use server.unixtime instead of time(NULL) where possible (cluster.c not checked though)

[redis.git] / src / aof.c

diff --git a/src/aof.c b/src/aof.c
index 167134818ac3292c8095a205378fa9e15bf110da..3f7cd10ffb87b1da7e4d0039b55c5d3a10c1dca3 100644 (file)
--- a/src/aof.c
+++ b/src/aof.c
@@ -1,4 +1,6 @@
 #include "redis.h"
 #include "redis.h"

+#include "bio.h"
+#include "rio.h"

 
 #include <signal.h>
 #include <fcntl.h>
 
 #include <signal.h>
 #include <fcntl.h>


@@ -8,45 +10,57 @@
 #include <sys/resource.h>
 #include <sys/wait.h>
 
 #include <sys/resource.h>
 #include <sys/wait.h>
 

+void aofUpdateCurrentSize(void);
+
+void aof_background_fsync(int fd) {
+    bioCreateBackgroundJob(REDIS_BIO_AOF_FSYNC,(void*)(long)fd,NULL,NULL);
+}
+

 /* Called when the user switches from "appendonly yes" to "appendonly no"
  * at runtime using the CONFIG command. */
 void stopAppendOnly(void) {
 /* Called when the user switches from "appendonly yes" to "appendonly no"
  * at runtime using the CONFIG command. */
 void stopAppendOnly(void) {

-    flushAppendOnlyFile();
-    aof_fsync(server.appendfd);
-    close(server.appendfd);
-
-    server.appendfd = -1;
-    server.appendseldb = -1;
-    server.appendonly = 0;
+    redisAssert(server.aof_state != REDIS_AOF_OFF);
+    flushAppendOnlyFile(1);
+    aof_fsync(server.aof_fd);
+    close(server.aof_fd);
+
+    server.aof_fd = -1;
+    server.aof_selected_db = -1;
+    server.aof_state = REDIS_AOF_OFF;

     /* rewrite operation in progress? kill it, wait child exit */
     /* rewrite operation in progress? kill it, wait child exit */

-    if (server.bgsavechildpid != -1) {
+    if (server.aof_child_pid != -1) {

         int statloc;
 
         int statloc;
 

-        if (kill(server.bgsavechildpid,SIGKILL) != -1)
+        redisLog(REDIS_NOTICE,"Killing running AOF rewrite child: %ld",
+            (long) server.aof_child_pid);
+        if (kill(server.aof_child_pid,SIGKILL) != -1)

             wait3(&statloc,0,NULL);
         /* reset the buffer accumulating changes while the child saves */
             wait3(&statloc,0,NULL);
         /* reset the buffer accumulating changes while the child saves */

-        sdsfree(server.bgrewritebuf);
-        server.bgrewritebuf = sdsempty();
-        server.bgsavechildpid = -1;
+        sdsfree(server.aof_rewrite_buf);
+        server.aof_rewrite_buf = sdsempty();
+        aofRemoveTempFile(server.aof_child_pid);
+        server.aof_child_pid = -1;

     }
 }
 
 /* Called when the user switches from "appendonly no" to "appendonly yes"
  * at runtime using the CONFIG command. */
 int startAppendOnly(void) {
     }
 }
 
 /* Called when the user switches from "appendonly no" to "appendonly yes"
  * at runtime using the CONFIG command. */
 int startAppendOnly(void) {

-    server.appendonly = 1;
-    server.lastfsync = time(NULL);
-    server.appendfd = open(server.appendfilename,O_WRONLY|O_APPEND|O_CREAT,0644);
-    if (server.appendfd == -1) {
-        redisLog(REDIS_WARNING,"Used tried to switch on AOF via CONFIG, but I can't open the AOF file: %s",strerror(errno));
+    server.aof_last_fsync = server.unixtime;
+    server.aof_fd = open(server.aof_filename,O_WRONLY|O_APPEND|O_CREAT,0644);
+    redisAssert(server.aof_state == REDIS_AOF_OFF);
+    if (server.aof_fd == -1) {
+        redisLog(REDIS_WARNING,"Redis needs to enable the AOF but can't open the append only file: %s",strerror(errno));

         return REDIS_ERR;
     }
     if (rewriteAppendOnlyFileBackground() == REDIS_ERR) {
         return REDIS_ERR;
     }
     if (rewriteAppendOnlyFileBackground() == REDIS_ERR) {

-        server.appendonly = 0;
-        close(server.appendfd);
-        redisLog(REDIS_WARNING,"Used tried to switch on AOF via CONFIG, I can't trigger a background AOF rewrite operation. Check the above logs for more info about the error.",strerror(errno));
+        close(server.aof_fd);
+        redisLog(REDIS_WARNING,"Redis needs to enable the AOF but can't trigger a background AOF rewrite operation. Check the above logs for more info about the error.");

         return REDIS_ERR;
     }
         return REDIS_ERR;
     }

+    /* We correctly switched on AOF, now wait for the rerwite to be complete
+     * in order to append data on disk. */
+    server.aof_state = REDIS_AOF_WAIT_REWRITE;

     return REDIS_OK;
 }
 
     return REDIS_OK;
 }
 


@@ -56,79 +70,161 @@ int startAppendOnly(void) {
  * and the only way the client socket can get a write is entering when the
  * the event loop, we accumulate all the AOF writes in a memory
  * buffer and write it on disk using this function just before entering
  * and the only way the client socket can get a write is entering when the
  * the event loop, we accumulate all the AOF writes in a memory
  * buffer and write it on disk using this function just before entering

- * the event loop again. */
-void flushAppendOnlyFile(void) {
-    time_t now;
+ * the event loop again.
+ *
+ * About the 'force' argument:
+ *
+ * When the fsync policy is set to 'everysec' we may delay the flush if there
+ * is still an fsync() going on in the background thread, since for instance
+ * on Linux write(2) will be blocked by the background fsync anyway.
+ * When this happens we remember that there is some aof buffer to be
+ * flushed ASAP, and will try to do that in the serverCron() function.
+ *
+ * However if force is set to 1 we'll write regardless of the background
+ * fsync. */
+void flushAppendOnlyFile(int force) {

     ssize_t nwritten;
     ssize_t nwritten;

-
-    if (sdslen(server.aofbuf) == 0) return;
+    int sync_in_progress = 0;
+
+    if (sdslen(server.aof_buf) == 0) return;
+
+    if (server.aof_fsync == AOF_FSYNC_EVERYSEC)
+        sync_in_progress = bioPendingJobsOfType(REDIS_BIO_AOF_FSYNC) != 0;
+
+    if (server.aof_fsync == AOF_FSYNC_EVERYSEC && !force) {
+        /* With this append fsync policy we do background fsyncing.
+         * If the fsync is still in progress we can try to delay
+         * the write for a couple of seconds. */
+        if (sync_in_progress) {
+            if (server.aof_flush_postponed_start == 0) {
+                /* No previous write postponinig, remember that we are
+                 * postponing the flush and return. */
+                server.aof_flush_postponed_start = server.unixtime;
+                return;
+            } else if (server.unixtime - server.aof_flush_postponed_start < 2) {
+                /* We were already waiting for fsync to finish, but for less
+                 * than two seconds this is still ok. Postpone again. */
+                return;
+            }
+            /* Otherwise fall trough, and go write since we can't wait
+             * over two seconds. */
+            server.aof_delayed_fsync++;
+            redisLog(REDIS_NOTICE,"Asynchronous AOF fsync is taking too long (disk is busy?). Writing the AOF buffer without waiting for fsync to complete, this may slow down Redis.");
+        }
+    }
+    /* If you are following this code path, then we are going to write so
+     * set reset the postponed flush sentinel to zero. */
+    server.aof_flush_postponed_start = 0;

 
     /* We want to perform a single write. This should be guaranteed atomic
      * at least if the filesystem we are writing is a real physical one.
      * While this will save us against the server being killed I don't think
      * there is much to do about the whole server stopping for power problems
      * or alike */
 
     /* We want to perform a single write. This should be guaranteed atomic
      * at least if the filesystem we are writing is a real physical one.
      * While this will save us against the server being killed I don't think
      * there is much to do about the whole server stopping for power problems
      * or alike */

-     nwritten = write(server.appendfd,server.aofbuf,sdslen(server.aofbuf));
-     if (nwritten != (signed)sdslen(server.aofbuf)) {
+    nwritten = write(server.aof_fd,server.aof_buf,sdslen(server.aof_buf));
+    if (nwritten != (signed)sdslen(server.aof_buf)) {

         /* Ooops, we are in troubles. The best thing to do for now is
          * aborting instead of giving the illusion that everything is
          * working as expected. */
         /* Ooops, we are in troubles. The best thing to do for now is
          * aborting instead of giving the illusion that everything is
          * working as expected. */

-         if (nwritten == -1) {
+        if (nwritten == -1) {

             redisLog(REDIS_WARNING,"Exiting on error writing to the append-only file: %s",strerror(errno));
             redisLog(REDIS_WARNING,"Exiting on error writing to the append-only file: %s",strerror(errno));

-         } else {
-            redisLog(REDIS_WARNING,"Exiting on short write while writing to the append-only file: %s",strerror(errno));
-         }
-         exit(1);
+        } else {
+            redisLog(REDIS_WARNING,"Exiting on short write while writing to "
+                                   "the append-only file: %s (nwritten=%ld, "
+                                   "expected=%ld)",
+                                   strerror(errno),
+                                   (long)nwritten,
+                                   (long)sdslen(server.aof_buf));
+        }
+        exit(1);
+    }
+    server.aof_current_size += nwritten;
+
+    /* Re-use AOF buffer when it is small enough. The maximum comes from the
+     * arena size of 4k minus some overhead (but is otherwise arbitrary). */
+    if ((sdslen(server.aof_buf)+sdsavail(server.aof_buf)) < 4000) {
+        sdsclear(server.aof_buf);
+    } else {
+        sdsfree(server.aof_buf);
+        server.aof_buf = sdsempty();

     }
     }

-    sdsfree(server.aofbuf);
-    server.aofbuf = sdsempty();

 
 

-    /* Don't Fsync if no-appendfsync-on-rewrite is set to yes and we have
-     * childs performing heavy I/O on disk. */
-    if (server.no_appendfsync_on_rewrite &&
-        (server.bgrewritechildpid != -1 || server.bgsavechildpid != -1))
+    /* Don't fsync if no-appendfsync-on-rewrite is set to yes and there are
+     * children doing I/O in the background. */
+    if (server.aof_no_fsync_on_rewrite &&
+        (server.aof_child_pid != -1 || server.rdb_child_pid != -1))

             return;
             return;

-    /* Fsync if needed */
-    now = time(NULL);
-    if (server.appendfsync == APPENDFSYNC_ALWAYS ||
-        (server.appendfsync == APPENDFSYNC_EVERYSEC &&
-         now-server.lastfsync > 1))
-    {
+
+    /* Perform the fsync if needed. */
+    if (server.aof_fsync == AOF_FSYNC_ALWAYS) {

         /* aof_fsync is defined as fdatasync() for Linux in order to avoid
          * flushing metadata. */
         /* aof_fsync is defined as fdatasync() for Linux in order to avoid
          * flushing metadata. */

-        aof_fsync(server.appendfd); /* Let's try to get this data on the disk */
-        server.lastfsync = now;
+        aof_fsync(server.aof_fd); /* Let's try to get this data on the disk */
+        server.aof_last_fsync = server.unixtime;
+    } else if ((server.aof_fsync == AOF_FSYNC_EVERYSEC &&
+                server.unixtime > server.aof_last_fsync)) {
+        if (!sync_in_progress) aof_background_fsync(server.aof_fd);
+        server.aof_last_fsync = server.unixtime;

     }
 }
 
     }
 }
 

-sds catAppendOnlyGenericCommand(sds buf, int argc, robj **argv) {
-    int j;
-    buf = sdscatprintf(buf,"*%d\r\n",argc);
+sds catAppendOnlyGenericCommand(sds dst, int argc, robj **argv) {
+    char buf[32];
+    int len, j;
+    robj *o;
+
+    buf[0] = '*';
+    len = 1+ll2string(buf+1,sizeof(buf)-1,argc);
+    buf[len++] = '\r';
+    buf[len++] = '\n';
+    dst = sdscatlen(dst,buf,len);
+

     for (j = 0; j < argc; j++) {
     for (j = 0; j < argc; j++) {

-        robj *o = getDecodedObject(argv[j]);
-        buf = sdscatprintf(buf,"$%lu\r\n",(unsigned long)sdslen(o->ptr));
-        buf = sdscatlen(buf,o->ptr,sdslen(o->ptr));
-        buf = sdscatlen(buf,"\r\n",2);
+        o = getDecodedObject(argv[j]);
+        buf[0] = '$';
+        len = 1+ll2string(buf+1,sizeof(buf)-1,sdslen(o->ptr));
+        buf[len++] = '\r';
+        buf[len++] = '\n';
+        dst = sdscatlen(dst,buf,len);
+        dst = sdscatlen(dst,o->ptr,sdslen(o->ptr));
+        dst = sdscatlen(dst,"\r\n",2);

         decrRefCount(o);
     }
         decrRefCount(o);
     }

-    return buf;
+    return dst;

 }
 
 }
 

-sds catAppendOnlyExpireAtCommand(sds buf, robj *key, robj *seconds) {
-    int argc = 3;
-    long when;
+/* Create the sds representation of an PEXPIREAT command, using
+ * 'seconds' as time to live and 'cmd' to understand what command
+ * we are translating into a PEXPIREAT.
+ *
+ * This command is used in order to translate EXPIRE and PEXPIRE commands
+ * into PEXPIREAT command so that we retain precision in the append only
+ * file, and the time is always absolute and not relative. */
+sds catAppendOnlyExpireAtCommand(sds buf, struct redisCommand *cmd, robj *key, robj *seconds) {
+    long long when;

     robj *argv[3];
 
     /* Make sure we can use strtol */
     seconds = getDecodedObject(seconds);
     robj *argv[3];
 
     /* Make sure we can use strtol */
     seconds = getDecodedObject(seconds);

-    when = time(NULL)+strtol(seconds->ptr,NULL,10);
+    when = strtoll(seconds->ptr,NULL,10);
+    /* Convert argument into milliseconds for EXPIRE, SETEX, EXPIREAT */
+    if (cmd->proc == expireCommand || cmd->proc == setexCommand ||
+        cmd->proc == expireatCommand)
+    {
+        when *= 1000;
+    }
+    /* Convert into absolute time for EXPIRE, PEXPIRE, SETEX, PSETEX */
+    if (cmd->proc == expireCommand || cmd->proc == pexpireCommand ||
+        cmd->proc == setexCommand || cmd->proc == psetexCommand)
+    {
+        when += mstime();
+    }

     decrRefCount(seconds);
 
     decrRefCount(seconds);
 

-    argv[0] = createStringObject("EXPIREAT",8);
+    argv[0] = createStringObject("PEXPIREAT",9);

     argv[1] = key;
     argv[1] = key;

-    argv[2] = createObject(REDIS_STRING,
-        sdscatprintf(sdsempty(),"%ld",when));
-    buf = catAppendOnlyGenericCommand(buf, argc, argv);
+    argv[2] = createStringObjectFromLongLong(when);
+    buf = catAppendOnlyGenericCommand(buf, 3, argv);

     decrRefCount(argv[0]);
     decrRefCount(argv[2]);
     return buf;
     decrRefCount(argv[0]);
     decrRefCount(argv[2]);
     return buf;


@@ -140,41 +236,46 @@ void feedAppendOnlyFile(struct redisCommand *cmd, int dictid, robj **argv, int a
 
     /* The DB this command was targetting is not the same as the last command
      * we appendend. To issue a SELECT command is needed. */
 
     /* The DB this command was targetting is not the same as the last command
      * we appendend. To issue a SELECT command is needed. */

-    if (dictid != server.appendseldb) {
+    if (dictid != server.aof_selected_db) {

         char seldb[64];
 
         snprintf(seldb,sizeof(seldb),"%d",dictid);
         buf = sdscatprintf(buf,"*2\r\n$6\r\nSELECT\r\n$%lu\r\n%s\r\n",
             (unsigned long)strlen(seldb),seldb);
         char seldb[64];
 
         snprintf(seldb,sizeof(seldb),"%d",dictid);
         buf = sdscatprintf(buf,"*2\r\n$6\r\nSELECT\r\n$%lu\r\n%s\r\n",
             (unsigned long)strlen(seldb),seldb);

-        server.appendseldb = dictid;
+        server.aof_selected_db = dictid;

     }
 
     }
 

-    if (cmd->proc == expireCommand) {
-        /* Translate EXPIRE into EXPIREAT */
-        buf = catAppendOnlyExpireAtCommand(buf,argv[1],argv[2]);
-    } else if (cmd->proc == setexCommand) {
-        /* Translate SETEX to SET and EXPIREAT */
+    if (cmd->proc == expireCommand || cmd->proc == pexpireCommand ||
+        cmd->proc == expireatCommand) {
+        /* Translate EXPIRE/PEXPIRE/EXPIREAT into PEXPIREAT */
+        buf = catAppendOnlyExpireAtCommand(buf,cmd,argv[1],argv[2]);
+    } else if (cmd->proc == setexCommand || cmd->proc == psetexCommand) {
+        /* Translate SETEX/PSETEX to SET and PEXPIREAT */

         tmpargv[0] = createStringObject("SET",3);
         tmpargv[1] = argv[1];
         tmpargv[2] = argv[3];
         buf = catAppendOnlyGenericCommand(buf,3,tmpargv);
         decrRefCount(tmpargv[0]);
         tmpargv[0] = createStringObject("SET",3);
         tmpargv[1] = argv[1];
         tmpargv[2] = argv[3];
         buf = catAppendOnlyGenericCommand(buf,3,tmpargv);
         decrRefCount(tmpargv[0]);

-        buf = catAppendOnlyExpireAtCommand(buf,argv[1],argv[2]);
+        buf = catAppendOnlyExpireAtCommand(buf,cmd,argv[1],argv[2]);

     } else {
     } else {

+        /* All the other commands don't need translation or need the
+         * same translation already operated in the command vector
+         * for the replication itself. */

         buf = catAppendOnlyGenericCommand(buf,argc,argv);
     }
 
     /* Append to the AOF buffer. This will be flushed on disk just before
      * of re-entering the event loop, so before the client will get a
      * positive reply about the operation performed. */
         buf = catAppendOnlyGenericCommand(buf,argc,argv);
     }
 
     /* Append to the AOF buffer. This will be flushed on disk just before
      * of re-entering the event loop, so before the client will get a
      * positive reply about the operation performed. */

-    server.aofbuf = sdscatlen(server.aofbuf,buf,sdslen(buf));
+    if (server.aof_state == REDIS_AOF_ON)
+        server.aof_buf = sdscatlen(server.aof_buf,buf,sdslen(buf));

 
     /* If a background append only file rewriting is in progress we want to
      * accumulate the differences between the child DB and the current one
      * in a buffer, so that when the child process will do its work we
      * can append the differences to the new append only file. */
 
     /* If a background append only file rewriting is in progress we want to
      * accumulate the differences between the child DB and the current one
      * in a buffer, so that when the child process will do its work we
      * can append the differences to the new append only file. */

-    if (server.bgrewritechildpid != -1)
-        server.bgrewritebuf = sdscatlen(server.bgrewritebuf,buf,sdslen(buf));
+    if (server.aof_child_pid != -1)
+        server.aof_rewrite_buf = sdscatlen(server.aof_rewrite_buf,buf,sdslen(buf));

 
     sdsfree(buf);
 }
 
     sdsfree(buf);
 }


@@ -187,6 +288,7 @@ struct redisClient *createFakeClient(void) {
     selectDb(c,0);
     c->fd = -1;
     c->querybuf = sdsempty();
     selectDb(c,0);
     c->fd = -1;
     c->querybuf = sdsempty();

+    c->querybuf_peak = 0;

     c->argc = 0;
     c->argv = NULL;
     c->bufpos = 0;
     c->argc = 0;
     c->argv = NULL;
     c->bufpos = 0;


@@ -195,6 +297,8 @@ struct redisClient *createFakeClient(void) {
      * so that Redis will not try to send replies to this client. */
     c->replstate = REDIS_REPL_WAIT_BGSAVE_START;
     c->reply = listCreate();
      * so that Redis will not try to send replies to this client. */
     c->replstate = REDIS_REPL_WAIT_BGSAVE_START;
     c->reply = listCreate();

+    c->reply_bytes = 0;
+    c->obuf_soft_limit_reached_time = 0;

     c->watched_keys = listCreate();
     listSetFreeMethod(c->reply,decrRefCount);
     listSetDupMethod(c->reply,dupClientReplyValue);
     c->watched_keys = listCreate();
     listSetFreeMethod(c->reply,decrRefCount);
     listSetDupMethod(c->reply,dupClientReplyValue);


@@ -217,10 +321,14 @@ int loadAppendOnlyFile(char *filename) {
     struct redisClient *fakeClient;
     FILE *fp = fopen(filename,"r");
     struct redis_stat sb;
     struct redisClient *fakeClient;
     FILE *fp = fopen(filename,"r");
     struct redis_stat sb;

-    int appendonly = server.appendonly;
+    int old_aof_state = server.aof_state;
+    long loops = 0;

 
 

-    if (redis_fstat(fileno(fp),&sb) != -1 && sb.st_size == 0)
+    if (fp && redis_fstat(fileno(fp),&sb) != -1 && sb.st_size == 0) {
+        server.aof_current_size = 0;
+        fclose(fp);

         return REDIS_ERR;
         return REDIS_ERR;

+    }

 
     if (fp == NULL) {
         redisLog(REDIS_WARNING,"Fatal error: can't open the append log file for reading: %s",strerror(errno));
 
     if (fp == NULL) {
         redisLog(REDIS_WARNING,"Fatal error: can't open the append log file for reading: %s",strerror(errno));


@@ -229,9 +337,11 @@ int loadAppendOnlyFile(char *filename) {
 
     /* Temporarily disable AOF, to prevent EXEC from feeding a MULTI
      * to the same file we're about to read. */
 
     /* Temporarily disable AOF, to prevent EXEC from feeding a MULTI
      * to the same file we're about to read. */

-    server.appendonly = 0;
+    server.aof_state = REDIS_AOF_OFF;

 
     fakeClient = createFakeClient();
 
     fakeClient = createFakeClient();

+    startLoading(fp);
+

     while(1) {
         int argc, j;
         unsigned long len;
     while(1) {
         int argc, j;
         unsigned long len;


@@ -239,7 +349,12 @@ int loadAppendOnlyFile(char *filename) {
         char buf[128];
         sds argsds;
         struct redisCommand *cmd;
         char buf[128];
         sds argsds;
         struct redisCommand *cmd;

-        int force_swapout;
+
+        /* Serve the clients from time to time */
+        if (!(loops++ % 1000)) {
+            loadingProgress(ftello(fp));
+            aeProcessEvents(server.el, AE_FILE_EVENTS|AE_DONT_WAIT);
+        }

 
         if (fgets(buf,sizeof(buf),fp) == NULL) {
             if (feof(fp))
 
         if (fgets(buf,sizeof(buf),fp) == NULL) {
             if (feof(fp))


@@ -249,6 +364,8 @@ int loadAppendOnlyFile(char *filename) {
         }
         if (buf[0] != '*') goto fmterr;
         argc = atoi(buf+1);
         }
         if (buf[0] != '*') goto fmterr;
         argc = atoi(buf+1);

+        if (argc < 1) goto fmterr;
+

         argv = zmalloc(sizeof(robj*)*argc);
         for (j = 0; j < argc; j++) {
             if (fgets(buf,sizeof(buf),fp) == NULL) goto readerr;
         argv = zmalloc(sizeof(robj*)*argc);
         for (j = 0; j < argc; j++) {
             if (fgets(buf,sizeof(buf),fp) == NULL) goto readerr;


@@ -266,9 +383,6 @@ int loadAppendOnlyFile(char *filename) {
             redisLog(REDIS_WARNING,"Unknown command '%s' reading the append only file", argv[0]->ptr);
             exit(1);
         }
             redisLog(REDIS_WARNING,"Unknown command '%s' reading the append only file", argv[0]->ptr);
             exit(1);
         }

-        /* Try object encoding */
-        if (cmd->flags & REDIS_CMD_BULK)
-            argv[argc-1] = tryObjectEncoding(argv[argc-1]);

         /* Run the command in the context of a fake client */
         fakeClient->argc = argc;
         fakeClient->argv = argv;
         /* Run the command in the context of a fake client */
         fakeClient->argc = argc;
         fakeClient->argv = argv;


@@ -276,21 +390,14 @@ int loadAppendOnlyFile(char *filename) {
 
         /* The fake client should not have a reply */
         redisAssert(fakeClient->bufpos == 0 && listLength(fakeClient->reply) == 0);
 
         /* The fake client should not have a reply */
         redisAssert(fakeClient->bufpos == 0 && listLength(fakeClient->reply) == 0);

-
-        /* Clean up, ready for the next command */
-        for (j = 0; j < argc; j++) decrRefCount(argv[j]);
-        zfree(argv);
-
-        /* Handle swapping while loading big datasets when VM is on */
-        force_swapout = 0;
-        if ((zmalloc_used_memory() - server.vm_max_memory) > 1024*1024*32)
-            force_swapout = 1;
-
-        if (server.vm_enabled && force_swapout) {
-            while (zmalloc_used_memory() > server.vm_max_memory) {
-                if (vmSwapOneObjectBlocking() == REDIS_ERR) break;
-            }
-        }
+        /* The fake client should never get blocked */
+        redisAssert((fakeClient->flags & REDIS_BLOCKED) == 0);
+
+        /* Clean up. Command code may have changed argv/argc so we use the
+         * argv/argc of the client instead of the local variables. */
+        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.


@@ -299,7 +406,10 @@ int loadAppendOnlyFile(char *filename) {
 
     fclose(fp);
     freeFakeClient(fakeClient);
 
     fclose(fp);
     freeFakeClient(fakeClient);

-    server.appendonly = appendonly;
+    server.aof_state = old_aof_state;
+    stopLoading();
+    aofUpdateCurrentSize();
+    server.aof_rewrite_base_size = server.aof_current_size;

     return REDIS_OK;
 
 readerr:
     return REDIS_OK;
 
 readerr:


@@ -314,244 +424,332 @@ fmterr:
     exit(1);
 }
 
     exit(1);
 }
 

-/* Write binary-safe string into a file in the bulkformat
- * $<count>\r\n<payload>\r\n */
-int fwriteBulkString(FILE *fp, char *s, unsigned long len) {
-    char cbuf[128];
-    int clen;
-    cbuf[0] = '$';
-    clen = 1+ll2string(cbuf+1,sizeof(cbuf)-1,len);
-    cbuf[clen++] = '\r';
-    cbuf[clen++] = '\n';
-    if (fwrite(cbuf,clen,1,fp) == 0) return 0;
-    if (len > 0 && fwrite(s,len,1,fp) == 0) return 0;
-    if (fwrite("\r\n",2,1,fp) == 0) return 0;
+/* Delegate writing an object to writing a bulk string or bulk long long.
+ * This is not placed in rio.c since that adds the redis.h dependency. */
+int rioWriteBulkObject(rio *r, robj *obj) {
+    /* Avoid using getDecodedObject to help copy-on-write (we are often
+     * in a child process when this function is called). */
+    if (obj->encoding == REDIS_ENCODING_INT) {
+        return rioWriteBulkLongLong(r,(long)obj->ptr);
+    } else if (obj->encoding == REDIS_ENCODING_RAW) {
+        return rioWriteBulkString(r,obj->ptr,sdslen(obj->ptr));
+    } else {
+        redisPanic("Unknown string encoding");
+    }
+}
+
+/* Emit the commands needed to rebuild a list object.
+ * The function returns 0 on error, 1 on success. */
+int rewriteListObject(rio *r, robj *key, robj *o) {
+    long long count = 0, items = listTypeLength(o);
+
+    if (o->encoding == REDIS_ENCODING_ZIPLIST) {
+        unsigned char *zl = o->ptr;
+        unsigned char *p = ziplistIndex(zl,0);
+        unsigned char *vstr;
+        unsigned int vlen;
+        long long vlong;
+
+        while(ziplistGet(p,&vstr,&vlen,&vlong)) {
+            if (count == 0) {
+                int cmd_items = (items > REDIS_AOF_REWRITE_ITEMS_PER_CMD) ?
+                    REDIS_AOF_REWRITE_ITEMS_PER_CMD : items;
+
+                if (rioWriteBulkCount(r,'*',2+cmd_items) == 0) return 0;
+                if (rioWriteBulkString(r,"RPUSH",5) == 0) return 0;
+                if (rioWriteBulkObject(r,key) == 0) return 0;
+            }
+            if (vstr) {
+                if (rioWriteBulkString(r,(char*)vstr,vlen) == 0) return 0;
+            } else {
+                if (rioWriteBulkLongLong(r,vlong) == 0) return 0;
+            }
+            p = ziplistNext(zl,p);
+            if (++count == REDIS_AOF_REWRITE_ITEMS_PER_CMD) count = 0;
+            items--;
+        }
+    } else if (o->encoding == REDIS_ENCODING_LINKEDLIST) {
+        list *list = o->ptr;
+        listNode *ln;
+        listIter li;
+
+        listRewind(list,&li);
+        while((ln = listNext(&li))) {
+            robj *eleobj = listNodeValue(ln);
+
+            if (count == 0) {
+                int cmd_items = (items > REDIS_AOF_REWRITE_ITEMS_PER_CMD) ?
+                    REDIS_AOF_REWRITE_ITEMS_PER_CMD : items;
+
+                if (rioWriteBulkCount(r,'*',2+cmd_items) == 0) return 0;
+                if (rioWriteBulkString(r,"RPUSH",5) == 0) return 0;
+                if (rioWriteBulkObject(r,key) == 0) return 0;
+            }
+            if (rioWriteBulkObject(r,eleobj) == 0) return 0;
+            if (++count == REDIS_AOF_REWRITE_ITEMS_PER_CMD) count = 0;
+            items--;
+        }
+    } else {
+        redisPanic("Unknown list encoding");
+    }

     return 1;
 }
 
     return 1;
 }
 

-/* Write a double value in bulk format $<count>\r\n<payload>\r\n */
-int fwriteBulkDouble(FILE *fp, double d) {
-    char buf[128], dbuf[128];
+/* Emit the commands needed to rebuild a set object.
+ * The function returns 0 on error, 1 on success. */
+int rewriteSetObject(rio *r, robj *key, robj *o) {
+    long long count = 0, items = setTypeSize(o);
+
+    if (o->encoding == REDIS_ENCODING_INTSET) {
+        int ii = 0;
+        int64_t llval;
+
+        while(intsetGet(o->ptr,ii++,&llval)) {
+            if (count == 0) {
+                int cmd_items = (items > REDIS_AOF_REWRITE_ITEMS_PER_CMD) ?
+                    REDIS_AOF_REWRITE_ITEMS_PER_CMD : items;

 
 

-    snprintf(dbuf,sizeof(dbuf),"%.17g\r\n",d);
-    snprintf(buf,sizeof(buf),"$%lu\r\n",(unsigned long)strlen(dbuf)-2);
-    if (fwrite(buf,strlen(buf),1,fp) == 0) return 0;
-    if (fwrite(dbuf,strlen(dbuf),1,fp) == 0) return 0;
+                if (rioWriteBulkCount(r,'*',2+cmd_items) == 0) return 0;
+                if (rioWriteBulkString(r,"SADD",4) == 0) return 0;
+                if (rioWriteBulkObject(r,key) == 0) return 0;
+            }
+            if (rioWriteBulkLongLong(r,llval) == 0) return 0;
+            if (++count == REDIS_AOF_REWRITE_ITEMS_PER_CMD) count = 0;
+            items--;
+        }
+    } else if (o->encoding == REDIS_ENCODING_HT) {
+        dictIterator *di = dictGetIterator(o->ptr);
+        dictEntry *de;
+
+        while((de = dictNext(di)) != NULL) {
+            robj *eleobj = dictGetKey(de);
+            if (count == 0) {
+                int cmd_items = (items > REDIS_AOF_REWRITE_ITEMS_PER_CMD) ?
+                    REDIS_AOF_REWRITE_ITEMS_PER_CMD : items;
+
+                if (rioWriteBulkCount(r,'*',2+cmd_items) == 0) return 0;
+                if (rioWriteBulkString(r,"SADD",4) == 0) return 0;
+                if (rioWriteBulkObject(r,key) == 0) return 0;
+            }
+            if (rioWriteBulkObject(r,eleobj) == 0) return 0;
+            if (++count == REDIS_AOF_REWRITE_ITEMS_PER_CMD) count = 0;
+            items--;
+        }
+        dictReleaseIterator(di);
+    } else {
+        redisPanic("Unknown set encoding");
+    }

     return 1;
 }
 
     return 1;
 }
 

-/* Write a long value in bulk format $<count>\r\n<payload>\r\n */
-int fwriteBulkLongLong(FILE *fp, long long l) {
-    char bbuf[128], lbuf[128];
-    unsigned int blen, llen;
-    llen = ll2string(lbuf,32,l);
-    blen = snprintf(bbuf,sizeof(bbuf),"$%u\r\n%s\r\n",llen,lbuf);
-    if (fwrite(bbuf,blen,1,fp) == 0) return 0;
+/* Emit the commands needed to rebuild a sorted set object.
+ * The function returns 0 on error, 1 on success. */
+int rewriteSortedSetObject(rio *r, robj *key, robj *o) {
+    long long count = 0, items = zsetLength(o);
+
+    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 (count == 0) {
+                int cmd_items = (items > REDIS_AOF_REWRITE_ITEMS_PER_CMD) ?
+                    REDIS_AOF_REWRITE_ITEMS_PER_CMD : items;
+
+                if (rioWriteBulkCount(r,'*',2+cmd_items*2) == 0) return 0;
+                if (rioWriteBulkString(r,"ZADD",4) == 0) return 0;
+                if (rioWriteBulkObject(r,key) == 0) return 0;
+            }
+            if (rioWriteBulkDouble(r,score) == 0) return 0;
+            if (vstr != NULL) {
+                if (rioWriteBulkString(r,(char*)vstr,vlen) == 0) return 0;
+            } else {
+                if (rioWriteBulkLongLong(r,vll) == 0) return 0;
+            }
+            zzlNext(zl,&eptr,&sptr);
+            if (++count == REDIS_AOF_REWRITE_ITEMS_PER_CMD) count = 0;
+            items--;
+        }
+    } else if (o->encoding == REDIS_ENCODING_SKIPLIST) {
+        zset *zs = o->ptr;
+        dictIterator *di = dictGetIterator(zs->dict);
+        dictEntry *de;
+
+        while((de = dictNext(di)) != NULL) {
+            robj *eleobj = dictGetKey(de);
+            double *score = dictGetVal(de);
+
+            if (count == 0) {
+                int cmd_items = (items > REDIS_AOF_REWRITE_ITEMS_PER_CMD) ?
+                    REDIS_AOF_REWRITE_ITEMS_PER_CMD : items;
+
+                if (rioWriteBulkCount(r,'*',2+cmd_items*2) == 0) return 0;
+                if (rioWriteBulkString(r,"ZADD",4) == 0) return 0;
+                if (rioWriteBulkObject(r,key) == 0) return 0;
+            }
+            if (rioWriteBulkDouble(r,*score) == 0) return 0;
+            if (rioWriteBulkObject(r,eleobj) == 0) return 0;
+            if (++count == REDIS_AOF_REWRITE_ITEMS_PER_CMD) count = 0;
+            items--;
+        }
+        dictReleaseIterator(di);
+    } else {
+        redisPanic("Unknown sorted zset encoding");
+    }

     return 1;
 }
 
     return 1;
 }
 

-/* Delegate writing an object to writing a bulk string or bulk long long. */
-int fwriteBulkObject(FILE *fp, robj *obj) {
-    /* Avoid using getDecodedObject to help copy-on-write (we are often
-     * in a child process when this function is called). */
-    if (obj->encoding == REDIS_ENCODING_INT) {
-        return fwriteBulkLongLong(fp,(long)obj->ptr);
-    } else if (obj->encoding == REDIS_ENCODING_RAW) {
-        return fwriteBulkString(fp,obj->ptr,sdslen(obj->ptr));
-    } else {
-        redisPanic("Unknown string encoding");
+/* Write either the key or the value of the currently selected item of an hash.
+ * The 'hi' argument passes a valid Redis hash iterator.
+ * The 'what' filed specifies if to write a key or a value and can be
+ * either REDIS_HASH_KEY or REDIS_HASH_VALUE.
+ *
+ * The function returns 0 on error, non-zero on success. */
+static int rioWriteHashIteratorCursor(rio *r, hashTypeIterator *hi, int what) {
+    if (hi->encoding == REDIS_ENCODING_ZIPLIST) {
+        unsigned char *vstr = NULL;
+        unsigned int vlen = UINT_MAX;
+        long long vll = LLONG_MAX;
+
+        hashTypeCurrentFromZiplist(hi, what, &vstr, &vlen, &vll);
+        if (vstr) {
+            return rioWriteBulkString(r, (char*)vstr, vlen);
+        } else {
+            return rioWriteBulkLongLong(r, vll);
+        }
+
+    } else if (hi->encoding == REDIS_ENCODING_HT) {
+        robj *value;
+
+        hashTypeCurrentFromHashTable(hi, what, &value);
+        return rioWriteBulkObject(r, value);

     }
     }

+
+    redisPanic("Unknown hash encoding");
+    return 0;
+}
+
+/* Emit the commands needed to rebuild a hash object.
+ * The function returns 0 on error, 1 on success. */
+int rewriteHashObject(rio *r, robj *key, robj *o) {
+    hashTypeIterator *hi;
+    long long count = 0, items = hashTypeLength(o);
+
+    hi = hashTypeInitIterator(o);
+    while (hashTypeNext(hi) != REDIS_ERR) {
+        if (count == 0) {
+            int cmd_items = (items > REDIS_AOF_REWRITE_ITEMS_PER_CMD) ?
+                REDIS_AOF_REWRITE_ITEMS_PER_CMD : items;
+
+            if (rioWriteBulkCount(r,'*',2+cmd_items*2) == 0) return 0;
+            if (rioWriteBulkString(r,"HMSET",5) == 0) return 0;
+            if (rioWriteBulkObject(r,key) == 0) return 0;
+        }
+
+        if (rioWriteHashIteratorCursor(r, hi, REDIS_HASH_KEY) == 0) return 0;
+        if (rioWriteHashIteratorCursor(r, hi, REDIS_HASH_VALUE) == 0) return 0;
+        if (++count == REDIS_AOF_REWRITE_ITEMS_PER_CMD) count = 0;
+        items--;
+    }
+
+    hashTypeReleaseIterator(hi);
+
+    return 1;

 }
 
 /* Write a sequence of commands able to fully rebuild the dataset into
 }
 
 /* Write a sequence of commands able to fully rebuild the dataset into

- * "filename". Used both by REWRITEAOF and BGREWRITEAOF. */
+ * "filename". Used both by REWRITEAOF and BGREWRITEAOF.
+ *
+ * In order to minimize the number of commands needed in the rewritten
+ * log Redis uses variadic commands when possible, such as RPUSH, SADD
+ * and ZADD. However at max REDIS_AOF_REWRITE_ITEMS_PER_CMD items per time
+ * are inserted using a single command. */

 int rewriteAppendOnlyFile(char *filename) {
     dictIterator *di = NULL;
     dictEntry *de;
 int rewriteAppendOnlyFile(char *filename) {
     dictIterator *di = NULL;
     dictEntry *de;

+    rio aof;

     FILE *fp;
     char tmpfile[256];
     int j;
     FILE *fp;
     char tmpfile[256];
     int j;

-    time_t now = time(NULL);
+    long long now = mstime();

 
     /* Note that we have to use a different temp name here compared to the
      * one used by rewriteAppendOnlyFileBackground() function. */
     snprintf(tmpfile,256,"temp-rewriteaof-%d.aof", (int) getpid());
     fp = fopen(tmpfile,"w");
     if (!fp) {
 
     /* Note that we have to use a different temp name here compared to the
      * one used by rewriteAppendOnlyFileBackground() function. */
     snprintf(tmpfile,256,"temp-rewriteaof-%d.aof", (int) getpid());
     fp = fopen(tmpfile,"w");
     if (!fp) {

-        redisLog(REDIS_WARNING, "Failed rewriting the append only file: %s", strerror(errno));
+        redisLog(REDIS_WARNING, "Opening the temp file for AOF rewrite in rewriteAppendOnlyFile(): %s", strerror(errno));

         return REDIS_ERR;
     }
         return REDIS_ERR;
     }

+
+    rioInitWithFile(&aof,fp);

     for (j = 0; j < server.dbnum; j++) {
         char selectcmd[] = "*2\r\n$6\r\nSELECT\r\n";
         redisDb *db = server.db+j;
         dict *d = db->dict;
         if (dictSize(d) == 0) continue;
     for (j = 0; j < server.dbnum; j++) {
         char selectcmd[] = "*2\r\n$6\r\nSELECT\r\n";
         redisDb *db = server.db+j;
         dict *d = db->dict;
         if (dictSize(d) == 0) continue;

-        di = dictGetIterator(d);
+        di = dictGetSafeIterator(d);

         if (!di) {
             fclose(fp);
             return REDIS_ERR;
         }
 
         /* SELECT the new DB */
         if (!di) {
             fclose(fp);
             return REDIS_ERR;
         }
 
         /* SELECT the new DB */

-        if (fwrite(selectcmd,sizeof(selectcmd)-1,1,fp) == 0) goto werr;
-        if (fwriteBulkLongLong(fp,j) == 0) goto werr;
+        if (rioWrite(&aof,selectcmd,sizeof(selectcmd)-1) == 0) goto werr;
+        if (rioWriteBulkLongLong(&aof,j) == 0) goto werr;

 
         /* Iterate this DB writing every entry */
         while((de = dictNext(di)) != NULL) {
 
         /* Iterate this DB writing every entry */
         while((de = dictNext(di)) != NULL) {

-            sds keystr = dictGetEntryKey(de);
+            sds keystr;

             robj key, *o;
             robj key, *o;

-            time_t expiretime;
-            int swapped;
+            long long expiretime;

 
 

-            keystr = dictGetEntryKey(de);
-            o = dictGetEntryVal(de);
+            keystr = dictGetKey(de);
+            o = dictGetVal(de);

             initStaticStringObject(key,keystr);
             initStaticStringObject(key,keystr);

-            /* If the value for this key is swapped, load a preview in memory.
-             * We use a "swapped" flag to remember if we need to free the
-             * value object instead to just increment the ref count anyway
-             * in order to avoid copy-on-write of pages if we are forked() */
-            if (!server.vm_enabled || o->storage == REDIS_VM_MEMORY ||
-                o->storage == REDIS_VM_SWAPPING) {
-                swapped = 0;
-            } else {
-                o = vmPreviewObject(o);
-                swapped = 1;
-            }
+

             expiretime = getExpire(db,&key);
 
             /* Save the key and associated value */
             if (o->type == REDIS_STRING) {
                 /* Emit a SET command */
                 char cmd[]="*3\r\n$3\r\nSET\r\n";
             expiretime = getExpire(db,&key);
 
             /* Save the key and associated value */
             if (o->type == REDIS_STRING) {
                 /* Emit a SET command */
                 char cmd[]="*3\r\n$3\r\nSET\r\n";

-                if (fwrite(cmd,sizeof(cmd)-1,1,fp) == 0) goto werr;
+                if (rioWrite(&aof,cmd,sizeof(cmd)-1) == 0) goto werr;

                 /* Key and value */
                 /* Key and value */

-                if (fwriteBulkObject(fp,&key) == 0) goto werr;
-                if (fwriteBulkObject(fp,o) == 0) goto werr;
+                if (rioWriteBulkObject(&aof,&key) == 0) goto werr;
+                if (rioWriteBulkObject(&aof,o) == 0) goto werr;

             } else if (o->type == REDIS_LIST) {
             } else if (o->type == REDIS_LIST) {

-                /* Emit the RPUSHes needed to rebuild the list */
-                char cmd[]="*3\r\n$5\r\nRPUSH\r\n";
-                if (o->encoding == REDIS_ENCODING_ZIPLIST) {
-                    unsigned char *zl = o->ptr;
-                    unsigned char *p = ziplistIndex(zl,0);
-                    unsigned char *vstr;
-                    unsigned int vlen;
-                    long long vlong;
-
-                    while(ziplistGet(p,&vstr,&vlen,&vlong)) {
-                        if (fwrite(cmd,sizeof(cmd)-1,1,fp) == 0) goto werr;
-                        if (fwriteBulkObject(fp,&key) == 0) goto werr;
-                        if (vstr) {
-                            if (fwriteBulkString(fp,(char*)vstr,vlen) == 0)
-                                goto werr;
-                        } else {
-                            if (fwriteBulkLongLong(fp,vlong) == 0)
-                                goto werr;
-                        }
-                        p = ziplistNext(zl,p);
-                    }
-                } else if (o->encoding == REDIS_ENCODING_LINKEDLIST) {
-                    list *list = o->ptr;
-                    listNode *ln;
-                    listIter li;
-
-                    listRewind(list,&li);
-                    while((ln = listNext(&li))) {
-                        robj *eleobj = listNodeValue(ln);
-
-                        if (fwrite(cmd,sizeof(cmd)-1,1,fp) == 0) goto werr;
-                        if (fwriteBulkObject(fp,&key) == 0) goto werr;
-                        if (fwriteBulkObject(fp,eleobj) == 0) goto werr;
-                    }
-                } else {
-                    redisPanic("Unknown list encoding");
-                }
+                if (rewriteListObject(&aof,&key,o) == 0) goto werr;

             } else if (o->type == REDIS_SET) {
             } else if (o->type == REDIS_SET) {

-                char cmd[]="*3\r\n$4\r\nSADD\r\n";
-
-                /* Emit the SADDs needed to rebuild the set */
-                if (o->encoding == REDIS_ENCODING_INTSET) {
-                    int ii = 0;
-                    int64_t llval;
-                    while(intsetGet(o->ptr,ii++,&llval)) {
-                        if (fwrite(cmd,sizeof(cmd)-1,1,fp) == 0) goto werr;
-                        if (fwriteBulkObject(fp,&key) == 0) goto werr;
-                        if (fwriteBulkLongLong(fp,llval) == 0) goto werr;
-                    }
-                } else if (o->encoding == REDIS_ENCODING_HT) {
-                    dictIterator *di = dictGetIterator(o->ptr);
-                    dictEntry *de;
-                    while((de = dictNext(di)) != NULL) {
-                        robj *eleobj = dictGetEntryKey(de);
-                        if (fwrite(cmd,sizeof(cmd)-1,1,fp) == 0) goto werr;
-                        if (fwriteBulkObject(fp,&key) == 0) goto werr;
-                        if (fwriteBulkObject(fp,eleobj) == 0) goto werr;
-                    }
-                    dictReleaseIterator(di);
-                } else {
-                    redisPanic("Unknown set encoding");
-                }
+                if (rewriteSetObject(&aof,&key,o) == 0) goto werr;

             } else if (o->type == REDIS_ZSET) {
             } else if (o->type == REDIS_ZSET) {

-                /* Emit the ZADDs needed to rebuild the sorted set */
-                zset *zs = o->ptr;
-                dictIterator *di = dictGetIterator(zs->dict);
-                dictEntry *de;
-
-                while((de = dictNext(di)) != NULL) {
-                    char cmd[]="*4\r\n$4\r\nZADD\r\n";
-                    robj *eleobj = dictGetEntryKey(de);
-                    double *score = dictGetEntryVal(de);
-
-                    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 (fwriteBulkObject(fp,eleobj) == 0) goto werr;
-                }
-                dictReleaseIterator(di);
+                if (rewriteSortedSetObject(&aof,&key,o) == 0) goto werr;

             } else if (o->type == REDIS_HASH) {
             } else if (o->type == REDIS_HASH) {

-                char cmd[]="*4\r\n$4\r\nHSET\r\n";
-
-                /* Emit the HSETs needed to rebuild the hash */
-                if (o->encoding == REDIS_ENCODING_ZIPMAP) {
-                    unsigned char *p = zipmapRewind(o->ptr);
-                    unsigned char *field, *val;
-                    unsigned int flen, vlen;
-
-                    while((p = zipmapNext(p,&field,&flen,&val,&vlen)) != NULL) {
-                        if (fwrite(cmd,sizeof(cmd)-1,1,fp) == 0) goto werr;
-                        if (fwriteBulkObject(fp,&key) == 0) goto werr;
-                        if (fwriteBulkString(fp,(char*)field,flen) == 0)
-                            goto werr;
-                        if (fwriteBulkString(fp,(char*)val,vlen) == 0)
-                            goto werr;
-                    }
-                } else {
-                    dictIterator *di = dictGetIterator(o->ptr);
-                    dictEntry *de;
-
-                    while((de = dictNext(di)) != NULL) {
-                        robj *field = dictGetEntryKey(de);
-                        robj *val = dictGetEntryVal(de);
-
-                        if (fwrite(cmd,sizeof(cmd)-1,1,fp) == 0) goto werr;
-                        if (fwriteBulkObject(fp,&key) == 0) goto werr;
-                        if (fwriteBulkObject(fp,field) == 0) goto werr;
-                        if (fwriteBulkObject(fp,val) == 0) goto werr;
-                    }
-                    dictReleaseIterator(di);
-                }
+                if (rewriteHashObject(&aof,&key,o) == 0) goto werr;

             } else {
                 redisPanic("Unknown object type");
             }
             /* Save the expire time */
             if (expiretime != -1) {
             } else {
                 redisPanic("Unknown object type");
             }
             /* Save the expire time */
             if (expiretime != -1) {

-                char cmd[]="*3\r\n$8\r\nEXPIREAT\r\n";
+                char cmd[]="*3\r\n$9\r\nPEXPIREAT\r\n";

                 /* If this key is already expired skip it */
                 if (expiretime < now) continue;
                 /* If this key is already expired skip it */
                 if (expiretime < now) continue;

-                if (fwrite(cmd,sizeof(cmd)-1,1,fp) == 0) goto werr;
-                if (fwriteBulkObject(fp,&key) == 0) goto werr;
-                if (fwriteBulkLongLong(fp,expiretime) == 0) goto werr;
+                if (rioWrite(&aof,cmd,sizeof(cmd)-1) == 0) goto werr;
+                if (rioWriteBulkObject(&aof,&key) == 0) goto werr;
+                if (rioWriteBulkLongLong(&aof,expiretime) == 0) goto werr;

             }
             }

-            if (swapped) decrRefCount(o);

         }
         dictReleaseIterator(di);
     }
         }
         dictReleaseIterator(di);
     }


@@ -584,24 +782,25 @@ werr:
  * 1) The user calls BGREWRITEAOF
  * 2) Redis calls this function, that forks():
  *    2a) the child rewrite the append only file in a temp file.
  * 1) The user calls BGREWRITEAOF
  * 2) Redis calls this function, that forks():
  *    2a) the child rewrite the append only file in a temp file.

- *    2b) the parent accumulates differences in server.bgrewritebuf.
+ *    2b) the parent accumulates differences in server.aof_rewrite_buf.

  * 3) When the child finished '2a' exists.
  * 4) The parent will trap the exit code, if it's OK, will append the
  * 3) When the child finished '2a' exists.
  * 4) The parent will trap the exit code, if it's OK, will append the

- *    data accumulated into server.bgrewritebuf into the temp file, and
+ *    data accumulated into server.aof_rewrite_buf into the temp file, and

  *    finally will rename(2) the temp file in the actual file name.
  *    The the new file is reopened as the new append only file. Profit!
  */
 int rewriteAppendOnlyFileBackground(void) {
     pid_t childpid;
  *    finally will rename(2) the temp file in the actual file name.
  *    The the new file is reopened as the new append only file. Profit!
  */
 int rewriteAppendOnlyFileBackground(void) {
     pid_t childpid;

+    long long start;

 
 

-    if (server.bgrewritechildpid != -1) return REDIS_ERR;
-    if (server.vm_enabled) waitEmptyIOJobsQueue();
+    if (server.aof_child_pid != -1) return REDIS_ERR;
+    start = ustime();

     if ((childpid = fork()) == 0) {
     if ((childpid = fork()) == 0) {

-        /* Child */

         char tmpfile[256];
 
         char tmpfile[256];
 

-        if (server.vm_enabled) vmReopenSwapFile();
-        close(server.fd);
+        /* Child */
+        if (server.ipfd > 0) close(server.ipfd);
+        if (server.sofd > 0) close(server.sofd);

         snprintf(tmpfile,256,"temp-rewriteaof-bg-%d.aof", (int) getpid());
         if (rewriteAppendOnlyFile(tmpfile) == REDIS_OK) {
             _exit(0);
         snprintf(tmpfile,256,"temp-rewriteaof-bg-%d.aof", (int) getpid());
         if (rewriteAppendOnlyFile(tmpfile) == REDIS_OK) {
             _exit(0);


@@ -610,6 +809,7 @@ int rewriteAppendOnlyFileBackground(void) {
         }
     } else {
         /* Parent */
         }
     } else {
         /* Parent */

+        server.stat_fork_time = ustime()-start;

         if (childpid == -1) {
             redisLog(REDIS_WARNING,
                 "Can't rewrite append only file in background: fork: %s",
         if (childpid == -1) {
             redisLog(REDIS_WARNING,
                 "Can't rewrite append only file in background: fork: %s",


@@ -618,24 +818,26 @@ int rewriteAppendOnlyFileBackground(void) {
         }
         redisLog(REDIS_NOTICE,
             "Background append only file rewriting started by pid %d",childpid);
         }
         redisLog(REDIS_NOTICE,
             "Background append only file rewriting started by pid %d",childpid);

-        server.bgrewritechildpid = childpid;
+        server.aof_rewrite_scheduled = 0;
+        server.aof_child_pid = childpid;

         updateDictResizePolicy();
         /* We set appendseldb to -1 in order to force the next call to the
          * feedAppendOnlyFile() to issue a SELECT command, so the differences
         updateDictResizePolicy();
         /* We set appendseldb to -1 in order to force the next call to the
          * feedAppendOnlyFile() to issue a SELECT command, so the differences

-         * accumulated by the parent into server.bgrewritebuf will start
+         * accumulated by the parent into server.aof_rewrite_buf will start

          * with a SELECT statement and it will be safe to merge. */
          * with a SELECT statement and it will be safe to merge. */

-        server.appendseldb = -1;
+        server.aof_selected_db = -1;

         return REDIS_OK;
     }
     return REDIS_OK; /* unreached */
 }
 
 void bgrewriteaofCommand(redisClient *c) {
         return REDIS_OK;
     }
     return REDIS_OK; /* unreached */
 }
 
 void bgrewriteaofCommand(redisClient *c) {

-    if (server.bgrewritechildpid != -1) {
+    if (server.aof_child_pid != -1) {

         addReplyError(c,"Background append only file rewriting already in progress");
         addReplyError(c,"Background append only file rewriting already in progress");

-        return;
-    }
-    if (rewriteAppendOnlyFileBackground() == REDIS_OK) {
+    } else if (server.rdb_child_pid != -1) {
+        server.aof_rewrite_scheduled = 1;
+        addReplyStatus(c,"Background append only file rewriting scheduled");
+    } else if (rewriteAppendOnlyFileBackground() == REDIS_OK) {

         addReplyStatus(c,"Background append only file rewriting started");
     } else {
         addReply(c,shared.err);
         addReplyStatus(c,"Background append only file rewriting started");
     } else {
         addReply(c,shared.err);


@@ -649,64 +851,155 @@ void aofRemoveTempFile(pid_t childpid) {
     unlink(tmpfile);
 }
 
     unlink(tmpfile);
 }
 

+/* Update the server.aof_current_size filed explicitly using stat(2)
+ * to check the size of the file. This is useful after a rewrite or after
+ * a restart, normally the size is updated just adding the write length
+ * to the current length, that is much faster. */
+void aofUpdateCurrentSize(void) {
+    struct redis_stat sb;
+
+    if (redis_fstat(server.aof_fd,&sb) == -1) {
+        redisLog(REDIS_WARNING,"Unable to obtain the AOF file length. stat: %s",
+            strerror(errno));
+    } else {
+        server.aof_current_size = sb.st_size;
+    }
+}
+

 /* A background append only file rewriting (BGREWRITEAOF) terminated its work.
  * Handle this. */
 /* A background append only file rewriting (BGREWRITEAOF) terminated its work.
  * Handle this. */

-void backgroundRewriteDoneHandler(int statloc) {
-    int exitcode = WEXITSTATUS(statloc);
-    int bysignal = WIFSIGNALED(statloc);
-
+void backgroundRewriteDoneHandler(int exitcode, int bysignal) {

     if (!bysignal && exitcode == 0) {
     if (!bysignal && exitcode == 0) {

-        int fd;
+        int newfd, oldfd;
+        int nwritten;

         char tmpfile[256];
         char tmpfile[256];

+        long long now = ustime();

 
         redisLog(REDIS_NOTICE,
 
         redisLog(REDIS_NOTICE,

-            "Background append only file rewriting terminated with success");
-        /* Now it's time to flush the differences accumulated by the parent */
-        snprintf(tmpfile,256,"temp-rewriteaof-bg-%d.aof", (int) server.bgrewritechildpid);
-        fd = open(tmpfile,O_WRONLY|O_APPEND);
-        if (fd == -1) {
-            redisLog(REDIS_WARNING, "Not able to open the temp append only file produced by the child: %s", strerror(errno));
+            "Background AOF rewrite terminated with success");
+
+        /* Flush the differences accumulated by the parent to the
+         * rewritten AOF. */
+        snprintf(tmpfile,256,"temp-rewriteaof-bg-%d.aof",
+            (int)server.aof_child_pid);
+        newfd = open(tmpfile,O_WRONLY|O_APPEND);
+        if (newfd == -1) {
+            redisLog(REDIS_WARNING,
+                "Unable to open the temporary AOF produced by the child: %s", strerror(errno));

             goto cleanup;
         }
             goto cleanup;
         }

-        /* Flush our data... */
-        if (write(fd,server.bgrewritebuf,sdslen(server.bgrewritebuf)) !=
-                (signed) sdslen(server.bgrewritebuf)) {
-            redisLog(REDIS_WARNING, "Error or short write trying to flush the parent diff of the append log file in the child temp file: %s", strerror(errno));
-            close(fd);
+
+        nwritten = write(newfd,server.aof_rewrite_buf,sdslen(server.aof_rewrite_buf));
+        if (nwritten != (signed)sdslen(server.aof_rewrite_buf)) {
+            if (nwritten == -1) {
+                redisLog(REDIS_WARNING,
+                    "Error trying to flush the parent diff to the rewritten AOF: %s", strerror(errno));
+            } else {
+                redisLog(REDIS_WARNING,
+                    "Short write trying to flush the parent diff to the rewritten AOF: %s", strerror(errno));
+            }
+            close(newfd);

             goto cleanup;
         }
             goto cleanup;
         }

-        redisLog(REDIS_NOTICE,"Parent diff flushed into the new append log file with success (%lu bytes)",sdslen(server.bgrewritebuf));
-        /* Now our work is to rename the temp file into the stable file. And
-         * switch the file descriptor used by the server for append only. */
-        if (rename(tmpfile,server.appendfilename) == -1) {
-            redisLog(REDIS_WARNING,"Can't rename the temp append only file into the stable one: %s", strerror(errno));
-            close(fd);
+
+        redisLog(REDIS_NOTICE,
+            "Parent diff successfully flushed to the rewritten AOF (%lu bytes)", nwritten);
+
+        /* The only remaining thing to do is to rename the temporary file to
+         * the configured file and switch the file descriptor used to do AOF
+         * writes. We don't want close(2) or rename(2) calls to block the
+         * server on old file deletion.
+         *
+         * There are two possible scenarios:
+         *
+         * 1) AOF is DISABLED and this was a one time rewrite. The temporary
+         * file will be renamed to the configured file. When this file already
+         * exists, it will be unlinked, which may block the server.
+         *
+         * 2) AOF is ENABLED and the rewritten AOF will immediately start
+         * receiving writes. After the temporary file is renamed to the
+         * configured file, the original AOF file descriptor will be closed.
+         * Since this will be the last reference to that file, closing it
+         * causes the underlying file to be unlinked, which may block the
+         * server.
+         *
+         * To mitigate the blocking effect of the unlink operation (either
+         * caused by rename(2) in scenario 1, or by close(2) in scenario 2), we
+         * use a background thread to take care of this. First, we
+         * make scenario 1 identical to scenario 2 by opening the target file
+         * when it exists. The unlink operation after the rename(2) will then
+         * be executed upon calling close(2) for its descriptor. Everything to
+         * guarantee atomicity for this switch has already happened by then, so
+         * we don't care what the outcome or duration of that close operation
+         * is, as long as the file descriptor is released again. */
+        if (server.aof_fd == -1) {
+            /* AOF disabled */
+
+             /* Don't care if this fails: oldfd will be -1 and we handle that.
+              * One notable case of -1 return is if the old file does
+              * not exist. */
+             oldfd = open(server.aof_filename,O_RDONLY|O_NONBLOCK);
+        } else {
+            /* AOF enabled */
+            oldfd = -1; /* We'll set this to the current AOF filedes later. */
+        }
+
+        /* Rename the temporary file. This will not unlink the target file if
+         * it exists, because we reference it with "oldfd". */
+        if (rename(tmpfile,server.aof_filename) == -1) {
+            redisLog(REDIS_WARNING,
+                "Error trying to rename the temporary AOF file: %s", strerror(errno));
+            close(newfd);
+            if (oldfd != -1) close(oldfd);

             goto cleanup;
         }
             goto cleanup;
         }

-        /* Mission completed... almost */
-        redisLog(REDIS_NOTICE,"Append only file successfully rewritten.");
-        if (server.appendfd != -1) {
-            /* If append only is actually enabled... */
-            close(server.appendfd);
-            server.appendfd = fd;
-            if (server.appendfsync != APPENDFSYNC_NO) aof_fsync(fd);
-            server.appendseldb = -1; /* Make sure it will issue SELECT */
-            redisLog(REDIS_NOTICE,"The new append only file was selected for future appends.");
+
+        if (server.aof_fd == -1) {
+            /* AOF disabled, we don't need to set the AOF file descriptor
+             * to this new file, so we can close it. */
+            close(newfd);

         } else {
         } else {

-            /* If append only is disabled we just generate a dump in this
-             * format. Why not? */
-            close(fd);
+            /* AOF enabled, replace the old fd with the new one. */
+            oldfd = server.aof_fd;
+            server.aof_fd = newfd;
+            if (server.aof_fsync == AOF_FSYNC_ALWAYS)
+                aof_fsync(newfd);
+            else if (server.aof_fsync == AOF_FSYNC_EVERYSEC)
+                aof_background_fsync(newfd);
+            server.aof_selected_db = -1; /* Make sure SELECT is re-issued */
+            aofUpdateCurrentSize();
+            server.aof_rewrite_base_size = server.aof_current_size;
+
+            /* Clear regular AOF buffer since its contents was just written to
+             * the new AOF from the background rewrite buffer. */
+            sdsfree(server.aof_buf);
+            server.aof_buf = sdsempty();

         }
         }

+
+        redisLog(REDIS_NOTICE, "Background AOF rewrite finished successfully");
+        /* Change state from WAIT_REWRITE to ON if needed */
+        if (server.aof_state == REDIS_AOF_WAIT_REWRITE)
+            server.aof_state = REDIS_AOF_ON;
+
+        /* Asynchronously close the overwritten AOF. */
+        if (oldfd != -1) bioCreateBackgroundJob(REDIS_BIO_CLOSE_FILE,(void*)(long)oldfd,NULL,NULL);
+
+        redisLog(REDIS_VERBOSE,
+            "Background AOF rewrite signal handler took %lldus", ustime()-now);

     } else if (!bysignal && exitcode != 0) {
     } else if (!bysignal && exitcode != 0) {

-        redisLog(REDIS_WARNING, "Background append only file rewriting error");
+        redisLog(REDIS_WARNING,
+            "Background AOF rewrite terminated with error");

     } else {
         redisLog(REDIS_WARNING,
     } else {
         redisLog(REDIS_WARNING,

-            "Background append only file rewriting terminated by signal %d",
-            WTERMSIG(statloc));
+            "Background AOF rewrite terminated by signal %d", bysignal);

     }
     }

+

 cleanup:
 cleanup:

-    sdsfree(server.bgrewritebuf);
-    server.bgrewritebuf = sdsempty();
-    aofRemoveTempFile(server.bgrewritechildpid);
-    server.bgrewritechildpid = -1;
+    sdsfree(server.aof_rewrite_buf);
+    server.aof_rewrite_buf = sdsempty();
+    aofRemoveTempFile(server.aof_child_pid);
+    server.aof_child_pid = -1;
+    /* Schedule a new rewrite if we are waiting for it to switch the AOF ON. */
+    if (server.aof_state == REDIS_AOF_WAIT_REWRITE)
+        server.aof_rewrite_scheduled = 1;

 }
 }