate = REDIS_REPL_CONNECTING;
     return REDIS_OK;
 }
 
@@ -471,6 +483,8 @@ void slaveofCommand(redisClient *c) {
             sdsfree(server.masterhost);
             server.masterhost = NULL;
             if (server.master) freeClient(server.master);
+            if (server.replstate == REDIS_REPL_TRANSFER)
+                replicationAbortSyncTransfer();
             server.replstate = REDIS_REPL_NONE;
             redisLog(REDIS_NOTICE,"MASTER MODE enabled (user request)");
         }
@@ -479,9 +493,70 @@ void slaveofCommand(redisClient *c) {
         server.masterhost = sdsdup(c->argv[1]->ptr);
         server.masterport = atoi(c->argv[2]->ptr);
         if (server.master) freeClient(server.master);
+        if (server.replstate == REDIS_REPL_TRANSFER)
+            replicationAbortSyncTransfer();
         server.replstate = REDIS_REPL_CONNECT;
         redisLog(REDIS_NOTICE,"SLAVE OF %s:%d enabled (user request)",
             server.masterhost, server.masterport);
     }
     addReply(c,shared.ok);
 }
+
+/* --------------------------- REPLICATION CRON  ---------------------------- */
+
+void replicationCron(void) {
+    /* Bulk transfer I/O timeout? */
+    if (server.masterhost && server.replstate == REDIS_REPL_TRANSFER &&
+        (time(NULL)-server.repl_transfer_lastio) > server.repl_timeout)
+    {
+        redisLog(REDIS_WARNING,"Timeout receiving bulk data from MASTER...");
+        replicationAbortSyncTransfer();
+    }
+
+    /* Timed out master when we are an already connected slave? */
+    if (server.masterhost && server.replstate == REDIS_REPL_CONNECTED &&
+        (time(NULL)-server.master->lastinteraction) > server.repl_timeout)
+    {
+        redisLog(REDIS_WARNING,"MASTER time out: no data nor PING received...");
+        freeClient(server.master);
+    }
+
+    /* Check if we should connect to a MASTER */
+    if (server.replstate == REDIS_REPL_CONNECT) {
+        redisLog(REDIS_NOTICE,"Connecting to MASTER...");
+        if (connectWithMaster() == REDIS_OK) {
+            redisLog(REDIS_NOTICE,"MASTER <-> SLAVE sync started");
+        }
+    }
+    
+    /* If we have attached slaves, PING them from time to time.
+     * So slaves can implement an explicit timeout to masters, and will
+     * be able to detect a link disconnection even if the TCP connection
+     * will not actually go down. */
+    if (!(server.cronloops % (server.repl_ping_slave_period*10))) {
+        listIter li;
+        listNode *ln;
+
+        listRewind(server.slaves,&li);
+        while((ln = listNext(&li))) {
+            redisClient *slave = ln->value;
+
+            /* Don't ping slaves that are in the middle of a bulk transfer
+             * with the master for first synchronization. */
+            if (slave->replstate == REDIS_REPL_SEND_BULK) continue;
+            if (slave->replstate == REDIS_REPL_ONLINE) {
+                /* If the slave is online send a normal ping */
+                addReplySds(slave,sdsnew("PING\r\n"));
+            } else {
+                /* Otherwise we are in the pre-synchronization stage.
+                 * Just a newline will do the work of refreshing the
+                 * connection last interaction time, and at the same time
+                 * we'll be sure that being a single char there are no
+                 * short-write problems. */
+                if (write(slave->fd, "\n", 1) == -1) {
+                    /* Don't worry, it's just a ping. */
+                }
+            }
+        }
+    }
+}