aof.c 30.5 KB
Newer Older
1
#include "redis.h"
2
#include "bio.h"
3 4 5 6

#include <signal.h>
#include <fcntl.h>
#include <sys/stat.h>
7 8 9 10
#include <sys/types.h>
#include <sys/time.h>
#include <sys/resource.h>
#include <sys/wait.h>
11

12 13
void aofUpdateCurrentSize(void);

14 15 16 17
void aof_background_fsync(int fd) {
    bioCreateBackgroundJob(REDIS_BIO_AOF_FSYNC,(void*)(long)fd,NULL);
}

18 19 20 21 22 23 24 25 26 27 28
/* 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;
    /* rewrite operation in progress? kill it, wait child exit */
29
    if (server.bgrewritechildpid != -1) {
30 31
        int statloc;

32
        if (kill(server.bgrewritechildpid,SIGKILL) != -1)
33 34 35 36
            wait3(&statloc,0,NULL);
        /* reset the buffer accumulating changes while the child saves */
        sdsfree(server.bgrewritebuf);
        server.bgrewritebuf = sdsempty();
37
        server.bgrewritechildpid = -1;
38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76
    }
}

/* 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));
        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));
        return REDIS_ERR;
    }
    return REDIS_OK;
}

/* Write the append only file buffer on disk.
 *
 * Since we are required to write the AOF before replying to the client,
 * 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) {
    ssize_t nwritten;

    if (sdslen(server.aofbuf) == 0) return;

    /* 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 */
P
Pieter Noordhuis 已提交
77 78
    nwritten = write(server.appendfd,server.aofbuf,sdslen(server.aofbuf));
    if (nwritten != (signed)sdslen(server.aofbuf)) {
79 80 81
        /* 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. */
P
Pieter Noordhuis 已提交
82
        if (nwritten == -1) {
83
            redisLog(REDIS_WARNING,"Exiting on error writing to the append-only file: %s",strerror(errno));
P
Pieter Noordhuis 已提交
84
        } else {
85
            redisLog(REDIS_WARNING,"Exiting on short write while writing to the append-only file: %s",strerror(errno));
P
Pieter Noordhuis 已提交
86 87
        }
        exit(1);
88
    }
89
    server.appendonly_current_size += nwritten;
90

91 92 93 94 95 96 97 98 99
    /* 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.aofbuf)+sdsavail(server.aofbuf)) < 4000) {
        sdsclear(server.aofbuf);
    } else {
        sdsfree(server.aofbuf);
        server.aofbuf = sdsempty();
    }

100 101
    /* Don't fsync if no-appendfsync-on-rewrite is set to yes and there are
     * children doing I/O in the background. */
102 103 104
    if (server.no_appendfsync_on_rewrite &&
        (server.bgrewritechildpid != -1 || server.bgsavechildpid != -1))
            return;
105 106

    /* Perform the fsync if needed. */
107 108
    if (server.appendfsync == APPENDFSYNC_ALWAYS ||
        (server.appendfsync == APPENDFSYNC_EVERYSEC &&
109
         server.unixtime > server.lastfsync))
110 111 112 113
    {
        /* 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 */
114
        server.lastfsync = server.unixtime;
115 116 117
    }
}

118 119 120 121 122 123 124 125 126 127 128
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);

129
    for (j = 0; j < argc; j++) {
130 131 132 133 134 135 136 137
        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);
138 139
        decrRefCount(o);
    }
140
    return dst;
141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217
}

sds catAppendOnlyExpireAtCommand(sds buf, robj *key, robj *seconds) {
    int argc = 3;
    long when;
    robj *argv[3];

    /* Make sure we can use strtol */
    seconds = getDecodedObject(seconds);
    when = time(NULL)+strtol(seconds->ptr,NULL,10);
    decrRefCount(seconds);

    argv[0] = createStringObject("EXPIREAT",8);
    argv[1] = key;
    argv[2] = createObject(REDIS_STRING,
        sdscatprintf(sdsempty(),"%ld",when));
    buf = catAppendOnlyGenericCommand(buf, argc, argv);
    decrRefCount(argv[0]);
    decrRefCount(argv[2]);
    return buf;
}

void feedAppendOnlyFile(struct redisCommand *cmd, int dictid, robj **argv, int argc) {
    sds buf = sdsempty();
    robj *tmpargv[3];

    /* 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) {
        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;
    }

    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 */
        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]);
    } else {
        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 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));

    sdsfree(buf);
}

/* In Redis commands are always executed in the context of a client, so in
 * order to load the append only file we need to create a fake client. */
struct redisClient *createFakeClient(void) {
    struct redisClient *c = zmalloc(sizeof(*c));

    selectDb(c,0);
    c->fd = -1;
    c->querybuf = sdsempty();
    c->argc = 0;
    c->argv = NULL;
218
    c->bufpos = 0;
219 220 221 222 223
    c->flags = 0;
    /* We set the fake client as a slave waiting for the synchronization
     * so that Redis will not try to send replies to this client. */
    c->replstate = REDIS_REPL_WAIT_BGSAVE_START;
    c->reply = listCreate();
224
    c->watched_keys = listCreate();
225 226 227 228 229 230 231 232 233
    listSetFreeMethod(c->reply,decrRefCount);
    listSetDupMethod(c->reply,dupClientReplyValue);
    initClientMultiState(c);
    return c;
}

void freeFakeClient(struct redisClient *c) {
    sdsfree(c->querybuf);
    listRelease(c->reply);
234
    listRelease(c->watched_keys);
235 236 237 238 239 240 241 242 243 244 245 246
    freeClientMultiState(c);
    zfree(c);
}

/* Replay the append log file. On error REDIS_OK is returned. On non fatal
 * error (the append only file is zero-length) REDIS_ERR is returned. On
 * fatal error an error message is logged and the program exists. */
int loadAppendOnlyFile(char *filename) {
    struct redisClient *fakeClient;
    FILE *fp = fopen(filename,"r");
    struct redis_stat sb;
    int appendonly = server.appendonly;
247
    long loops = 0;
248

A
antirez 已提交
249
    if (fp && redis_fstat(fileno(fp),&sb) != -1 && sb.st_size == 0) {
250
        server.appendonly_current_size = 0;
A
antirez 已提交
251
        fclose(fp);
252
        return REDIS_ERR;
A
antirez 已提交
253
    }
254 255 256 257 258 259 260 261 262 263 264

    if (fp == NULL) {
        redisLog(REDIS_WARNING,"Fatal error: can't open the append log file for reading: %s",strerror(errno));
        exit(1);
    }

    /* Temporarily disable AOF, to prevent EXEC from feeding a MULTI
     * to the same file we're about to read. */
    server.appendonly = 0;

    fakeClient = createFakeClient();
265 266
    startLoading(fp);

267 268 269 270 271 272 273 274
    while(1) {
        int argc, j;
        unsigned long len;
        robj **argv;
        char buf[128];
        sds argsds;
        struct redisCommand *cmd;

275 276 277 278 279 280
        /* Serve the clients from time to time */
        if (!(loops++ % 1000)) {
            loadingProgress(ftello(fp));
            aeProcessEvents(server.el, AE_FILE_EVENTS|AE_DONT_WAIT);
        }

281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309
        if (fgets(buf,sizeof(buf),fp) == NULL) {
            if (feof(fp))
                break;
            else
                goto readerr;
        }
        if (buf[0] != '*') goto fmterr;
        argc = atoi(buf+1);
        argv = zmalloc(sizeof(robj*)*argc);
        for (j = 0; j < argc; j++) {
            if (fgets(buf,sizeof(buf),fp) == NULL) goto readerr;
            if (buf[0] != '$') goto fmterr;
            len = strtol(buf+1,NULL,10);
            argsds = sdsnewlen(NULL,len);
            if (len && fread(argsds,len,1,fp) == 0) goto fmterr;
            argv[j] = createObject(REDIS_STRING,argsds);
            if (fread(buf,2,1,fp) == 0) goto fmterr; /* discard CRLF */
        }

        /* Command lookup */
        cmd = lookupCommand(argv[0]->ptr);
        if (!cmd) {
            redisLog(REDIS_WARNING,"Unknown command '%s' reading the append only file", argv[0]->ptr);
            exit(1);
        }
        /* Run the command in the context of a fake client */
        fakeClient->argc = argc;
        fakeClient->argv = argv;
        cmd->proc(fakeClient);
310 311 312

        /* The fake client should not have a reply */
        redisAssert(fakeClient->bufpos == 0 && listLength(fakeClient->reply) == 0);
313 314
        /* The fake client should never get blocked */
        redisAssert((fakeClient->flags & REDIS_BLOCKED) == 0);
315

316 317 318 319 320
        /* 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);
321 322 323 324 325 326 327 328 329
    }

    /* This point can only be reached when EOF is reached without errors.
     * If the client is in the middle of a MULTI/EXEC, log error and quit. */
    if (fakeClient->flags & REDIS_MULTI) goto readerr;

    fclose(fp);
    freeFakeClient(fakeClient);
    server.appendonly = appendonly;
330
    stopLoading();
331
    aofUpdateCurrentSize();
332
    server.auto_aofrewrite_base_size = server.appendonly_current_size;
333 334 335 336 337 338 339 340 341 342
    return REDIS_OK;

readerr:
    if (feof(fp)) {
        redisLog(REDIS_WARNING,"Unexpected end of file reading the append only file");
    } else {
        redisLog(REDIS_WARNING,"Unrecoverable error reading the append only file: %s", strerror(errno));
    }
    exit(1);
fmterr:
A
antirez 已提交
343
    redisLog(REDIS_WARNING,"Bad file format reading the append only file: make a backup of your AOF file, then use ./redis-check-aof --fix <filename>");
344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369
    exit(1);
}

/* Write a sequence of commands able to fully rebuild the dataset into
 * "filename". Used both by REWRITEAOF and BGREWRITEAOF. */
int rewriteAppendOnlyFile(char *filename) {
    dictIterator *di = NULL;
    dictEntry *de;
    FILE *fp;
    char tmpfile[256];
    int j;
    time_t now = time(NULL);

    /* 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));
        return REDIS_ERR;
    }
    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;
370
        di = dictGetSafeIterator(d);
371 372 373 374 375 376 377 378 379 380 381
        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;

        /* Iterate this DB writing every entry */
        while((de = dictNext(di)) != NULL) {
A
antirez 已提交
382
            sds keystr;
383 384 385 386 387 388
            robj key, *o;
            time_t expiretime;

            keystr = dictGetEntryKey(de);
            o = dictGetEntryVal(de);
            initStaticStringObject(key,keystr);
389

390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438
            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;
                /* Key and value */
                if (fwriteBulkObject(fp,&key) == 0) goto werr;
                if (fwriteBulkObject(fp,o) == 0) goto werr;
            } 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");
                }
            } else if (o->type == REDIS_SET) {
439
                char cmd[]="*3\r\n$4\r\nSADD\r\n";
440

441 442 443
                /* Emit the SADDs needed to rebuild the set */
                if (o->encoding == REDIS_ENCODING_INTSET) {
                    int ii = 0;
444
                    int64_t llval;
445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461
                    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");
462 463 464
                }
            } else if (o->type == REDIS_ZSET) {
                /* Emit the ZADDs needed to rebuild the sorted set */
465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495
                char cmd[]="*4\r\n$4\r\nZADD\r\n";

                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 (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 (vstr != NULL) {
                            if (fwriteBulkString(fp,(char*)vstr,vlen) == 0)
                                goto werr;
                        } else {
                            if (fwriteBulkLongLong(fp,vll) == 0)
                                goto werr;
                        }
                        zzlNext(zl,&eptr,&sptr);
                    }
496
                } else if (o->encoding == REDIS_ENCODING_SKIPLIST) {
497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512
                    zset *zs = o->ptr;
                    dictIterator *di = dictGetIterator(zs->dict);
                    dictEntry *de;

                    while((de = dictNext(di)) != NULL) {
                        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);
                } else {
                    redisPanic("Unknown sorted set encoding");
513 514 515 516 517 518 519 520 521 522 523 524 525
                }
            } 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;
526
                        if (fwriteBulkString(fp,(char*)field,flen) == 0)
527
                            goto werr;
528
                        if (fwriteBulkString(fp,(char*)val,vlen) == 0)
529
                            goto werr;
530 531 532 533 534 535 536 537 538 539 540
                    }
                } 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;
541 542
                        if (fwriteBulkObject(fp,field) == 0) goto werr;
                        if (fwriteBulkObject(fp,val) == 0) goto werr;
543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598
                    }
                    dictReleaseIterator(di);
                }
            } else {
                redisPanic("Unknown object type");
            }
            /* Save the expire time */
            if (expiretime != -1) {
                char cmd[]="*3\r\n$8\r\nEXPIREAT\r\n";
                /* 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;
            }
        }
        dictReleaseIterator(di);
    }

    /* Make sure data will not remain on the OS's output buffers */
    fflush(fp);
    aof_fsync(fileno(fp));
    fclose(fp);

    /* Use RENAME to make sure the DB file is changed atomically only
     * if the generate DB file is ok. */
    if (rename(tmpfile,filename) == -1) {
        redisLog(REDIS_WARNING,"Error moving temp append only file on the final destination: %s", strerror(errno));
        unlink(tmpfile);
        return REDIS_ERR;
    }
    redisLog(REDIS_NOTICE,"SYNC append only file rewrite performed");
    return REDIS_OK;

werr:
    fclose(fp);
    unlink(tmpfile);
    redisLog(REDIS_WARNING,"Write error writing append only file on disk: %s", strerror(errno));
    if (di) dictReleaseIterator(di);
    return REDIS_ERR;
}

/* This is how rewriting of the append only file in background works:
 *
 * 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.
 * 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
 *    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;
599
    long long start;
600 601

    if (server.bgrewritechildpid != -1) return REDIS_ERR;
602
    start = ustime();
603 604 605
    if ((childpid = fork()) == 0) {
        char tmpfile[256];

606
        /* Child */
607 608
        if (server.ipfd > 0) close(server.ipfd);
        if (server.sofd > 0) close(server.sofd);
609 610 611 612 613 614 615 616
        snprintf(tmpfile,256,"temp-rewriteaof-bg-%d.aof", (int) getpid());
        if (rewriteAppendOnlyFile(tmpfile) == REDIS_OK) {
            _exit(0);
        } else {
            _exit(1);
        }
    } else {
        /* Parent */
617
        server.stat_fork_time = ustime()-start;
618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639
        if (childpid == -1) {
            redisLog(REDIS_WARNING,
                "Can't rewrite append only file in background: fork: %s",
                strerror(errno));
            return REDIS_ERR;
        }
        redisLog(REDIS_NOTICE,
            "Background append only file rewriting started by pid %d",childpid);
        server.bgrewritechildpid = 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
         * accumulated by the parent into server.bgrewritebuf will start
         * with a SELECT statement and it will be safe to merge. */
        server.appendseldb = -1;
        return REDIS_OK;
    }
    return REDIS_OK; /* unreached */
}

void bgrewriteaofCommand(redisClient *c) {
    if (server.bgrewritechildpid != -1) {
640
        addReplyError(c,"Background append only file rewriting already in progress");
641 642
    } else if (server.bgsavechildpid != -1) {
        server.aofrewrite_scheduled = 1;
643
        addReplyStatus(c,"Background append only file rewriting scheduled");
644
    } else if (rewriteAppendOnlyFileBackground() == REDIS_OK) {
645
        addReplyStatus(c,"Background append only file rewriting started");
646 647 648 649 650 651 652 653 654 655 656 657
    } else {
        addReply(c,shared.err);
    }
}

void aofRemoveTempFile(pid_t childpid) {
    char tmpfile[256];

    snprintf(tmpfile,256,"temp-rewriteaof-bg-%d.aof", (int) childpid);
    unlink(tmpfile);
}

658 659 660 661 662 663 664 665 666 667 668 669 670 671 672
/* Update the server.appendonly_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 lenght, that is much faster. */
void aofUpdateCurrentSize(void) {
    struct redis_stat sb;

    if (redis_fstat(server.appendfd,&sb) == -1) {
        redisLog(REDIS_WARNING,"Unable to check the AOF length: %s",
            strerror(errno));
    } else {
        server.appendonly_current_size = sb.st_size;
    }
}

673 674
/* A background append only file rewriting (BGREWRITEAOF) terminated its work.
 * Handle this. */
675
void backgroundRewriteDoneHandler(int exitcode, int bysignal) {
676
    if (!bysignal && exitcode == 0) {
677 678
        int newfd, oldfd;
        int nwritten;
679
        char tmpfile[256];
680
        long long now = ustime();
681 682

        redisLog(REDIS_NOTICE,
683 684
            "Background AOF rewrite terminated with success");

685 686 687 688
        /* Flush the differences accumulated by the parent to the
         * rewritten AOF. */
        snprintf(tmpfile,256,"temp-rewriteaof-bg-%d.aof",
            (int)server.bgrewritechildpid);
689 690 691 692
        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));
693 694
            goto cleanup;
        }
695 696 697 698 699 700 701 702 703 704 705

        nwritten = write(newfd,server.bgrewritebuf,sdslen(server.bgrewritebuf));
        if (nwritten != (signed)sdslen(server.bgrewritebuf)) {
            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);
706 707
            goto cleanup;
        }
708 709 710 711 712 713

        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
714 715 716 717
         * writes. We don't want close(2) or rename(2) calls to block the
         * server on old file deletion.
         *
         * There are two possible scenarios:
718 719 720 721 722 723 724 725 726 727 728 729 730 731
         *
         * 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
732
         * use a background thread to take care of this. First, we
733 734 735 736 737 738 739 740 741
         * 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.appendfd == -1) {
            /* AOF disabled */

742 743 744 745
             /* 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.appendfilename,O_RDONLY|O_NONBLOCK);
746 747
        } else {
            /* AOF enabled */
748
            oldfd = -1; /* We'll set this to the current AOF filedes later. */
749 750 751 752
        }

        /* Rename the temporary file. This will not unlink the target file if
         * it exists, because we reference it with "oldfd". */
753
        if (rename(tmpfile,server.appendfilename) == -1) {
754 755 756
            redisLog(REDIS_WARNING,
                "Error trying to rename the temporary AOF: %s", strerror(errno));
            close(newfd);
757
            if (oldfd != -1) close(oldfd);
758 759
            goto cleanup;
        }
760 761

        if (server.appendfd == -1) {
762 763
            /* AOF disabled, we don't need to set the AOF file descriptor
             * to this new file, so we can close it. */
764 765
            close(newfd);
        } else {
766
            /* AOF enabled, replace the old fd with the new one. */
767 768
            oldfd = server.appendfd;
            server.appendfd = newfd;
769 770 771 772
            if (server.appendfsync == APPENDFSYNC_ALWAYS)
                aof_fsync(newfd);
            else if (server.appendfsync == APPENDFSYNC_EVERYSEC)
                aof_background_fsync(newfd);
773
            server.appendseldb = -1; /* Make sure SELECT is re-issued */
774
            aofUpdateCurrentSize();
775
            server.auto_aofrewrite_base_size = server.appendonly_current_size;
776
        }
777 778 779 780

        redisLog(REDIS_NOTICE, "Background AOF rewrite successful");

        /* Asynchronously close the overwritten AOF. */
781
        if (oldfd != -1) bioCreateBackgroundJob(REDIS_BIO_CLOSE_FILE,(void*)(long)oldfd,NULL,NULL);
782 783 784

        redisLog(REDIS_VERBOSE,
            "Background AOF rewrite signal handler took %lldus", ustime()-now);
785
    } else if (!bysignal && exitcode != 0) {
786 787
        redisLog(REDIS_WARNING,
            "Background AOF rewrite terminated with error");
788 789
    } else {
        redisLog(REDIS_WARNING,
790
            "Background AOF rewrite terminated by signal %d", bysignal);
791
    }
792

793 794 795 796 797 798
cleanup:
    sdsfree(server.bgrewritebuf);
    server.bgrewritebuf = sdsempty();
    aofRemoveTempFile(server.bgrewritechildpid);
    server.bgrewritechildpid = -1;
}