Add ziplist test for deleting next to last entries

[redis.git] / src / replication.c

diff --git a/src/replication.c b/src/replication.c
index dd42fcf17f4e39a3857f81cd482403eafd3ee73b..45a223b8e36d4dbc333d83122052442ce22c747d 100644 (file)
--- a/src/replication.c
+++ b/src/replication.c
@@ -25,24 +25,15 @@ void replicationFeedSlaves(list *slaves, int dictid, robj **argv, int argc) {
         if (slave->slaveseldb != dictid) {
             robj *selectcmd;
 
         if (slave->slaveseldb != dictid) {
             robj *selectcmd;
 

-            switch(dictid) {
-            case 0: selectcmd = shared.select0; break;
-            case 1: selectcmd = shared.select1; break;
-            case 2: selectcmd = shared.select2; break;
-            case 3: selectcmd = shared.select3; break;
-            case 4: selectcmd = shared.select4; break;
-            case 5: selectcmd = shared.select5; break;
-            case 6: selectcmd = shared.select6; break;
-            case 7: selectcmd = shared.select7; break;
-            case 8: selectcmd = shared.select8; break;
-            case 9: selectcmd = shared.select9; break;
-            default:
+            if (dictid >= 0 && dictid < REDIS_SHARED_SELECT_CMDS) {
+                selectcmd = shared.select[dictid];
+                incrRefCount(selectcmd);
+            } else {

                 selectcmd = createObject(REDIS_STRING,
                     sdscatprintf(sdsempty(),"select %d\r\n",dictid));
                 selectcmd = createObject(REDIS_STRING,
                     sdscatprintf(sdsempty(),"select %d\r\n",dictid));

-                selectcmd->refcount = 0;
-                break;

             }
             addReply(slave,selectcmd);
             }
             addReply(slave,selectcmd);

+            decrRefCount(selectcmd);

             slave->slaveseldb = dictid;
         }
         addReplyMultiBulkLen(slave,argc);
             slave->slaveseldb = dictid;
         }
         addReplyMultiBulkLen(slave,argc);


@@ -50,17 +41,23 @@ void replicationFeedSlaves(list *slaves, int dictid, robj **argv, int argc) {
     }
 }
 
     }
 }
 

-void replicationFeedMonitors(list *monitors, int dictid, robj **argv, int argc) {
+void replicationFeedMonitors(redisClient *c, list *monitors, int dictid, robj **argv, int argc) {

     listNode *ln;
     listIter li;
     listNode *ln;
     listIter li;

-    int j;
+    int j, port;

     sds cmdrepr = sdsnew("+");
     robj *cmdobj;
     sds cmdrepr = sdsnew("+");
     robj *cmdobj;

+    char ip[32];

     struct timeval tv;
 
     gettimeofday(&tv,NULL);
     cmdrepr = sdscatprintf(cmdrepr,"%ld.%06ld ",(long)tv.tv_sec,(long)tv.tv_usec);
     struct timeval tv;
 
     gettimeofday(&tv,NULL);
     cmdrepr = sdscatprintf(cmdrepr,"%ld.%06ld ",(long)tv.tv_sec,(long)tv.tv_usec);

-    if (dictid != 0) cmdrepr = sdscatprintf(cmdrepr,"(db %d) ", dictid);
+    if (c->flags & REDIS_LUA_CLIENT) {
+        cmdrepr = sdscatprintf(cmdrepr,"[%d lua] ", dictid);
+    } else {
+        anetPeerToString(c->fd,ip,&port);
+        cmdrepr = sdscatprintf(cmdrepr,"[%d %s:%d] ", dictid,ip,port);
+    }

 
     for (j = 0; j < argc; j++) {
         if (argv[j]->encoding == REDIS_ENCODING_INT) {
 
     for (j = 0; j < argc; j++) {
         if (argv[j]->encoding == REDIS_ENCODING_INT) {


@@ -148,6 +145,46 @@ void syncCommand(redisClient *c) {
     return;
 }
 
     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);
 void sendBulkToSlave(aeEventLoop *el, int fd, void *privdata, int mask) {
     redisClient *slave = privdata;
     REDIS_NOTUSED(el);


@@ -197,7 +234,6 @@ void sendBulkToSlave(aeEventLoop *el, int fd, void *privdata, int mask) {
             freeClient(slave);
             return;
         }
             freeClient(slave);
             return;
         }

-        addReplySds(slave,sdsempty());

         redisLog(REDIS_NOTICE,"Synchronization with slave succeeded");
     }
 }
         redisLog(REDIS_NOTICE,"Synchronization with slave succeeded");
     }
 }


@@ -285,7 +321,7 @@ void readSyncBulkPayload(aeEventLoop *el, int fd, void *privdata, int 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 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) {
+        if (syncReadLine(fd,buf,1024,server.repl_syncio_timeout*1000) == -1) {

             redisLog(REDIS_WARNING,
                 "I/O error reading bulk count from MASTER: %s",
                 strerror(errno));
             redisLog(REDIS_WARNING,
                 "I/O error reading bulk count from MASTER: %s",
                 strerror(errno));


@@ -301,7 +337,7 @@ void readSyncBulkPayload(aeEventLoop *el, int fd, void *privdata, int mask) {
             /* At this stage just a newline works as a PING in order to take
              * the connection live. So we refresh our last interaction
              * timestamp. */
             /* 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);
+            server.repl_transfer_lastio = server.unixtime;

             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?");
             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?");


@@ -324,9 +360,9 @@ void readSyncBulkPayload(aeEventLoop *el, int fd, void *privdata, int mask) {
         replicationAbortSyncTransfer();
         return;
     }
         replicationAbortSyncTransfer();
         return;
     }

-    server.repl_transfer_lastio = time(NULL);
+    server.repl_transfer_lastio = server.unixtime;

     if (write(server.repl_transfer_fd,buf,nread) != nread) {
     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));
+        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;
         goto error;
     }
     server.repl_transfer_left -= nread;


@@ -382,8 +418,54 @@ error:
     return;
 }
 
     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) {
 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);
     int dfd, maxtries = 5;
     REDIS_NOTUSED(el);
     REDIS_NOTUSED(privdata);


@@ -404,29 +486,31 @@ void syncWithMaster(aeEventLoop *el, int fd, void *privdata, int mask) {
 
     /* AUTH with the master if required. */
     if(server.masterauth) {
 
     /* 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) == -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) == -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;
         }
             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);

         }
     }
 
     /* Issue the SYNC command */
         }
     }
 
     /* Issue the SYNC command */

-    if (syncWrite(fd,"SYNC \r\n",7,server.repl_syncio_timeout) == -1) {
+    if (syncWrite(fd,"SYNC\r\n",6,server.repl_syncio_timeout*1000) == -1) {

         redisLog(REDIS_WARNING,"I/O error writing to MASTER: %s",
             strerror(errno));
         goto error;
         redisLog(REDIS_WARNING,"I/O error writing to MASTER: %s",
             strerror(errno));
         goto error;


@@ -435,7 +519,7 @@ void syncWithMaster(aeEventLoop *el, int fd, void *privdata, int mask) {
     /* Prepare a suitable temp file for bulk transfer */
     while(maxtries--) {
         snprintf(tmpfile,256,
     /* Prepare a suitable temp file for bulk transfer */
     while(maxtries--) {
         snprintf(tmpfile,256,

-            "temp-%d.%ld.rdb",(int)time(NULL),(long int)getpid());
+            "temp-%d.%ld.rdb",(int)server.unixtime,(long int)getpid());

         dfd = open(tmpfile,O_CREAT|O_WRONLY|O_EXCL,0644);
         if (dfd != -1) break;
         sleep(1);
         dfd = open(tmpfile,O_CREAT|O_WRONLY|O_EXCL,0644);
         if (dfd != -1) break;
         sleep(1);


@@ -456,7 +540,7 @@ void syncWithMaster(aeEventLoop *el, int fd, void *privdata, int mask) {
     server.repl_state = REDIS_REPL_TRANSFER;
     server.repl_transfer_left = -1;
     server.repl_transfer_fd = dfd;
     server.repl_state = REDIS_REPL_TRANSFER;
     server.repl_transfer_left = -1;
     server.repl_transfer_fd = dfd;

-    server.repl_transfer_lastio = time(NULL);
+    server.repl_transfer_lastio = server.unixtime;

     server.repl_transfer_tmpfile = zstrdup(tmpfile);
     return;
 
     server.repl_transfer_tmpfile = zstrdup(tmpfile);
     return;
 


@@ -484,7 +568,7 @@ int connectWithMaster(void) {
         return REDIS_ERR;
     }
 
         return REDIS_ERR;
     }
 

-    server.repl_transfer_lastio = time(NULL);
+    server.repl_transfer_lastio = server.unixtime;

     server.repl_transfer_s = fd;
     server.repl_state = REDIS_REPL_CONNECTING;
     return REDIS_OK;
     server.repl_transfer_s = fd;
     server.repl_state = REDIS_REPL_CONNECTING;
     return REDIS_OK;


@@ -535,6 +619,7 @@ void slaveofCommand(redisClient *c) {
         server.masterhost = sdsdup(c->argv[1]->ptr);
         server.masterport = port;
         if (server.master) freeClient(server.master);
         server.masterhost = sdsdup(c->argv[1]->ptr);
         server.masterport = port;
         if (server.master) freeClient(server.master);

+        disconnectSlaves(); /* Force our slaves to resync with us as well. */

         if (server.repl_state == REDIS_REPL_TRANSFER)
             replicationAbortSyncTransfer();
         server.repl_state = REDIS_REPL_CONNECT;
         if (server.repl_state == REDIS_REPL_TRANSFER)
             replicationAbortSyncTransfer();
         server.repl_state = REDIS_REPL_CONNECT;


@@ -596,7 +681,7 @@ void replicationCron(void) {
             if (slave->replstate == REDIS_REPL_SEND_BULK) continue;
             if (slave->replstate == REDIS_REPL_ONLINE) {
                 /* If the slave is online send a normal ping */
             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"));
+                addReplySds(slave,sdsnew("*1\r\n$4\r\nPING\r\n"));

             } else {
                 /* Otherwise we are in the pre-synchronization stage.
                  * Just a newline will do the work of refreshing the
             } else {
                 /* Otherwise we are in the pre-synchronization stage.
                  * Just a newline will do the work of refreshing the