redis.c 106.4 KB
Newer Older
1
/*
2
 * Copyright (c) 2009-2012, Salvatore Sanfilippo <antirez at gmail dot com>
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions are met:
 *
 *   * Redistributions of source code must retain the above copyright notice,
 *     this list of conditions and the following disclaimer.
 *   * Redistributions in binary form must reproduce the above copyright
 *     notice, this list of conditions and the following disclaimer in the
 *     documentation and/or other materials provided with the distribution.
 *   * Neither the name of Redis nor the names of its contributors may be used
 *     to endorse or promote products derived from this software without
 *     specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
 * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
 * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
 * POSSIBILITY OF SUCH DAMAGE.
 */

#include "redis.h"
31
#include "slowlog.h"
32
#include "bio.h"
33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49

#include <time.h>
#include <signal.h>
#include <sys/wait.h>
#include <errno.h>
#include <assert.h>
#include <ctype.h>
#include <stdarg.h>
#include <arpa/inet.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <sys/time.h>
#include <sys/resource.h>
#include <sys/uio.h>
#include <limits.h>
#include <float.h>
#include <math.h>
50
#include <sys/resource.h>
51
#include <sys/utsname.h>
52 53 54 55 56

/* Our shared "common" objects */

struct sharedObjectsStruct shared;

57
/* Global vars that are actually used as constants. The following double
58 59 60 61 62 63 64 65 66 67
 * values are used for double on-disk serialization, and are initialized
 * at runtime to avoid strange compiler optimizations. */

double R_Zero, R_PosInf, R_NegInf, R_Nan;

/*================================= Globals ================================= */

/* Global vars */
struct redisServer server; /* server global state */
struct redisCommand *commandTable;
68

69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93
/* Our command table.
 *
 * Every entry is composed of the following fields:
 *
 * name: a string representing the command name.
 * function: pointer to the C function implementing the command.
 * arity: number of arguments, it is possible to use -N to say >= N
 * sflags: command flags as string. See below for a table of flags.
 * flags: flags as bitmask. Computed by Redis using the 'sflags' field.
 * get_keys_proc: an optional function to get key arguments from a command.
 *                This is only used when the following three fields are not
 *                enough to specify what arguments are keys.
 * first_key_index: first argument that is a key
 * last_key_index: last argument that is a key
 * key_step: step to get all the keys from first to last argument. For instance
 *           in MSET the step is two since arguments are key,val,key,val,...
 * microseconds: microseconds of total execution time for this command.
 * calls: total number of calls of this command.
 *
 * The flags, microseconds and calls fields are computed by Redis and should
 * always be set to zero.
 *
 * Command flags are expressed using strings where every character represents
 * a flag. Later the populateCommandTable() function will take care of
 * populating the real 'flags' field using this characters.
94 95 96 97 98 99 100 101
 *
 * This is the meaning of the flags:
 *
 * w: write command (may modify the key space).
 * r: read command  (will never modify the key space).
 * m: may increase memory usage once called. Don't allow if out of memory.
 * a: admin command, like SAVE or SHUTDOWN.
 * p: Pub/Sub related command.
G
guiquanz 已提交
102
 * f: force replication of this command, regardless of server.dirty.
103
 * s: command not allowed in scripts.
104
 * R: random command. Command is not deterministic, that is, the same command
105
 *    with the same arguments, with the same key space, may have different
106 107 108
 *    results. For instance SPOP and RANDOMKEY are two random commands.
 * S: Sort command output array if called from script, so that the output
 *    is deterministic.
109 110 111 112
 * l: Allow command while loading the database.
 * t: Allow command while a slave has stale data but is not allowed to
 *    server this data. Normally no command is accepted in this condition
 *    but just a few.
113
 * M: Do not automatically propagate the command on MONITOR.
114 115
 * k: Perform an implicit ASKING for this command, so the command will be
 *    accepted in cluster mode if the slot is marked as 'importing'.
116
 */
117
struct redisCommand redisCommandTable[] = {
118 119 120
    {"get",getCommand,2,"r",0,NULL,1,1,1,0,0},
    {"set",setCommand,3,"wm",0,noPreloadGetKeys,1,1,1,0,0},
    {"setnx",setnxCommand,3,"wm",0,noPreloadGetKeys,1,1,1,0,0},
121 122
    {"setex",setexCommand,4,"wm",0,noPreloadGetKeys,1,1,1,0,0},
    {"psetex",psetexCommand,4,"wm",0,noPreloadGetKeys,1,1,1,0,0},
123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141
    {"append",appendCommand,3,"wm",0,NULL,1,1,1,0,0},
    {"strlen",strlenCommand,2,"r",0,NULL,1,1,1,0,0},
    {"del",delCommand,-2,"w",0,noPreloadGetKeys,1,-1,1,0,0},
    {"exists",existsCommand,2,"r",0,NULL,1,1,1,0,0},
    {"setbit",setbitCommand,4,"wm",0,NULL,1,1,1,0,0},
    {"getbit",getbitCommand,3,"r",0,NULL,1,1,1,0,0},
    {"setrange",setrangeCommand,4,"wm",0,NULL,1,1,1,0,0},
    {"getrange",getrangeCommand,4,"r",0,NULL,1,1,1,0,0},
    {"substr",getrangeCommand,4,"r",0,NULL,1,1,1,0,0},
    {"incr",incrCommand,2,"wm",0,NULL,1,1,1,0,0},
    {"decr",decrCommand,2,"wm",0,NULL,1,1,1,0,0},
    {"mget",mgetCommand,-2,"r",0,NULL,1,-1,1,0,0},
    {"rpush",rpushCommand,-3,"wm",0,NULL,1,1,1,0,0},
    {"lpush",lpushCommand,-3,"wm",0,NULL,1,1,1,0,0},
    {"rpushx",rpushxCommand,3,"wm",0,NULL,1,1,1,0,0},
    {"lpushx",lpushxCommand,3,"wm",0,NULL,1,1,1,0,0},
    {"linsert",linsertCommand,5,"wm",0,NULL,1,1,1,0,0},
    {"rpop",rpopCommand,2,"w",0,NULL,1,1,1,0,0},
    {"lpop",lpopCommand,2,"w",0,NULL,1,1,1,0,0},
A
antirez 已提交
142 143 144
    {"brpop",brpopCommand,-3,"ws",0,NULL,1,1,1,0,0},
    {"brpoplpush",brpoplpushCommand,4,"wms",0,NULL,1,2,1,0,0},
    {"blpop",blpopCommand,-3,"ws",0,NULL,1,-2,1,0,0},
145 146 147 148 149 150 151 152 153 154 155 156
    {"llen",llenCommand,2,"r",0,NULL,1,1,1,0,0},
    {"lindex",lindexCommand,3,"r",0,NULL,1,1,1,0,0},
    {"lset",lsetCommand,4,"wm",0,NULL,1,1,1,0,0},
    {"lrange",lrangeCommand,4,"r",0,NULL,1,1,1,0,0},
    {"ltrim",ltrimCommand,4,"w",0,NULL,1,1,1,0,0},
    {"lrem",lremCommand,4,"w",0,NULL,1,1,1,0,0},
    {"rpoplpush",rpoplpushCommand,3,"wm",0,NULL,1,2,1,0,0},
    {"sadd",saddCommand,-3,"wm",0,NULL,1,1,1,0,0},
    {"srem",sremCommand,-3,"w",0,NULL,1,1,1,0,0},
    {"smove",smoveCommand,4,"w",0,NULL,1,2,1,0,0},
    {"sismember",sismemberCommand,3,"r",0,NULL,1,1,1,0,0},
    {"scard",scardCommand,2,"r",0,NULL,1,1,1,0,0},
157
    {"spop",spopCommand,2,"wRs",0,NULL,1,1,1,0,0},
158
    {"srandmember",srandmemberCommand,-2,"rR",0,NULL,1,1,1,0,0},
159
    {"sinter",sinterCommand,-2,"rS",0,NULL,1,-1,1,0,0},
160
    {"sinterstore",sinterstoreCommand,-3,"wm",0,NULL,1,-1,1,0,0},
161
    {"sunion",sunionCommand,-2,"rS",0,NULL,1,-1,1,0,0},
162
    {"sunionstore",sunionstoreCommand,-3,"wm",0,NULL,1,-1,1,0,0},
163
    {"sdiff",sdiffCommand,-2,"rS",0,NULL,1,-1,1,0,0},
164
    {"sdiffstore",sdiffstoreCommand,-3,"wm",0,NULL,1,-1,1,0,0},
165
    {"smembers",sinterCommand,2,"rS",0,NULL,1,1,1,0,0},
166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187
    {"zadd",zaddCommand,-4,"wm",0,NULL,1,1,1,0,0},
    {"zincrby",zincrbyCommand,4,"wm",0,NULL,1,1,1,0,0},
    {"zrem",zremCommand,-3,"w",0,NULL,1,1,1,0,0},
    {"zremrangebyscore",zremrangebyscoreCommand,4,"w",0,NULL,1,1,1,0,0},
    {"zremrangebyrank",zremrangebyrankCommand,4,"w",0,NULL,1,1,1,0,0},
    {"zunionstore",zunionstoreCommand,-4,"wm",0,zunionInterGetKeys,0,0,0,0,0},
    {"zinterstore",zinterstoreCommand,-4,"wm",0,zunionInterGetKeys,0,0,0,0,0},
    {"zrange",zrangeCommand,-4,"r",0,NULL,1,1,1,0,0},
    {"zrangebyscore",zrangebyscoreCommand,-4,"r",0,NULL,1,1,1,0,0},
    {"zrevrangebyscore",zrevrangebyscoreCommand,-4,"r",0,NULL,1,1,1,0,0},
    {"zcount",zcountCommand,4,"r",0,NULL,1,1,1,0,0},
    {"zrevrange",zrevrangeCommand,-4,"r",0,NULL,1,1,1,0,0},
    {"zcard",zcardCommand,2,"r",0,NULL,1,1,1,0,0},
    {"zscore",zscoreCommand,3,"r",0,NULL,1,1,1,0,0},
    {"zrank",zrankCommand,3,"r",0,NULL,1,1,1,0,0},
    {"zrevrank",zrevrankCommand,3,"r",0,NULL,1,1,1,0,0},
    {"hset",hsetCommand,4,"wm",0,NULL,1,1,1,0,0},
    {"hsetnx",hsetnxCommand,4,"wm",0,NULL,1,1,1,0,0},
    {"hget",hgetCommand,3,"r",0,NULL,1,1,1,0,0},
    {"hmset",hmsetCommand,-4,"wm",0,NULL,1,1,1,0,0},
    {"hmget",hmgetCommand,-3,"r",0,NULL,1,1,1,0,0},
    {"hincrby",hincrbyCommand,4,"wm",0,NULL,1,1,1,0,0},
A
antirez 已提交
188
    {"hincrbyfloat",hincrbyfloatCommand,4,"wm",0,NULL,1,1,1,0,0},
189 190
    {"hdel",hdelCommand,-3,"w",0,NULL,1,1,1,0,0},
    {"hlen",hlenCommand,2,"r",0,NULL,1,1,1,0,0},
191 192
    {"hkeys",hkeysCommand,2,"rS",0,NULL,1,1,1,0,0},
    {"hvals",hvalsCommand,2,"rS",0,NULL,1,1,1,0,0},
193 194 195 196
    {"hgetall",hgetallCommand,2,"r",0,NULL,1,1,1,0,0},
    {"hexists",hexistsCommand,3,"r",0,NULL,1,1,1,0,0},
    {"incrby",incrbyCommand,3,"wm",0,NULL,1,1,1,0,0},
    {"decrby",decrbyCommand,3,"wm",0,NULL,1,1,1,0,0},
A
antirez 已提交
197
    {"incrbyfloat",incrbyfloatCommand,3,"wm",0,NULL,1,1,1,0,0},
198 199 200
    {"getset",getsetCommand,3,"wm",0,NULL,1,1,1,0,0},
    {"mset",msetCommand,-3,"wm",0,NULL,1,-1,2,0,0},
    {"msetnx",msetnxCommand,-3,"wm",0,NULL,1,-1,2,0,0},
201
    {"randomkey",randomkeyCommand,1,"rR",0,NULL,0,0,0,0,0},
202 203 204 205
    {"select",selectCommand,2,"r",0,NULL,0,0,0,0,0},
    {"move",moveCommand,3,"w",0,NULL,1,1,1,0,0},
    {"rename",renameCommand,3,"w",0,renameGetKeys,1,2,1,0,0},
    {"renamenx",renamenxCommand,3,"w",0,renameGetKeys,1,2,1,0,0},
206 207 208 209
    {"expire",expireCommand,3,"w",0,NULL,1,1,1,0,0},
    {"expireat",expireatCommand,3,"w",0,NULL,1,1,1,0,0},
    {"pexpire",pexpireCommand,3,"w",0,NULL,1,1,1,0,0},
    {"pexpireat",pexpireatCommand,3,"w",0,NULL,1,1,1,0,0},
210
    {"keys",keysCommand,2,"rS",0,NULL,0,0,0,0,0},
211
    {"dbsize",dbsizeCommand,1,"r",0,NULL,0,0,0,0,0},
212
    {"auth",authCommand,2,"rsl",0,NULL,0,0,0,0,0},
213 214
    {"ping",pingCommand,1,"r",0,NULL,0,0,0,0,0},
    {"echo",echoCommand,2,"r",0,NULL,0,0,0,0,0},
A
antirez 已提交
215
    {"save",saveCommand,1,"ars",0,NULL,0,0,0,0,0},
216 217
    {"bgsave",bgsaveCommand,1,"ar",0,NULL,0,0,0,0,0},
    {"bgrewriteaof",bgrewriteaofCommand,1,"ar",0,NULL,0,0,0,0,0},
218
    {"shutdown",shutdownCommand,-1,"ar",0,NULL,0,0,0,0,0},
A
antirez 已提交
219
    {"lastsave",lastsaveCommand,1,"rR",0,NULL,0,0,0,0,0},
220
    {"type",typeCommand,2,"r",0,NULL,1,1,1,0,0},
221
    {"multi",multiCommand,1,"rs",0,NULL,0,0,0,0,0},
222
    {"exec",execCommand,1,"sM",0,NULL,0,0,0,0,0},
223 224
    {"discard",discardCommand,1,"rs",0,NULL,0,0,0,0,0},
    {"sync",syncCommand,1,"ars",0,NULL,0,0,0,0,0},
225
    {"psync",syncCommand,3,"ars",0,NULL,0,0,0,0,0},
A
antirez 已提交
226
    {"replconf",replconfCommand,-1,"ars",0,NULL,0,0,0,0,0},
227 228
    {"flushdb",flushdbCommand,1,"w",0,NULL,0,0,0,0,0},
    {"flushall",flushallCommand,1,"w",0,NULL,0,0,0,0,0},
229
    {"sort",sortCommand,-2,"wm",0,NULL,1,1,1,0,0},
230
    {"info",infoCommand,-1,"rlt",0,NULL,0,0,0,0,0},
231
    {"monitor",monitorCommand,1,"ars",0,NULL,0,0,0,0,0},
232 233
    {"ttl",ttlCommand,2,"r",0,NULL,1,1,1,0,0},
    {"pttl",pttlCommand,2,"r",0,NULL,1,1,1,0,0},
234
    {"persist",persistCommand,2,"w",0,NULL,1,1,1,0,0},
235
    {"slaveof",slaveofCommand,3,"ast",0,NULL,0,0,0,0,0},
236
    {"debug",debugCommand,-2,"as",0,NULL,0,0,0,0,0},
237
    {"config",configCommand,-2,"ar",0,NULL,0,0,0,0,0},
238 239 240 241 242
    {"subscribe",subscribeCommand,-2,"rpslt",0,NULL,0,0,0,0,0},
    {"unsubscribe",unsubscribeCommand,-1,"rpslt",0,NULL,0,0,0,0,0},
    {"psubscribe",psubscribeCommand,-2,"rpslt",0,NULL,0,0,0,0,0},
    {"punsubscribe",punsubscribeCommand,-1,"rpslt",0,NULL,0,0,0,0,0},
    {"publish",publishCommand,3,"pflt",0,NULL,0,0,0,0,0},
243 244
    {"watch",watchCommand,-2,"rs",0,noPreloadGetKeys,1,-1,1,0,0},
    {"unwatch",unwatchCommand,1,"rs",0,NULL,0,0,0,0,0},
245
    {"cluster",clusterCommand,-2,"ar",0,NULL,0,0,0,0,0},
A
antirez 已提交
246
    {"restore",restoreCommand,-4,"awm",0,NULL,1,1,1,0,0},
247
    {"restore-asking",restoreCommand,-4,"awmk",0,NULL,1,1,1,0,0},
A
antirez 已提交
248
    {"migrate",migrateCommand,-6,"aw",0,NULL,0,0,0,0,0},
249
    {"asking",askingCommand,1,"r",0,NULL,0,0,0,0,0},
250 251
    {"dump",dumpCommand,2,"ar",0,NULL,1,1,1,0,0},
    {"object",objectCommand,-2,"r",0,NULL,2,2,2,0,0},
252
    {"client",clientCommand,-2,"ar",0,NULL,0,0,0,0,0},
253 254
    {"eval",evalCommand,-3,"s",0,zunionInterGetKeys,0,0,0,0,0},
    {"evalsha",evalShaCommand,-3,"s",0,zunionInterGetKeys,0,0,0,0,0},
255
    {"slowlog",slowlogCommand,-2,"r",0,NULL,0,0,0,0,0},
A
antirez 已提交
256
    {"script",scriptCommand,-2,"ras",0,NULL,0,0,0,0,0},
A
antirez 已提交
257 258 259
    {"time",timeCommand,1,"rR",0,NULL,0,0,0,0,0},
    {"bitop",bitopCommand,-4,"wm",0,NULL,2,-1,1,0,0},
    {"bitcount",bitcountCommand,-2,"r",0,NULL,1,1,1,0,0}
260 261 262 263
};

/*============================ Utility functions ============================ */

264 265 266
/* Low level logging. To use only for very big messages, otherwise
 * redisLog() is to prefer. */
void redisLogRaw(int level, const char *msg) {
J
Jonah H. Harris 已提交
267 268
    const int syslogLevelMap[] = { LOG_DEBUG, LOG_INFO, LOG_NOTICE, LOG_WARNING };
    const char *c = ".-*#";
269
    FILE *fp;
270
    char buf[64];
A
antirez 已提交
271
    int rawmode = (level & REDIS_LOG_RAW);
272

A
antirez 已提交
273
    level &= 0xff; /* clear flags */
274
    if (level < server.verbosity) return;
275 276 277 278

    fp = (server.logfile == NULL) ? stdout : fopen(server.logfile,"a");
    if (!fp) return;

A
antirez 已提交
279 280 281
    if (rawmode) {
        fprintf(fp,"%s",msg);
    } else {
282 283 284 285 286 287
        int off;
        struct timeval tv;

        gettimeofday(&tv,NULL);
        off = strftime(buf,sizeof(buf),"%d %b %H:%M:%S.",localtime(&tv.tv_sec));
        snprintf(buf+off,sizeof(buf)-off,"%03d",(int)tv.tv_usec/1000);
A
antirez 已提交
288 289
        fprintf(fp,"[%d] %s %c %s\n",(int)getpid(),buf,c[level],msg);
    }
J
Jonah H. Harris 已提交
290 291
    fflush(fp);

292
    if (server.logfile) fclose(fp);
J
Jonah H. Harris 已提交
293 294

    if (server.syslog_enabled) syslog(syslogLevelMap[level], "%s", msg);
295 296
}

G
guiquanz 已提交
297
/* Like redisLogRaw() but with printf-alike support. This is the function that
298 299 300 301 302 303
 * is used across the code. The raw version is only used in order to dump
 * the INFO output on crash. */
void redisLog(int level, const char *fmt, ...) {
    va_list ap;
    char msg[REDIS_MAX_LOGMSG_LEN];

A
antirez 已提交
304
    if ((level&0xff) < server.verbosity) return;
305 306 307 308 309 310 311 312

    va_start(ap, fmt);
    vsnprintf(msg, sizeof(msg), fmt, ap);
    va_end(ap);

    redisLogRaw(level,msg);
}

A
antirez 已提交
313 314 315 316 317 318 319 320 321 322 323 324 325 326
/* Log a fixed message without printf-alike capabilities, in a way that is
 * safe to call from a signal handler.
 *
 * We actually use this only for signals that are not fatal from the point
 * of view of Redis. Signals that are going to kill the server anyway and
 * where we need printf-alike features are served by redisLog(). */
void redisLogFromHandler(int level, const char *msg) {
    int fd;
    char buf[64];

    if ((level&0xff) < server.verbosity ||
        (server.logfile == NULL && server.daemonize)) return;
    fd = server.logfile ?
        open(server.logfile, O_APPEND|O_CREAT|O_WRONLY, 0644) :
A
antirez 已提交
327
        STDOUT_FILENO;
A
antirez 已提交
328 329
    if (fd == -1) return;
    ll2string(buf,sizeof(buf),getpid());
330 331 332
    if (write(fd,"[",1) == -1) goto err;
    if (write(fd,buf,strlen(buf)) == -1) goto err;
    if (write(fd," | signal handler] (",20) == -1) goto err;
A
antirez 已提交
333
    ll2string(buf,sizeof(buf),time(NULL));
334 335 336 337 338
    if (write(fd,buf,strlen(buf)) == -1) goto err;
    if (write(fd,") ",2) == -1) goto err;
    if (write(fd,msg,strlen(msg)) == -1) goto err;
    if (write(fd,"\n",1) == -1) goto err;
err:
A
antirez 已提交
339
    if (server.logfile) close(fd);
A
antirez 已提交
340 341
}

A
antirez 已提交
342 343 344 345 346 347 348 349 350 351 352
/* Return the UNIX time in microseconds */
long long ustime(void) {
    struct timeval tv;
    long long ust;

    gettimeofday(&tv, NULL);
    ust = ((long long)tv.tv_sec)*1000000;
    ust += tv.tv_usec;
    return ust;
}

353 354 355 356 357
/* Return the UNIX time in milliseconds */
long long mstime(void) {
    return ustime()/1000;
}

358 359 360 361 362 363 364 365 366 367 368 369
/* After an RDB dump or AOF rewrite we exit from children using _exit() instead of
 * exit(), because the latter may interact with the same file objects used by
 * the parent process. However if we are testing the coverage normal exit() is
 * used in order to obtain the right coverage information. */
void exitFromChild(int retcode) {
#ifdef COVERAGE_TEST
    exit(retcode);
#else
    _exit(retcode);
#endif
}

370 371
/*====================== Hash table type implementation  ==================== */

G
guiquanz 已提交
372
/* This is an hash table type that uses the SDS dynamic strings library as
373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399
 * keys and radis objects as values (objects can hold SDS strings,
 * lists, sets). */

void dictVanillaFree(void *privdata, void *val)
{
    DICT_NOTUSED(privdata);
    zfree(val);
}

void dictListDestructor(void *privdata, void *val)
{
    DICT_NOTUSED(privdata);
    listRelease((list*)val);
}

int dictSdsKeyCompare(void *privdata, const void *key1,
        const void *key2)
{
    int l1,l2;
    DICT_NOTUSED(privdata);

    l1 = sdslen((sds)key1);
    l2 = sdslen((sds)key2);
    if (l1 != l2) return 0;
    return memcmp(key1, key2, l1) == 0;
}

A
antirez 已提交
400 401
/* A case insensitive version used for the command lookup table and other
 * places where case insensitive non binary-safe comparison is needed. */
402 403 404 405 406 407 408 409
int dictSdsKeyCaseCompare(void *privdata, const void *key1,
        const void *key2)
{
    DICT_NOTUSED(privdata);

    return strcasecmp(key1, key2) == 0;
}

410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440
void dictRedisObjectDestructor(void *privdata, void *val)
{
    DICT_NOTUSED(privdata);

    if (val == NULL) return; /* Values of swapped out keys as set to NULL */
    decrRefCount(val);
}

void dictSdsDestructor(void *privdata, void *val)
{
    DICT_NOTUSED(privdata);

    sdsfree(val);
}

int dictObjKeyCompare(void *privdata, const void *key1,
        const void *key2)
{
    const robj *o1 = key1, *o2 = key2;
    return dictSdsKeyCompare(privdata,o1->ptr,o2->ptr);
}

unsigned int dictObjHash(const void *key) {
    const robj *o = key;
    return dictGenHashFunction(o->ptr, sdslen((sds)o->ptr));
}

unsigned int dictSdsHash(const void *key) {
    return dictGenHashFunction((unsigned char*)key, sdslen((char*)key));
}

441 442 443 444
unsigned int dictSdsCaseHash(const void *key) {
    return dictGenCaseHashFunction((unsigned char*)key, sdslen((char*)key));
}

445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485
int dictEncObjKeyCompare(void *privdata, const void *key1,
        const void *key2)
{
    robj *o1 = (robj*) key1, *o2 = (robj*) key2;
    int cmp;

    if (o1->encoding == REDIS_ENCODING_INT &&
        o2->encoding == REDIS_ENCODING_INT)
            return o1->ptr == o2->ptr;

    o1 = getDecodedObject(o1);
    o2 = getDecodedObject(o2);
    cmp = dictSdsKeyCompare(privdata,o1->ptr,o2->ptr);
    decrRefCount(o1);
    decrRefCount(o2);
    return cmp;
}

unsigned int dictEncObjHash(const void *key) {
    robj *o = (robj*) key;

    if (o->encoding == REDIS_ENCODING_RAW) {
        return dictGenHashFunction(o->ptr, sdslen((sds)o->ptr));
    } else {
        if (o->encoding == REDIS_ENCODING_INT) {
            char buf[32];
            int len;

            len = ll2string(buf,32,(long)o->ptr);
            return dictGenHashFunction((unsigned char*)buf, len);
        } else {
            unsigned int hash;

            o = getDecodedObject(o);
            hash = dictGenHashFunction(o->ptr, sdslen((sds)o->ptr));
            decrRefCount(o);
            return hash;
        }
    }
}

486
/* Sets type hash table */
487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502
dictType setDictType = {
    dictEncObjHash,            /* hash function */
    NULL,                      /* key dup */
    NULL,                      /* val dup */
    dictEncObjKeyCompare,      /* key compare */
    dictRedisObjectDestructor, /* key destructor */
    NULL                       /* val destructor */
};

/* Sorted sets hash (note: a skiplist is used in addition to the hash table) */
dictType zsetDictType = {
    dictEncObjHash,            /* hash function */
    NULL,                      /* key dup */
    NULL,                      /* val dup */
    dictEncObjKeyCompare,      /* key compare */
    dictRedisObjectDestructor, /* key destructor */
503
    NULL                       /* val destructor */
504 505 506 507 508 509 510 511 512 513 514 515
};

/* Db->dict, keys are sds strings, vals are Redis objects. */
dictType dbDictType = {
    dictSdsHash,                /* hash function */
    NULL,                       /* key dup */
    NULL,                       /* val dup */
    dictSdsKeyCompare,          /* key compare */
    dictSdsDestructor,          /* key destructor */
    dictRedisObjectDestructor   /* val destructor */
};

A
antirez 已提交
516 517 518 519 520 521 522 523 524 525
/* server.lua_scripts sha (as sds string) -> scripts (as robj) cache. */
dictType shaScriptObjectDictType = {
    dictSdsCaseHash,            /* hash function */
    NULL,                       /* key dup */
    NULL,                       /* val dup */
    dictSdsKeyCaseCompare,      /* key compare */
    dictSdsDestructor,          /* key destructor */
    dictRedisObjectDestructor   /* val destructor */
};

526 527 528 529 530 531 532 533 534 535
/* Db->expires */
dictType keyptrDictType = {
    dictSdsHash,               /* hash function */
    NULL,                      /* key dup */
    NULL,                      /* val dup */
    dictSdsKeyCompare,         /* key compare */
    NULL,                      /* key destructor */
    NULL                       /* val destructor */
};

536 537 538 539 540 541 542 543 544 545
/* Command table. sds string -> command struct pointer. */
dictType commandTableDictType = {
    dictSdsCaseHash,           /* hash function */
    NULL,                      /* key dup */
    NULL,                      /* val dup */
    dictSdsKeyCaseCompare,     /* key compare */
    dictSdsDestructor,         /* key destructor */
    NULL                       /* val destructor */
};

546
/* Hash type hash table (note that small hashes are represented with ziplists) */
547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567
dictType hashDictType = {
    dictEncObjHash,             /* hash function */
    NULL,                       /* key dup */
    NULL,                       /* val dup */
    dictEncObjKeyCompare,       /* key compare */
    dictRedisObjectDestructor,  /* key destructor */
    dictRedisObjectDestructor   /* val destructor */
};

/* Keylist hash table type has unencoded redis objects as keys and
 * lists as values. It's used for blocking operations (BLPOP) and to
 * map swapped keys to a list of clients waiting for this keys to be loaded. */
dictType keylistDictType = {
    dictObjHash,                /* hash function */
    NULL,                       /* key dup */
    NULL,                       /* val dup */
    dictObjKeyCompare,          /* key compare */
    dictRedisObjectDestructor,  /* key destructor */
    dictListDestructor          /* val destructor */
};

A
antirez 已提交
568 569 570 571 572 573 574 575 576 577 578
/* Cluster nodes hash table, mapping nodes addresses 1.2.3.4:6379 to
 * clusterNode structures. */
dictType clusterNodesDictType = {
    dictSdsHash,                /* hash function */
    NULL,                       /* key dup */
    NULL,                       /* val dup */
    dictSdsKeyCompare,          /* key compare */
    dictSdsDestructor,          /* key destructor */
    NULL                        /* val destructor */
};

A
antirez 已提交
579 580 581 582 583 584 585 586 587 588
/* Migrate cache dict type. */
dictType migrateCacheDictType = {
    dictSdsHash,                /* hash function */
    NULL,                       /* key dup */
    NULL,                       /* val dup */
    dictSdsKeyCompare,          /* key compare */
    dictSdsDestructor,          /* key destructor */
    NULL                        /* val destructor */
};

589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618
int htNeedsResize(dict *dict) {
    long long size, used;

    size = dictSlots(dict);
    used = dictSize(dict);
    return (size && used && size > DICT_HT_INITIAL_SIZE &&
            (used*100/size < REDIS_HT_MINFILL));
}

/* If the percentage of used slots in the HT reaches REDIS_HT_MINFILL
 * we resize the hash table to save memory */
void tryResizeHashTables(void) {
    int j;

    for (j = 0; j < server.dbnum; j++) {
        if (htNeedsResize(server.db[j].dict))
            dictResize(server.db[j].dict);
        if (htNeedsResize(server.db[j].expires))
            dictResize(server.db[j].expires);
    }
}

/* Our hash table implementation performs rehashing incrementally while
 * we write/read from the hash table. Still if the server is idle, the hash
 * table will use two tables for a long time. So we try to use 1 millisecond
 * of CPU time at every serverCron() loop in order to rehash some key. */
void incrementallyRehash(void) {
    int j;

    for (j = 0; j < server.dbnum; j++) {
619
        /* Keys dictionary */
620 621 622 623
        if (dictIsRehashing(server.db[j].dict)) {
            dictRehashMilliseconds(server.db[j].dict,1);
            break; /* already used our millisecond for this loop... */
        }
624 625 626 627 628
        /* Expires */
        if (dictIsRehashing(server.db[j].expires)) {
            dictRehashMilliseconds(server.db[j].expires,1);
            break; /* already used our millisecond for this loop... */
        }
629 630 631 632 633 634 635 636 637 638
    }
}

/* This function is called once a background process of some kind terminates,
 * as we want to avoid resizing the hash tables when there is a child in order
 * to play well with copy-on-write (otherwise when a resize happens lots of
 * memory pages are copied). The goal of this function is to update the ability
 * for dict.c to resize the hash tables accordingly to the fact we have o not
 * running childs. */
void updateDictResizePolicy(void) {
A
antirez 已提交
639
    if (server.rdb_child_pid == -1 && server.aof_child_pid == -1)
640 641 642 643 644 645 646
        dictEnableResize();
    else
        dictDisableResize();
}

/* ======================= Cron: called every 100 ms ======================== */

647 648 649 650 651
/* Try to expire a few timed out keys. The algorithm used is adaptive and
 * will use few CPU cycles if there are few expiring keys, otherwise
 * it will get more aggressive to avoid that too much memory is used by
 * keys that can be removed from the keyspace. */
void activeExpireCycle(void) {
652
    int j, iteration = 0;
653
    long long start = ustime(), timelimit;
654 655 656

    /* We can use at max REDIS_EXPIRELOOKUPS_TIME_PERC percentage of CPU time
     * per iteration. Since this function gets called with a frequency of
657
     * server.hz times per second, the following is the max amount of
658
     * microseconds we can spend in this function. */
659
    timelimit = 1000000*REDIS_EXPIRELOOKUPS_TIME_PERC/server.hz/100;
660
    if (timelimit <= 0) timelimit = 1;
661 662

    for (j = 0; j < server.dbnum; j++) {
663
        int expired;
664 665 666 667 668
        redisDb *db = server.db+j;

        /* Continue to expire if at the end of the cycle more than 25%
         * of the keys were expired. */
        do {
669 670
            unsigned long num = dictSize(db->expires);
            unsigned long slots = dictSlots(db->expires);
671
            long long now = mstime();
672

673 674 675 676 677 678
            /* When there are less than 1% filled slots getting random
             * keys is expensive, so stop here waiting for better times...
             * The dictionary will be resized asap. */
            if (num && slots > DICT_HT_INITIAL_SIZE &&
                (num*100/slots < 1)) break;

679 680
            /* The main collection cycle. Sample random keys among keys
             * with an expire set, checking for expired ones. */
681 682 683 684 685
            expired = 0;
            if (num > REDIS_EXPIRELOOKUPS_PER_CRON)
                num = REDIS_EXPIRELOOKUPS_PER_CRON;
            while (num--) {
                dictEntry *de;
686
                long long t;
687 688

                if ((de = dictGetRandomKey(db->expires)) == NULL) break;
689
                t = dictGetSignedIntegerVal(de);
690
                if (now > t) {
691
                    sds key = dictGetKey(de);
692 693 694 695
                    robj *keyobj = createStringObject(key,sdslen(key));

                    propagateExpire(db,keyobj);
                    dbDelete(db,keyobj);
696 697
                    notifyKeyspaceEvent(REDIS_NOTIFY_EXPIRED,
                        "expired",keyobj,db->id);
698 699 700 701 702
                    decrRefCount(keyobj);
                    expired++;
                    server.stat_expiredkeys++;
                }
            }
703 704 705 706
            /* We can't block forever here even if there are many keys to
             * expire. So after a given amount of milliseconds return to the
             * caller waiting for the other active expire cycle. */
            iteration++;
707
            if ((iteration & 0xf) == 0 && /* check once every 16 cycles. */
708
                (ustime()-start) > timelimit) return;
709 710 711 712
        } while (expired > REDIS_EXPIRELOOKUPS_PER_CRON/4);
    }
}

713
void updateLRUClock(void) {
714
    server.lruclock = (server.unixtime/REDIS_LRU_CLOCK_RESOLUTION) &
715 716
                                                REDIS_LRU_CLOCK_MAX;
}
717

718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742

/* Add a sample to the operations per second array of samples. */
void trackOperationsPerSecond(void) {
    long long t = mstime() - server.ops_sec_last_sample_time;
    long long ops = server.stat_numcommands - server.ops_sec_last_sample_ops;
    long long ops_sec;

    ops_sec = t > 0 ? (ops*1000/t) : 0;

    server.ops_sec_samples[server.ops_sec_idx] = ops_sec;
    server.ops_sec_idx = (server.ops_sec_idx+1) % REDIS_OPS_SEC_SAMPLES;
    server.ops_sec_last_sample_time = mstime();
    server.ops_sec_last_sample_ops = server.stat_numcommands;
}

/* Return the mean of all the samples. */
long long getOperationsPerSecond(void) {
    int j;
    long long sum = 0;

    for (j = 0; j < REDIS_OPS_SEC_SAMPLES; j++)
        sum += server.ops_sec_samples[j];
    return sum / REDIS_OPS_SEC_SAMPLES;
}

A
antirez 已提交
743 744
/* Check for timeouts. Returns non-zero if the client was terminated */
int clientsCronHandleTimeout(redisClient *c) {
745
    time_t now = server.unixtime;
746 747 748 749 750 751 752 753 754 755 756

    if (server.maxidletime &&
        !(c->flags & REDIS_SLAVE) &&    /* no timeout for slaves */
        !(c->flags & REDIS_MASTER) &&   /* no timeout for masters */
        !(c->flags & REDIS_BLOCKED) &&  /* no timeout for BLPOP */
        dictSize(c->pubsub_channels) == 0 && /* no timeout for pubsub */
        listLength(c->pubsub_patterns) == 0 &&
        (now - c->lastinteraction > server.maxidletime))
    {
        redisLog(REDIS_VERBOSE,"Closing idle client");
        freeClient(c);
A
antirez 已提交
757
        return 1;
758 759 760 761 762 763
    } else if (c->flags & REDIS_BLOCKED) {
        if (c->bpop.timeout != 0 && c->bpop.timeout < now) {
            addReply(c,shared.nullmultibulk);
            unblockClientWaitingData(c);
        }
    }
A
antirez 已提交
764
    return 0;
765 766
}

767
/* The client query buffer is an sds.c string that can end with a lot of
A
antirez 已提交
768 769
 * free space not used, this function reclaims space if needed.
 *
G
guiquanz 已提交
770
 * The function always returns 0 as it never terminates the client. */
A
antirez 已提交
771
int clientsCronResizeQueryBuffer(redisClient *c) {
772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789
    size_t querybuf_size = sdsAllocSize(c->querybuf);
    time_t idletime = server.unixtime - c->lastinteraction;

    /* There are two conditions to resize the query buffer:
     * 1) Query buffer is > BIG_ARG and too big for latest peak.
     * 2) Client is inactive and the buffer is bigger than 1k. */
    if (((querybuf_size > REDIS_MBULK_BIG_ARG) &&
         (querybuf_size/(c->querybuf_peak+1)) > 2) ||
         (querybuf_size > 1024 && idletime > 2))
    {
        /* Only resize the query buffer if it is actually wasting space. */
        if (sdsavail(c->querybuf) > 1024) {
            c->querybuf = sdsRemoveFreeSpace(c->querybuf);
        }
    }
    /* Reset the peak again to capture the peak memory usage in the next
     * cycle. */
    c->querybuf_peak = 0;
A
antirez 已提交
790
    return 0;
791 792
}

793
void clientsCron(void) {
794 795
    /* Make sure to process at least 1/(server.hz*10) of clients per call.
     * Since this function is called server.hz times per second we are sure that
796 797 798
     * in the worst case we process all the clients in 10 seconds.
     * In normal conditions (a reasonable number of clients) we process
     * all the clients in a shorter time. */
799
    int numclients = listLength(server.clients);
800
    int iterations = numclients/(server.hz*10);
801

802 803
    if (iterations < 50)
        iterations = (numclients < 50) ? numclients : 50;
804 805 806 807 808 809 810 811 812 813
    while(listLength(server.clients) && iterations--) {
        redisClient *c;
        listNode *head;

        /* Rotate the list, take the current head, process.
         * This way if the client must be removed from the list it's the
         * first element and we don't incur into O(N) computation. */
        listRotate(server.clients);
        head = listFirst(server.clients);
        c = listNodeValue(head);
A
antirez 已提交
814 815 816 817 818
        /* The following functions do different service checks on the client.
         * The protocol is that they return non-zero if the client was
         * terminated. */
        if (clientsCronHandleTimeout(c)) continue;
        if (clientsCronResizeQueryBuffer(c)) continue;
819 820 821
    }
}

822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842
/* This function handles 'background' operations we are required to do
 * incrementally in Redis databases, such as active key expiring, resizing,
 * rehashing. */
void databasesCron(void) {
    /* Expire a few keys per cycle, only if this is a master.
     * On slaves we wait for DEL operations synthesized by the master
     * in order to guarantee a strict consistency. */
    if (server.masterhost == NULL) activeExpireCycle();

    /* We don't want to resize the hash tables while a background saving
     * is in progress: the saving child is created using fork() that is
     * implemented with a copy-on-write semantic in most modern systems, so
     * if we resize the HT while there is the saving child at work actually
     * a lot of memory movements in the parent will cause a lot of pages
     * copied. */
    if (server.rdb_child_pid == -1 && server.aof_child_pid == -1) {
        tryResizeHashTables();
        if (server.activerehashing) incrementallyRehash();
    }
}

843
/* This is our timer interrupt, called server.hz times per second.
844 845 846 847 848
 * Here is where we do a number of things that need to be done asynchronously.
 * For instance:
 *
 * - Active expired keys collection (it is also performed in a lazy way on
 *   lookup).
G
guiquanz 已提交
849
 * - Software watchdog.
850 851 852
 * - Update some statistic.
 * - Incremental rehashing of the DBs hash tables.
 * - Triggering BGSAVE / AOF rewrite, and handling of terminated children.
G
guiquanz 已提交
853
 * - Clients timeout of different kinds.
854 855 856
 * - Replication reconnection.
 * - Many more...
 *
857
 * Everything directly called here will be called server.hz times per second,
858 859 860 861
 * so in order to throttle execution of things we want to do less frequently
 * a macro is used: run_with_period(milliseconds) { .... }
 */

862
int serverCron(struct aeEventLoop *eventLoop, long long id, void *clientData) {
863
    int j;
864 865 866 867
    REDIS_NOTUSED(eventLoop);
    REDIS_NOTUSED(id);
    REDIS_NOTUSED(clientData);

A
antirez 已提交
868 869 870 871
    /* Software watchdog: deliver the SIGALRM that will reach the signal
     * handler if we don't return here fast enough. */
    if (server.watchdog_period) watchdogScheduleSignal(server.watchdog_period);

872 873 874 875 876
    /* We take a cached value of the unix time in the global state because
     * with virtual memory and aging there is to store the current time
     * in objects at every object access, and accuracy is not needed.
     * To access a global var is faster than calling time(NULL) */
    server.unixtime = time(NULL);
A
antirez 已提交
877

878
    run_with_period(100) trackOperationsPerSecond();
879

880
    /* We have just 22 bits per object for LRU information.
881
     * So we use an (eventually wrapping) LRU clock with 10 seconds resolution.
G
guiquanz 已提交
882
     * 2^22 bits with 10 seconds resolution is more or less 1.5 years.
883
     *
884
     * Note that even if this will wrap after 1.5 years it's not a problem,
885
     * everything will still work but just some object will appear younger
886 887 888 889 890
     * to Redis. But for this to happen a given object should never be touched
     * for 1.5 years.
     *
     * Note that you can change the resolution altering the
     * REDIS_LRU_CLOCK_RESOLUTION define.
891
     */
892
    updateLRUClock();
893

894 895 896 897
    /* Record the max memory used since the server was started. */
    if (zmalloc_used_memory() > server.stat_peak_memory)
        server.stat_peak_memory = zmalloc_used_memory();

898 899 900
    /* We received a SIGTERM, shutting down here in a safe way, as it is
     * not ok doing so inside the signal handler. */
    if (server.shutdown_asap) {
901
        if (prepareForShutdown(0) == REDIS_OK) exit(0);
902
        redisLog(REDIS_WARNING,"SIGTERM received but errors trying to shut down the server, check the logs for more information");
903
        server.shutdown_asap = 0;
904 905 906
    }

    /* Show some info about non-empty databases */
907 908 909 910 911 912 913 914 915 916 917
    run_with_period(5000) {
        for (j = 0; j < server.dbnum; j++) {
            long long size, used, vkeys;

            size = dictSlots(server.db[j].dict);
            used = dictSize(server.db[j].dict);
            vkeys = dictSize(server.db[j].expires);
            if (used || vkeys) {
                redisLog(REDIS_VERBOSE,"DB %d: %lld keys (%lld volatile) in %lld slots HT.",j,used,vkeys,size);
                /* dictPrintStats(server.dict); */
            }
918 919 920 921
        }
    }

    /* Show information about connected clients */
922 923 924
    if (!server.sentinel_mode) {
        run_with_period(5000) {
            redisLog(REDIS_VERBOSE,
925
                "%lu clients connected (%lu slaves), %zu bytes in use",
926 927 928 929
                listLength(server.clients)-listLength(server.slaves),
                listLength(server.slaves),
                zmalloc_used_memory());
        }
930 931
    }

932 933
    /* We need to do a few operations on clients asynchronously. */
    clientsCron();
934

935 936 937
    /* Handle background operations on Redis databases. */
    databasesCron();

938 939
    /* Start a scheduled AOF rewrite if this was requested by the user while
     * a BGSAVE was in progress. */
A
antirez 已提交
940
    if (server.rdb_child_pid == -1 && server.aof_child_pid == -1 &&
941
        server.aof_rewrite_scheduled)
942 943 944 945
    {
        rewriteAppendOnlyFileBackground();
    }

A
antirez 已提交
946
    /* Check if a background saving or AOF rewrite in progress terminated. */
A
antirez 已提交
947
    if (server.rdb_child_pid != -1 || server.aof_child_pid != -1) {
948 949 950 951
        int statloc;
        pid_t pid;

        if ((pid = wait3(&statloc,WNOHANG,NULL)) != 0) {
952 953 954 955 956
            int exitcode = WEXITSTATUS(statloc);
            int bysignal = 0;
            
            if (WIFSIGNALED(statloc)) bysignal = WTERMSIG(statloc);

A
antirez 已提交
957
            if (pid == server.rdb_child_pid) {
958
                backgroundSaveDoneHandler(exitcode,bysignal);
959
            } else if (pid == server.aof_child_pid) {
960
                backgroundRewriteDoneHandler(exitcode,bysignal);
961 962 963 964
            } else {
                redisLog(REDIS_WARNING,
                    "Warning, detected child with unmatched pid: %ld",
                    (long)pid);
965 966 967
            }
            updateDictResizePolicy();
        }
A
antirez 已提交
968
    } else {
969 970
        /* If there is not a background saving/rewrite in progress check if
         * we have to save/rewrite now */
971 972 973 974
         for (j = 0; j < server.saveparamslen; j++) {
            struct saveparam *sp = server.saveparams+j;

            if (server.dirty >= sp->changes &&
975
                server.unixtime-server.lastsave > sp->seconds) {
976
                redisLog(REDIS_NOTICE,"%d changes in %d seconds. Saving...",
977
                    sp->changes, (int)sp->seconds);
A
antirez 已提交
978
                rdbSaveBackground(server.rdb_filename);
979 980 981
                break;
            }
         }
982 983

         /* Trigger an AOF rewrite if needed */
A
antirez 已提交
984
         if (server.rdb_child_pid == -1 &&
A
antirez 已提交
985
             server.aof_child_pid == -1 &&
986 987
             server.aof_rewrite_perc &&
             server.aof_current_size > server.aof_rewrite_min_size)
988
         {
989 990 991 992
            long long base = server.aof_rewrite_base_size ?
                            server.aof_rewrite_base_size : 1;
            long long growth = (server.aof_current_size*100/base) - 100;
            if (growth >= server.aof_rewrite_perc) {
993
                redisLog(REDIS_NOTICE,"Starting automatic rewriting of AOF on %lld%% growth",growth);
994 995 996
                rewriteAppendOnlyFileBackground();
            }
         }
997 998
    }

999 1000 1001 1002 1003

    /* If we postponed an AOF buffer flush, let's try to do it every time the
     * cron function is called. */
    if (server.aof_flush_postponed_start) flushAppendOnlyFile(0);

1004 1005 1006
    /* Close clients that need to be closed asynchronous */
    freeClientsInAsyncFreeQueue();

1007 1008
    /* Replication cron function -- used to reconnect to master and
     * to detect transfer failures. */
1009
    run_with_period(1000) replicationCron();
1010

A
antirez 已提交
1011
    /* Run the Redis Cluster cron. */
1012 1013 1014
    run_with_period(1000) {
        if (server.cluster_enabled) clusterCron();
    }
A
antirez 已提交
1015

A
antirez 已提交
1016
    /* Run the Sentinel timer if we are in sentinel mode. */
1017 1018 1019 1020
    run_with_period(100) {
        if (server.sentinel_mode) sentinelTimer();
    }

A
antirez 已提交
1021 1022 1023 1024 1025
    /* Cleanup expired MIGRATE cached sockets. */
    run_with_period(1000) {
        migrateCloseTimedoutSockets();
    }

1026
    server.cronloops++;
1027
    return 1000/server.hz;
1028 1029 1030 1031 1032 1033 1034
}

/* This function gets called every time Redis is entering the
 * main loop of the event driven library, that is, before to sleep
 * for ready file descriptors. */
void beforeSleep(struct aeEventLoop *eventLoop) {
    REDIS_NOTUSED(eventLoop);
1035 1036
    listNode *ln;
    redisClient *c;
1037

1038 1039 1040 1041 1042 1043
    /* Try to process pending commands for clients that were just unblocked. */
    while (listLength(server.unblocked_clients)) {
        ln = listFirst(server.unblocked_clients);
        redisAssert(ln != NULL);
        c = ln->value;
        listDelNode(server.unblocked_clients,ln);
1044
        c->flags &= ~REDIS_UNBLOCKED;
1045 1046

        /* Process remaining data in the input buffer. */
1047 1048
        if (c->querybuf && sdslen(c->querybuf) > 0) {
            server.current_client = c;
1049
            processInputBuffer(c);
1050 1051
            server.current_client = NULL;
        }
1052 1053
    }

1054
    /* Write the AOF buffer on disk */
1055
    flushAppendOnlyFile(0);
1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075
}

/* =========================== Server initialization ======================== */

void createSharedObjects(void) {
    int j;

    shared.crlf = createObject(REDIS_STRING,sdsnew("\r\n"));
    shared.ok = createObject(REDIS_STRING,sdsnew("+OK\r\n"));
    shared.err = createObject(REDIS_STRING,sdsnew("-ERR\r\n"));
    shared.emptybulk = createObject(REDIS_STRING,sdsnew("$0\r\n\r\n"));
    shared.czero = createObject(REDIS_STRING,sdsnew(":0\r\n"));
    shared.cone = createObject(REDIS_STRING,sdsnew(":1\r\n"));
    shared.cnegone = createObject(REDIS_STRING,sdsnew(":-1\r\n"));
    shared.nullbulk = createObject(REDIS_STRING,sdsnew("$-1\r\n"));
    shared.nullmultibulk = createObject(REDIS_STRING,sdsnew("*-1\r\n"));
    shared.emptymultibulk = createObject(REDIS_STRING,sdsnew("*0\r\n"));
    shared.pong = createObject(REDIS_STRING,sdsnew("+PONG\r\n"));
    shared.queued = createObject(REDIS_STRING,sdsnew("+QUEUED\r\n"));
    shared.wrongtypeerr = createObject(REDIS_STRING,sdsnew(
1076
        "-WRONGTYPE Operation against a key holding the wrong kind of value\r\n"));
1077 1078 1079 1080 1081 1082 1083 1084
    shared.nokeyerr = createObject(REDIS_STRING,sdsnew(
        "-ERR no such key\r\n"));
    shared.syntaxerr = createObject(REDIS_STRING,sdsnew(
        "-ERR syntax error\r\n"));
    shared.sameobjecterr = createObject(REDIS_STRING,sdsnew(
        "-ERR source and destination objects are the same\r\n"));
    shared.outofrangeerr = createObject(REDIS_STRING,sdsnew(
        "-ERR index out of range\r\n"));
A
antirez 已提交
1085 1086
    shared.noscripterr = createObject(REDIS_STRING,sdsnew(
        "-NOSCRIPT No matching script. Please use EVAL.\r\n"));
1087 1088
    shared.loadingerr = createObject(REDIS_STRING,sdsnew(
        "-LOADING Redis is loading the dataset in memory\r\n"));
1089
    shared.slowscripterr = createObject(REDIS_STRING,sdsnew(
1090
        "-BUSY Redis is busy running a script. You can only call SCRIPT KILL or SHUTDOWN NOSAVE.\r\n"));
1091 1092
    shared.masterdownerr = createObject(REDIS_STRING,sdsnew(
        "-MASTERDOWN Link with MASTER is down and slave-serve-stale-data is set to 'no'.\r\n"));
1093
    shared.bgsaveerr = createObject(REDIS_STRING,sdsnew(
1094 1095 1096
        "-MISCONF Redis is configured to save RDB snapshots, but is currently not able to persist on disk. Commands that may modify the data set are disabled. Please check Redis logs for details about the error.\r\n"));
    shared.roslaveerr = createObject(REDIS_STRING,sdsnew(
        "-READONLY You can't write against a read only slave.\r\n"));
1097 1098
    shared.noautherr = createObject(REDIS_STRING,sdsnew(
        "-NOAUTH Authentication required.\r\n"));
1099 1100
    shared.oomerr = createObject(REDIS_STRING,sdsnew(
        "-OOM command not allowed when used memory > 'maxmemory'.\r\n"));
A
antirez 已提交
1101 1102
    shared.execaborterr = createObject(REDIS_STRING,sdsnew(
        "-EXECABORT Transaction discarded because of previous errors.\r\n"));
1103 1104 1105
    shared.space = createObject(REDIS_STRING,sdsnew(" "));
    shared.colon = createObject(REDIS_STRING,sdsnew(":"));
    shared.plus = createObject(REDIS_STRING,sdsnew("+"));
1106 1107

    for (j = 0; j < REDIS_SHARED_SELECT_CMDS; j++) {
1108 1109 1110 1111
        char dictid_str[64];
        int dictid_len;

        dictid_len = ll2string(dictid_str,sizeof(dictid_str),j);
1112
        shared.select[j] = createObject(REDIS_STRING,
1113 1114 1115
            sdscatprintf(sdsempty(),
                "*2\r\n$6\r\nSELECT\r\n$%d\r\n%s\r\n",
                dictid_len, dictid_str));
1116
    }
1117 1118 1119 1120 1121 1122
    shared.messagebulk = createStringObject("$7\r\nmessage\r\n",13);
    shared.pmessagebulk = createStringObject("$8\r\npmessage\r\n",14);
    shared.subscribebulk = createStringObject("$9\r\nsubscribe\r\n",15);
    shared.unsubscribebulk = createStringObject("$11\r\nunsubscribe\r\n",18);
    shared.psubscribebulk = createStringObject("$10\r\npsubscribe\r\n",17);
    shared.punsubscribebulk = createStringObject("$12\r\npunsubscribe\r\n",19);
1123
    shared.del = createStringObject("DEL",3);
1124 1125
    shared.rpop = createStringObject("RPOP",4);
    shared.lpop = createStringObject("LPOP",4);
1126
    shared.lpush = createStringObject("LPUSH",5);
1127 1128 1129 1130
    for (j = 0; j < REDIS_SHARED_INTEGERS; j++) {
        shared.integers[j] = createObject(REDIS_STRING,(void*)(long)j);
        shared.integers[j]->encoding = REDIS_ENCODING_INT;
    }
1131 1132 1133 1134 1135 1136
    for (j = 0; j < REDIS_SHARED_BULKHDR_LEN; j++) {
        shared.mbulkhdr[j] = createObject(REDIS_STRING,
            sdscatprintf(sdsempty(),"*%d\r\n",j));
        shared.bulkhdr[j] = createObject(REDIS_STRING,
            sdscatprintf(sdsempty(),"$%d\r\n",j));
    }
1137 1138 1139
}

void initServerConfig() {
A
antirez 已提交
1140
    getRandomHexChars(server.runid,REDIS_RUN_ID_SIZE);
1141
    server.hz = REDIS_DEFAULT_HZ;
A
antirez 已提交
1142
    server.runid[REDIS_RUN_ID_SIZE] = '\0';
1143
    server.arch_bits = (sizeof(long) == 8) ? 64 : 32;
1144
    server.port = REDIS_SERVERPORT;
1145
    server.bindaddr = NULL;
1146
    server.unixsocket = NULL;
1147
    server.unixsocketperm = 0;
1148 1149 1150
    server.ipfd = -1;
    server.sofd = -1;
    server.dbnum = REDIS_DEFAULT_DBNUM;
A
antirez 已提交
1151
    server.verbosity = REDIS_NOTICE;
1152
    server.maxidletime = REDIS_MAXIDLETIME;
1153
    server.tcpkeepalive = 0;
1154
    server.client_max_querybuf_len = REDIS_MAX_QUERYBUF_LEN;
1155
    server.saveparams = NULL;
1156
    server.loading = 0;
1157
    server.logfile = NULL; /* NULL = log on standard output */
J
Jonah H. Harris 已提交
1158 1159 1160
    server.syslog_enabled = 0;
    server.syslog_ident = zstrdup("redis");
    server.syslog_facility = LOG_LOCAL0;
1161
    server.daemonize = 0;
1162
    server.aof_state = REDIS_AOF_OFF;
1163 1164 1165 1166 1167 1168
    server.aof_fsync = AOF_FSYNC_EVERYSEC;
    server.aof_no_fsync_on_rewrite = 0;
    server.aof_rewrite_perc = REDIS_AOF_REWRITE_PERC;
    server.aof_rewrite_min_size = REDIS_AOF_REWRITE_MIN_SIZE;
    server.aof_rewrite_base_size = 0;
    server.aof_rewrite_scheduled = 0;
A
antirez 已提交
1169
    server.aof_last_fsync = time(NULL);
1170 1171
    server.aof_rewrite_time_last = -1;
    server.aof_rewrite_time_start = -1;
1172
    server.aof_lastbgrewrite_status = REDIS_OK;
1173
    server.aof_delayed_fsync = 0;
A
antirez 已提交
1174 1175
    server.aof_fd = -1;
    server.aof_selected_db = -1; /* Make sure the first time will not match */
1176
    server.aof_flush_postponed_start = 0;
1177
    server.pidfile = zstrdup("/var/run/redis.pid");
A
antirez 已提交
1178
    server.rdb_filename = zstrdup("dump.rdb");
1179
    server.aof_filename = zstrdup("appendonly.aof");
1180
    server.requirepass = NULL;
A
antirez 已提交
1181
    server.rdb_compression = 1;
1182
    server.rdb_checksum = 1;
1183
    server.activerehashing = 1;
A
antirez 已提交
1184
    server.notify_keyspace_events = 0;
1185
    server.maxclients = REDIS_MAX_CLIENTS;
1186
    server.bpop_blocked_clients = 0;
1187
    server.maxmemory = 0;
1188 1189
    server.maxmemory_policy = REDIS_MAXMEMORY_VOLATILE_LRU;
    server.maxmemory_samples = 3;
1190 1191
    server.hash_max_ziplist_entries = REDIS_HASH_MAX_ZIPLIST_ENTRIES;
    server.hash_max_ziplist_value = REDIS_HASH_MAX_ZIPLIST_VALUE;
1192 1193
    server.list_max_ziplist_entries = REDIS_LIST_MAX_ZIPLIST_ENTRIES;
    server.list_max_ziplist_value = REDIS_LIST_MAX_ZIPLIST_VALUE;
1194
    server.set_max_intset_entries = REDIS_SET_MAX_INTSET_ENTRIES;
1195 1196
    server.zset_max_ziplist_entries = REDIS_ZSET_MAX_ZIPLIST_ENTRIES;
    server.zset_max_ziplist_value = REDIS_ZSET_MAX_ZIPLIST_VALUE;
1197
    server.shutdown_asap = 0;
A
1fe4cd5  
antirez 已提交
1198 1199
    server.repl_ping_slave_period = REDIS_REPL_PING_SLAVE_PERIOD;
    server.repl_timeout = REDIS_REPL_TIMEOUT;
A
antirez 已提交
1200
    server.cluster_enabled = 0;
1201
    server.cluster_configfile = zstrdup("nodes.conf");
1202
    server.lua_caller = NULL;
A
antirez 已提交
1203
    server.lua_time_limit = REDIS_LUA_TIME_LIMIT;
1204
    server.lua_client = NULL;
1205
    server.lua_timedout = 0;
A
antirez 已提交
1206
    server.migrate_cached_sockets = dictCreate(&migrateCacheDictType,NULL);
1207

A
antirez 已提交
1208
    updateLRUClock();
1209 1210 1211 1212 1213 1214 1215 1216 1217 1218
    resetServerSaveParams();

    appendServerSaveParams(60*60,1);  /* save after 1 hour and 1 change */
    appendServerSaveParams(300,100);  /* save after 5 minutes and 100 changes */
    appendServerSaveParams(60,10000); /* save after 1 minute and 10000 changes */
    /* Replication related */
    server.masterauth = NULL;
    server.masterhost = NULL;
    server.masterport = 6379;
    server.master = NULL;
1219 1220
    server.cached_master = NULL;
    server.repl_master_initial_offset = -1;
A
antirez 已提交
1221
    server.repl_state = REDIS_REPL_NONE;
1222
    server.repl_syncio_timeout = REDIS_REPL_SYNCIO_TIMEOUT;
1223
    server.repl_serve_stale_data = 1;
1224
    server.repl_slave_ro = 1;
1225
    server.repl_down_since = time(NULL);
1226
    server.repl_disable_tcp_nodelay = 0;
1227
    server.slave_priority = REDIS_DEFAULT_SLAVE_PRIORITY;
1228 1229 1230 1231 1232 1233 1234 1235 1236 1237
    server.master_repl_offset = 0;

    /* Replication partial resync backlog */
    server.repl_backlog = NULL;
    server.repl_backlog_size = REDIS_DEFAULT_REPL_BACKLOG_SIZE;
    server.repl_backlog_histlen = 0;
    server.repl_backlog_idx = 0;
    server.repl_backlog_off = 0;
    server.repl_backlog_time_limit = REDIS_DEFAULT_REPL_BACKLOG_TIME_LIMIT;
    server.repl_no_slaves_since = time(NULL);
1238

1239 1240 1241 1242
    /* Client output buffer limits */
    server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_NORMAL].hard_limit_bytes = 0;
    server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_NORMAL].soft_limit_bytes = 0;
    server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_NORMAL].soft_limit_seconds = 0;
1243 1244
    server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_SLAVE].hard_limit_bytes = 1024*1024*256;
    server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_SLAVE].soft_limit_bytes = 1024*1024*64;
1245
    server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_SLAVE].soft_limit_seconds = 60;
1246 1247
    server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_PUBSUB].hard_limit_bytes = 1024*1024*32;
    server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_PUBSUB].soft_limit_bytes = 1024*1024*8;
1248 1249
    server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_PUBSUB].soft_limit_seconds = 60;

1250 1251 1252 1253 1254
    /* Double constants initialization */
    R_Zero = 0.0;
    R_PosInf = 1.0/R_Zero;
    R_NegInf = -1.0/R_Zero;
    R_Nan = R_Zero/R_Zero;
1255

G
guiquanz 已提交
1256
    /* Command table -- we initiialize it here as it is part of the
1257 1258 1259
     * initial configuration, since command names may be changed via
     * redis.conf using the rename-command directive. */
    server.commands = dictCreate(&commandTableDictType,NULL);
1260
    server.orig_commands = dictCreate(&commandTableDictType,NULL);
1261 1262 1263
    populateCommandTable();
    server.delCommand = lookupCommandByCString("del");
    server.multiCommand = lookupCommandByCString("multi");
1264
    server.lpushCommand = lookupCommandByCString("lpush");
1265 1266
    server.lpopCommand = lookupCommandByCString("lpop");
    server.rpopCommand = lookupCommandByCString("rpop");
1267 1268 1269 1270
    
    /* Slow log */
    server.slowlog_log_slower_than = REDIS_SLOWLOG_LOG_SLOWER_THAN;
    server.slowlog_max_len = REDIS_SLOWLOG_MAX_LEN;
A
antirez 已提交
1271

A
antirez 已提交
1272
    /* Debugging */
A
antirez 已提交
1273 1274 1275 1276
    server.assert_failed = "<no assertion failed>";
    server.assert_file = "<no file>";
    server.assert_line = 0;
    server.bug_report_start = 0;
A
antirez 已提交
1277
    server.watchdog_period = 0;
1278 1279
}

1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301
/* This function will try to raise the max number of open files accordingly to
 * the configured max number of clients. It will also account for 32 additional
 * file descriptors as we need a few more for persistence, listening
 * sockets, log files and so forth.
 *
 * If it will not be possible to set the limit accordingly to the configured
 * max number of clients, the function will do the reverse setting
 * server.maxclients to the value that we can actually handle. */
void adjustOpenFilesLimit(void) {
    rlim_t maxfiles = server.maxclients+32;
    struct rlimit limit;

    if (getrlimit(RLIMIT_NOFILE,&limit) == -1) {
        redisLog(REDIS_WARNING,"Unable to obtain the current NOFILE limit (%s), assuming 1024 and setting the max clients configuration accordingly.",
            strerror(errno));
        server.maxclients = 1024-32;
    } else {
        rlim_t oldlimit = limit.rlim_cur;

        /* Set the max number of files if the current limit is not enough
         * for our needs. */
        if (oldlimit < maxfiles) {
1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313
            rlim_t f;
            
            f = maxfiles;
            while(f > oldlimit) {
                limit.rlim_cur = f;
                limit.rlim_max = f;
                if (setrlimit(RLIMIT_NOFILE,&limit) != -1) break;
                f -= 128;
            }
            if (f < oldlimit) f = oldlimit;
            if (f != maxfiles) {
                server.maxclients = f-32;
1314 1315 1316 1317 1318 1319 1320 1321 1322 1323
                redisLog(REDIS_WARNING,"Unable to set the max number of files limit to %d (%s), setting the max clients configuration to %d.",
                    (int) maxfiles, strerror(errno), (int) server.maxclients);
            } else {
                redisLog(REDIS_NOTICE,"Max number of open files set to %d",
                    (int) maxfiles);
            }
        }
    }
}

1324 1325 1326 1327 1328
void initServer() {
    int j;

    signal(SIGHUP, SIG_IGN);
    signal(SIGPIPE, SIG_IGN);
1329
    setupSignalHandlers();
1330

J
Jonah H. Harris 已提交
1331 1332 1333 1334 1335
    if (server.syslog_enabled) {
        openlog(server.syslog_ident, LOG_PID | LOG_NDELAY | LOG_NOWAIT,
            server.syslog_facility);
    }

1336
    server.current_client = NULL;
1337
    server.clients = listCreate();
1338
    server.clients_to_close = listCreate();
1339 1340
    server.slaves = listCreate();
    server.monitors = listCreate();
1341
    server.slaveseldb = -1; /* Force to emit the first SELECT command. */
1342
    server.unblocked_clients = listCreate();
1343
    server.ready_keys = listCreate();
1344

1345
    createSharedObjects();
1346 1347
    adjustOpenFilesLimit();
    server.el = aeCreateEventLoop(server.maxclients+1024);
1348
    server.db = zmalloc(sizeof(redisDb)*server.dbnum);
1349

1350
    if (server.port != 0) {
1351
        server.ipfd = anetTcpServer(server.neterr,server.port,server.bindaddr);
1352
        if (server.ipfd == ANET_ERR) {
1353 1354
            redisLog(REDIS_WARNING, "Opening port %d: %s",
                server.port, server.neterr);
1355 1356
            exit(1);
        }
1357
    }
1358 1359
    if (server.unixsocket != NULL) {
        unlink(server.unixsocket); /* don't care if this fails */
1360
        server.sofd = anetUnixServer(server.neterr,server.unixsocket,server.unixsocketperm);
1361 1362 1363 1364
        if (server.sofd == ANET_ERR) {
            redisLog(REDIS_WARNING, "Opening socket: %s", server.neterr);
            exit(1);
        }
1365
    }
1366 1367
    if (server.ipfd < 0 && server.sofd < 0) {
        redisLog(REDIS_WARNING, "Configured to not listen anywhere, exiting.");
1368 1369 1370 1371 1372 1373
        exit(1);
    }
    for (j = 0; j < server.dbnum; j++) {
        server.db[j].dict = dictCreate(&dbDictType,NULL);
        server.db[j].expires = dictCreate(&keyptrDictType,NULL);
        server.db[j].blocking_keys = dictCreate(&keylistDictType,NULL);
1374
        server.db[j].ready_keys = dictCreate(&setDictType,NULL);
1375 1376 1377 1378 1379 1380 1381 1382
        server.db[j].watched_keys = dictCreate(&keylistDictType,NULL);
        server.db[j].id = j;
    }
    server.pubsub_channels = dictCreate(&keylistDictType,NULL);
    server.pubsub_patterns = listCreate();
    listSetFreeMethod(server.pubsub_patterns,freePubsubPattern);
    listSetMatchMethod(server.pubsub_patterns,listMatchPubsubPattern);
    server.cronloops = 0;
A
antirez 已提交
1383
    server.rdb_child_pid = -1;
A
antirez 已提交
1384
    server.aof_child_pid = -1;
1385
    aofRewriteBufferReset();
A
antirez 已提交
1386
    server.aof_buf = sdsempty();
1387
    server.lastsave = time(NULL);
1388 1389
    server.rdb_save_time_last = -1;
    server.rdb_save_time_start = -1;
1390 1391 1392 1393
    server.dirty = 0;
    server.stat_numcommands = 0;
    server.stat_numconnections = 0;
    server.stat_expiredkeys = 0;
1394
    server.stat_evictedkeys = 0;
1395
    server.stat_starttime = time(NULL);
1396 1397
    server.stat_keyspace_misses = 0;
    server.stat_keyspace_hits = 0;
1398
    server.stat_peak_memory = 0;
1399
    server.stat_fork_time = 0;
1400
    server.stat_rejected_conn = 0;
1401 1402 1403
    server.stat_sync_full = 0;
    server.stat_sync_partial_ok = 0;
    server.stat_sync_partial_err = 0;
1404 1405 1406 1407
    memset(server.ops_sec_samples,0,sizeof(server.ops_sec_samples));
    server.ops_sec_idx = 0;
    server.ops_sec_last_sample_time = mstime();
    server.ops_sec_last_sample_ops = 0;
1408
    server.unixtime = time(NULL);
1409
    server.lastbgsave_status = REDIS_OK;
1410
    server.stop_writes_on_bgsave_err = 1;
stamhe's avatar
stamhe 已提交
1411
    if(aeCreateTimeEvent(server.el, 1, serverCron, NULL, NULL) == AE_ERR) {
1412
        redisPanic("Can't create the serverCron time event.");
stamhe's avatar
stamhe 已提交
1413 1414
        exit(1);
    }
1415
    if (server.ipfd > 0 && aeCreateFileEvent(server.el,server.ipfd,AE_READABLE,
A
antirez 已提交
1416
        acceptTcpHandler,NULL) == AE_ERR) redisPanic("Unrecoverable error creating server.ipfd file event.");
1417
    if (server.sofd > 0 && aeCreateFileEvent(server.el,server.sofd,AE_READABLE,
A
antirez 已提交
1418
        acceptUnixHandler,NULL) == AE_ERR) redisPanic("Unrecoverable error creating server.sofd file event.");
1419

1420
    if (server.aof_state == REDIS_AOF_ON) {
A
antirez 已提交
1421
        server.aof_fd = open(server.aof_filename,
1422
                               O_WRONLY|O_APPEND|O_CREAT,0644);
A
antirez 已提交
1423
        if (server.aof_fd == -1) {
1424 1425 1426 1427 1428 1429
            redisLog(REDIS_WARNING, "Can't open the append-only file: %s",
                strerror(errno));
            exit(1);
        }
    }

1430 1431
    /* 32 bit instances are limited to 4GB of address space, so if there is
     * no explicit limit in the user provided configuration we set a limit
1432 1433
     * at 3 GB using maxmemory with 'noeviction' policy'. This avoids
     * useless crashes of the Redis instance for out of memory. */
1434
    if (server.arch_bits == 32 && server.maxmemory == 0) {
1435 1436
        redisLog(REDIS_WARNING,"Warning: 32 bit instance detected but no memory limit set. Setting 3 GB maxmemory limit with 'noeviction' policy now.");
        server.maxmemory = 3072LL*(1024*1024); /* 3 GB */
1437 1438 1439
        server.maxmemory_policy = REDIS_MAXMEMORY_NO_EVICTION;
    }

A
antirez 已提交
1440
    if (server.cluster_enabled) clusterInit();
1441
    scriptingInit();
1442
    slowlogInit();
1443
    bioInit();
1444 1445
}

1446 1447 1448 1449
/* Populates the Redis Command Table starting from the hard coded list
 * we have on top of redis.c file. */
void populateCommandTable(void) {
    int j;
1450
    int numcommands = sizeof(redisCommandTable)/sizeof(struct redisCommand);
1451 1452

    for (j = 0; j < numcommands; j++) {
1453
        struct redisCommand *c = redisCommandTable+j;
1454
        char *f = c->sflags;
1455
        int retval1, retval2;
1456

1457 1458 1459 1460 1461 1462 1463 1464
        while(*f != '\0') {
            switch(*f) {
            case 'w': c->flags |= REDIS_CMD_WRITE; break;
            case 'r': c->flags |= REDIS_CMD_READONLY; break;
            case 'm': c->flags |= REDIS_CMD_DENYOOM; break;
            case 'a': c->flags |= REDIS_CMD_ADMIN; break;
            case 'p': c->flags |= REDIS_CMD_PUBSUB; break;
            case 'f': c->flags |= REDIS_CMD_FORCE_REPLICATION; break;
1465 1466
            case 's': c->flags |= REDIS_CMD_NOSCRIPT; break;
            case 'R': c->flags |= REDIS_CMD_RANDOM; break;
1467
            case 'S': c->flags |= REDIS_CMD_SORT_FOR_SCRIPT; break;
1468 1469
            case 'l': c->flags |= REDIS_CMD_LOADING; break;
            case 't': c->flags |= REDIS_CMD_STALE; break;
1470
            case 'M': c->flags |= REDIS_CMD_SKIP_MONITOR; break;
1471
            case 'k': c->flags |= REDIS_CMD_ASKING; break;
1472 1473 1474 1475 1476
            default: redisPanic("Unsupported command flag"); break;
            }
            f++;
        }

1477 1478 1479 1480 1481
        retval1 = dictAdd(server.commands, sdsnew(c->name), c);
        /* Populate an additional dictionary that will be unaffected
         * by rename-command statements in redis.conf. */
        retval2 = dictAdd(server.orig_commands, sdsnew(c->name), c);
        redisAssert(retval1 == DICT_OK && retval2 == DICT_OK);
1482
    }
1483 1484
}

1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496
void resetCommandTableStats(void) {
    int numcommands = sizeof(redisCommandTable)/sizeof(struct redisCommand);
    int j;

    for (j = 0; j < numcommands; j++) {
        struct redisCommand *c = redisCommandTable+j;

        c->microseconds = 0;
        c->calls = 0;
    }
}

1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533
/* ========================== Redis OP Array API ============================ */

void redisOpArrayInit(redisOpArray *oa) {
    oa->ops = NULL;
    oa->numops = 0;
}

int redisOpArrayAppend(redisOpArray *oa, struct redisCommand *cmd, int dbid,
                       robj **argv, int argc, int target)
{
    redisOp *op;

    oa->ops = zrealloc(oa->ops,sizeof(redisOp)*(oa->numops+1));
    op = oa->ops+oa->numops;
    op->cmd = cmd;
    op->dbid = dbid;
    op->argv = argv;
    op->argc = argc;
    op->target = target;
    oa->numops++;
    return oa->numops;
}

void redisOpArrayFree(redisOpArray *oa) {
    while(oa->numops) {
        int j;
        redisOp *op;

        oa->numops--;
        op = oa->ops+oa->numops;
        for (j = 0; j < op->argc; j++)
            decrRefCount(op->argv[j]);
        zfree(op->argv);
    }
    zfree(oa->ops);
}

1534 1535
/* ====================== Commands lookup and execution ===================== */

1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546
struct redisCommand *lookupCommand(sds name) {
    return dictFetchValue(server.commands, name);
}

struct redisCommand *lookupCommandByCString(char *s) {
    struct redisCommand *cmd;
    sds name = sdsnew(s);

    cmd = dictFetchValue(server.commands, name);
    sdsfree(name);
    return cmd;
1547 1548
}

1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562
/* Lookup the command in the current table, if not found also check in
 * the original table containing the original command names unaffected by
 * redis.conf rename-command statement.
 *
 * This is used by functions rewriting the argument vector such as
 * rewriteClientCommandVector() in order to set client->cmd pointer
 * correctly even if the command was renamed. */
struct redisCommand *lookupCommandOrOriginal(sds name) {
    struct redisCommand *cmd = dictFetchValue(server.commands, name);

    if (!cmd) cmd = dictFetchValue(server.orig_commands,name);
    return cmd;
}

A
antirez 已提交
1563
/* Propagate the specified command (in the context of the specified database id)
1564
 * to AOF and Slaves.
A
antirez 已提交
1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575
 *
 * flags are an xor between:
 * + REDIS_PROPAGATE_NONE (no propagation of command at all)
 * + REDIS_PROPAGATE_AOF (propagate into the AOF file if is enabled)
 * + REDIS_PROPAGATE_REPL (propagate into the replication link)
 */
void propagate(struct redisCommand *cmd, int dbid, robj **argv, int argc,
               int flags)
{
    if (server.aof_state != REDIS_AOF_OFF && flags & REDIS_PROPAGATE_AOF)
        feedAppendOnlyFile(cmd,dbid,argv,argc);
1576
    if (flags & REDIS_PROPAGATE_REPL)
A
antirez 已提交
1577 1578 1579
        replicationFeedSlaves(server.slaves,dbid,argv,argc);
}

1580 1581
/* Used inside commands to schedule the propagation of additional commands
 * after the current command is propagated to AOF / Replication. */
1582 1583 1584
void alsoPropagate(struct redisCommand *cmd, int dbid, robj **argv, int argc,
                   int target)
{
1585
    redisOpArrayAppend(&server.also_propagate,cmd,dbid,argv,argc,target);
1586 1587
}

1588
/* Call() is the core of Redis execution of a command */
1589
void call(redisClient *c, int flags) {
1590
    long long dirty, start = ustime(), duration;
1591

A
antirez 已提交
1592
    /* Sent the command to clients in MONITOR mode, only if the commands are
G
guiquanz 已提交
1593
     * not generated from reading an AOF. */
1594 1595 1596 1597
    if (listLength(server.monitors) &&
        !server.loading &&
        !(c->cmd->flags & REDIS_CMD_SKIP_MONITOR))
    {
1598
        replicationFeedMonitors(c,server.monitors,c->db->id,c->argv,c->argc);
1599
    }
A
antirez 已提交
1600 1601

    /* Call the command. */
1602
    redisOpArrayInit(&server.also_propagate);
1603
    dirty = server.dirty;
1604
    c->cmd->proc(c);
1605
    dirty = server.dirty-dirty;
1606
    duration = ustime()-start;
1607 1608 1609 1610 1611 1612

    /* When EVAL is called loading the AOF we don't want commands called
     * from Lua to go into the slowlog or to populate statistics. */
    if (server.loading && c->flags & REDIS_LUA_CLIENT)
        flags &= ~(REDIS_CALL_SLOWLOG | REDIS_CALL_STATS);

A
antirez 已提交
1613 1614
    /* Log the command into the Slow log if needed, and populate the
     * per-command statistics that we show in INFO commandstats. */
1615
    if (flags & REDIS_CALL_SLOWLOG && c->cmd->proc != execCommand)
1616 1617 1618 1619 1620
        slowlogPushEntryIfNeeded(c->argv,c->argc,duration);
    if (flags & REDIS_CALL_STATS) {
        c->cmd->microseconds += duration;
        c->cmd->calls++;
    }
A
antirez 已提交
1621 1622

    /* Propagate the command into the AOF and replication link */
1623
    if (flags & REDIS_CALL_PROPAGATE) {
A
antirez 已提交
1624 1625 1626 1627 1628 1629 1630 1631
        int flags = REDIS_PROPAGATE_NONE;

        if (c->cmd->flags & REDIS_CMD_FORCE_REPLICATION)
            flags |= REDIS_PROPAGATE_REPL;
        if (dirty)
            flags |= (REDIS_PROPAGATE_REPL | REDIS_PROPAGATE_AOF);
        if (flags != REDIS_PROPAGATE_NONE)
            propagate(c->cmd,c->db->id,c->argv,c->argc,flags);
1632
    }
1633 1634 1635

    /* Handle the alsoPropagate() API to handle commands that want to propagate
     * multiple separated commands. */
1636
    if (server.also_propagate.numops) {
1637
        int j;
1638
        redisOp *rop;
1639

1640 1641 1642 1643 1644
        for (j = 0; j < server.also_propagate.numops; j++) {
            rop = &server.also_propagate.ops[j];
            propagate(rop->cmd, rop->dbid, rop->argv, rop->argc, rop->target);
        }
        redisOpArrayFree(&server.also_propagate);
1645
    }
1646 1647 1648 1649
    server.stat_numcommands++;
}

/* If this function gets called we already read a whole
1650
 * command, arguments are in the client argv/argc fields.
1651 1652 1653 1654
 * processCommand() execute the command or prepare the
 * server for a bulk read from the client.
 *
 * If 1 is returned the client is still alive and valid and
G
guiquanz 已提交
1655 1656
 * other operations can be performed by the caller. Otherwise
 * if 0 is returned the client was destroyed (i.e. after QUIT). */
1657
int processCommand(redisClient *c) {
P
Pieter Noordhuis 已提交
1658 1659 1660 1661
    /* The QUIT command is handled separately. Normal command procs will
     * go through checking for replication and QUIT will cause trouble
     * when FORCE_REPLICATION is enabled and would be implemented in
     * a regular command proc. */
1662
    if (!strcasecmp(c->argv[0]->ptr,"quit")) {
P
Pieter Noordhuis 已提交
1663
        addReply(c,shared.ok);
1664
        c->flags |= REDIS_CLOSE_AFTER_REPLY;
1665
        return REDIS_ERR;
1666 1667 1668
    }

    /* Now lookup the command and check ASAP about trivial error conditions
1669
     * such as wrong arity, bad command name and so forth. */
1670
    c->cmd = c->lastcmd = lookupCommand(c->argv[0]->ptr);
1671
    if (!c->cmd) {
A
antirez 已提交
1672
        flagTransaction(c);
1673 1674
        addReplyErrorFormat(c,"unknown command '%s'",
            (char*)c->argv[0]->ptr);
1675
        return REDIS_OK;
1676 1677
    } else if ((c->cmd->arity > 0 && c->cmd->arity != c->argc) ||
               (c->argc < -c->cmd->arity)) {
A
antirez 已提交
1678
        flagTransaction(c);
1679
        addReplyErrorFormat(c,"wrong number of arguments for '%s' command",
1680
            c->cmd->name);
1681
        return REDIS_OK;
1682 1683 1684
    }

    /* Check if the user is authenticated */
1685 1686
    if (server.requirepass && !c->authenticated && c->cmd->proc != authCommand)
    {
A
antirez 已提交
1687
        flagTransaction(c);
1688
        addReply(c,shared.noautherr);
1689
        return REDIS_OK;
1690 1691
    }

1692 1693 1694 1695
    /* If cluster is enabled perform the cluster redirection here.
     * However we don't perform the redirection if:
     * 1) The sender of this command is our master.
     * 2) The command has no key arguments. */
A
antirez 已提交
1696
    if (server.cluster_enabled &&
1697 1698 1699
        !(c->flags & REDIS_MASTER) &&
        !(c->cmd->getkeys_proc == NULL && c->cmd->firstkey == 0))
    {
A
antirez 已提交
1700 1701
        int hashslot;

1702
        if (server.cluster->state != REDIS_CLUSTER_OK) {
1703
            addReplySds(c,sdsnew("-CLUSTERDOWN The cluster is down. Use CLUSTER INFO for more information\r\n"));
A
antirez 已提交
1704 1705
            return REDIS_OK;
        } else {
1706
            int ask;
1707
            clusterNode *n = getNodeByQuery(c,c->cmd,c->argv,c->argc,&hashslot,&ask);
A
antirez 已提交
1708
            if (n == NULL) {
1709
                addReplyError(c,"Multi keys request invalid in cluster");
A
antirez 已提交
1710
                return REDIS_OK;
1711
            } else if (n != server.cluster->myself) {
A
antirez 已提交
1712
                addReplySds(c,sdscatprintf(sdsempty(),
1713 1714
                    "-%s %d %s:%d\r\n", ask ? "ASK" : "MOVED",
                    hashslot,n->ip,n->port));
A
antirez 已提交
1715 1716 1717 1718 1719
                return REDIS_OK;
            }
        }
    }

1720 1721 1722 1723 1724
    /* Handle the maxmemory directive.
     *
     * First we try to free some memory if possible (if there are volatile
     * keys in the dataset). If there are not the only thing we can do
     * is returning an error. */
1725 1726 1727
    if (server.maxmemory) {
        int retval = freeMemoryIfNeeded();
        if ((c->cmd->flags & REDIS_CMD_DENYOOM) && retval == REDIS_ERR) {
A
antirez 已提交
1728
            flagTransaction(c);
1729
            addReply(c, shared.oomerr);
1730 1731
            return REDIS_OK;
        }
1732 1733
    }

1734
    /* Don't accept write commands if there are problems persisting on disk. */
1735 1736 1737
    if (server.stop_writes_on_bgsave_err &&
        server.saveparamslen > 0
        && server.lastbgsave_status == REDIS_ERR &&
1738 1739
        c->cmd->flags & REDIS_CMD_WRITE)
    {
A
antirez 已提交
1740
        flagTransaction(c);
1741 1742 1743 1744
        addReply(c, shared.bgsaveerr);
        return REDIS_OK;
    }

1745
    /* Don't accept write commands if this is a read only slave. But
1746 1747 1748 1749 1750 1751 1752 1753 1754
     * accept write commands if this is our master. */
    if (server.masterhost && server.repl_slave_ro &&
        !(c->flags & REDIS_MASTER) &&
        c->cmd->flags & REDIS_CMD_WRITE)
    {
        addReply(c, shared.roslaveerr);
        return REDIS_OK;
    }

1755 1756 1757
    /* Only allow SUBSCRIBE and UNSUBSCRIBE in the context of Pub/Sub */
    if ((dictSize(c->pubsub_channels) > 0 || listLength(c->pubsub_patterns) > 0)
        &&
1758 1759 1760 1761
        c->cmd->proc != subscribeCommand &&
        c->cmd->proc != unsubscribeCommand &&
        c->cmd->proc != psubscribeCommand &&
        c->cmd->proc != punsubscribeCommand) {
1762
        addReplyError(c,"only (P)SUBSCRIBE / (P)UNSUBSCRIBE / QUIT allowed in this context");
1763
        return REDIS_OK;
1764 1765
    }

1766 1767
    /* Only allow INFO and SLAVEOF when slave-serve-stale-data is no and
     * we are a slave with a broken link with master. */
A
antirez 已提交
1768
    if (server.masterhost && server.repl_state != REDIS_REPL_CONNECTED &&
1769
        server.repl_serve_stale_data == 0 &&
1770
        !(c->cmd->flags & REDIS_CMD_STALE))
1771
    {
A
antirez 已提交
1772
        flagTransaction(c);
1773
        addReply(c, shared.masterdownerr);
1774 1775 1776
        return REDIS_OK;
    }

1777 1778 1779
    /* Loading DB? Return an error if the command has not the
     * REDIS_CMD_LOADING flag. */
    if (server.loading && !(c->cmd->flags & REDIS_CMD_LOADING)) {
1780 1781 1782 1783
        addReply(c, shared.loadingerr);
        return REDIS_OK;
    }

1784
    /* Lua script too slow? Only allow commands with REDIS_CMD_STALE flag. */
1785
    if (server.lua_timedout &&
1786
          c->cmd->proc != authCommand &&
1787
        !(c->cmd->proc == shutdownCommand &&
1788 1789 1790 1791 1792 1793
          c->argc == 2 &&
          tolower(((char*)c->argv[1]->ptr)[0]) == 'n') &&
        !(c->cmd->proc == scriptCommand &&
          c->argc == 2 &&
          tolower(((char*)c->argv[1]->ptr)[0]) == 'k'))
    {
A
antirez 已提交
1794
        flagTransaction(c);
1795 1796 1797 1798
        addReply(c, shared.slowscripterr);
        return REDIS_OK;
    }

1799 1800
    /* Exec the command */
    if (c->flags & REDIS_MULTI &&
1801 1802
        c->cmd->proc != execCommand && c->cmd->proc != discardCommand &&
        c->cmd->proc != multiCommand && c->cmd->proc != watchCommand)
1803
    {
1804
        queueMultiCommand(c);
1805 1806
        addReply(c,shared.queued);
    } else {
1807
        call(c,REDIS_CALL_FULL);
1808 1809
        if (listLength(server.ready_keys))
            handleClientsBlockedOnLists();
1810
    }
1811
    return REDIS_OK;
1812 1813 1814 1815
}

/*================================== Shutdown =============================== */

1816 1817 1818 1819
int prepareForShutdown(int flags) {
    int save = flags & REDIS_SHUTDOWN_SAVE;
    int nosave = flags & REDIS_SHUTDOWN_NOSAVE;

1820
    redisLog(REDIS_WARNING,"User requested shutdown...");
1821 1822 1823
    /* Kill the saving child if there is a background saving in progress.
       We want to avoid race conditions, for instance our saving child may
       overwrite the synchronous saving did by SHUTDOWN. */
A
antirez 已提交
1824
    if (server.rdb_child_pid != -1) {
1825
        redisLog(REDIS_WARNING,"There is a child saving an .rdb. Killing it!");
1826
        kill(server.rdb_child_pid,SIGUSR1);
A
antirez 已提交
1827
        rdbRemoveTempFile(server.rdb_child_pid);
1828
    }
1829
    if (server.aof_state != REDIS_AOF_OFF) {
1830 1831
        /* Kill the AOF saving child as the AOF we already have may be longer
         * but contains the full dataset anyway. */
A
antirez 已提交
1832
        if (server.aof_child_pid != -1) {
1833 1834
            redisLog(REDIS_WARNING,
                "There is a child rewriting the AOF. Killing it!");
1835
            kill(server.aof_child_pid,SIGUSR1);
1836
        }
1837
        /* Append only file: fsync() the AOF and exit */
1838
        redisLog(REDIS_NOTICE,"Calling fsync() on the AOF file.");
A
antirez 已提交
1839
        aof_fsync(server.aof_fd);
1840
    }
1841
    if ((server.saveparamslen > 0 && !nosave) || save) {
1842
        redisLog(REDIS_NOTICE,"Saving the final RDB snapshot before exiting.");
1843
        /* Snapshotting. Perform a SYNC SAVE and exit */
A
antirez 已提交
1844
        if (rdbSave(server.rdb_filename) != REDIS_OK) {
1845 1846 1847 1848 1849
            /* Ooops.. error saving! The best we can do is to continue
             * operating. Note that if there was a background saving process,
             * in the next cron() Redis will be notified that the background
             * saving aborted, handling special stuff like slaves pending for
             * synchronization... */
1850
            redisLog(REDIS_WARNING,"Error trying to save the DB, can't exit.");
1851 1852 1853
            return REDIS_ERR;
        }
    }
1854 1855 1856 1857
    if (server.daemonize) {
        redisLog(REDIS_NOTICE,"Removing the pid file.");
        unlink(server.pidfile);
    }
1858 1859 1860
    /* Close the listening sockets. Apparently this allows faster restarts. */
    if (server.ipfd != -1) close(server.ipfd);
    if (server.sofd != -1) close(server.sofd);
1861 1862 1863 1864
    if (server.unixsocket) {
        redisLog(REDIS_NOTICE,"Removing the unix socket file.");
        unlink(server.unixsocket); /* don't care if this fails */
    }
1865

1866
    redisLog(REDIS_WARNING,"Redis is now ready to exit, bye bye...");
1867 1868 1869 1870 1871
    return REDIS_OK;
}

/*================================== Commands =============================== */

1872
/* Return zero if strings are the same, non-zero if they are not.
1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913
 * The comparison is performed in a way that prevents an attacker to obtain
 * information about the nature of the strings just monitoring the execution
 * time of the function.
 *
 * Note that limiting the comparison length to strings up to 512 bytes we
 * can avoid leaking any information about the password length and any
 * possible branch misprediction related leak.
 */
int time_independent_strcmp(char *a, char *b) {
    char bufa[REDIS_AUTHPASS_MAX_LEN], bufb[REDIS_AUTHPASS_MAX_LEN];
    /* The above two strlen perform len(a) + len(b) operations where either
     * a or b are fixed (our password) length, and the difference is only
     * relative to the length of the user provided string, so no information
     * leak is possible in the following two lines of code. */
    int alen = strlen(a);
    int blen = strlen(b);
    int j;
    int diff = 0;

    /* We can't compare strings longer than our static buffers.
     * Note that this will never pass the first test in practical circumstances
     * so there is no info leak. */
    if (alen > sizeof(bufa) || blen > sizeof(bufb)) return 1;

    memset(bufa,0,sizeof(bufa));        /* Constant time. */
    memset(bufb,0,sizeof(bufb));        /* Constant time. */
    /* Again the time of the following two copies is proportional to
     * len(a) + len(b) so no info is leaked. */
    memcpy(bufa,a,alen);
    memcpy(bufb,b,blen);

    /* Always compare all the chars in the two buffers without
     * conditional expressions. */
    for (j = 0; j < sizeof(bufa); j++) {
        diff |= (bufa[j] ^ bufb[j]);
    }
    /* Length must be equal as well. */
    diff |= alen ^ blen;
    return diff; /* If zero strings are the same. */
}

1914
void authCommand(redisClient *c) {
1915 1916
    if (!server.requirepass) {
        addReplyError(c,"Client sent AUTH, but no password is set");
1917
    } else if (!time_independent_strcmp(c->argv[1]->ptr, server.requirepass)) {
1918 1919 1920 1921
      c->authenticated = 1;
      addReply(c,shared.ok);
    } else {
      c->authenticated = 0;
1922
      addReplyError(c,"invalid password");
1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933
    }
}

void pingCommand(redisClient *c) {
    addReply(c,shared.pong);
}

void echoCommand(redisClient *c) {
    addReplyBulk(c,c->argv[1]);
}

A
antirez 已提交
1934 1935 1936
void timeCommand(redisClient *c) {
    struct timeval tv;

G
guiquanz 已提交
1937
    /* gettimeofday() can only fail if &tv is a bad address so we
A
antirez 已提交
1938 1939 1940 1941 1942 1943 1944
     * don't check for errors. */
    gettimeofday(&tv,NULL);
    addReplyMultiBulkLen(c,2);
    addReplyBulkLongLong(c,tv.tv_sec);
    addReplyBulkLongLong(c,tv.tv_usec);
}

1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968
/* Convert an amount of bytes into a human readable string in the form
 * of 100B, 2G, 100M, 4K, and so forth. */
void bytesToHuman(char *s, unsigned long long n) {
    double d;

    if (n < 1024) {
        /* Bytes */
        sprintf(s,"%lluB",n);
        return;
    } else if (n < (1024*1024)) {
        d = (double)n/(1024);
        sprintf(s,"%.2fK",d);
    } else if (n < (1024LL*1024*1024)) {
        d = (double)n/(1024*1024);
        sprintf(s,"%.2fM",d);
    } else if (n < (1024LL*1024*1024*1024)) {
        d = (double)n/(1024LL*1024*1024);
        sprintf(s,"%.2fG",d);
    }
}

/* Create the string returned by the INFO command. This is decoupled
 * by the INFO command itself as we need to report the same information
 * on memory corruption problems. */
1969 1970
sds genRedisInfoString(char *section) {
    sds info = sdsempty();
1971
    time_t uptime = server.unixtime-server.stat_starttime;
A
antirez 已提交
1972
    int j, numcommands;
1973
    struct rusage self_ru, c_ru;
1974
    unsigned long lol, bib;
1975 1976
    int allsections = 0, defsections = 0;
    int sections = 0;
1977

1978 1979
    if (section) {
        allsections = strcasecmp(section,"all") == 0;
1980
        defsections = strcasecmp(section,"default") == 0;
1981
    }
1982 1983 1984

    getrusage(RUSAGE_SELF, &self_ru);
    getrusage(RUSAGE_CHILDREN, &c_ru);
1985
    getClientsMaxBuffers(&lol,&bib);
1986 1987 1988

    /* Server */
    if (allsections || defsections || !strcasecmp(section,"server")) {
1989
        struct utsname name;
1990
        char *mode;
1991

1992 1993 1994 1995
        if (server.cluster_enabled) mode = "cluster";
        else if (server.sentinel_mode) mode = "sentinel";
        else mode = "standalone";
    
1996
        if (sections++) info = sdscat(info,"\r\n");
1997
        uname(&name);
1998
        info = sdscatprintf(info,
1999 2000 2001 2002
            "# Server\r\n"
            "redis_version:%s\r\n"
            "redis_git_sha1:%s\r\n"
            "redis_git_dirty:%d\r\n"
2003
            "redis_build_id:%llx\r\n"
2004
            "redis_mode:%s\r\n"
2005
            "os:%s %s %s\r\n"
2006
            "arch_bits:%d\r\n"
2007
            "multiplexing_api:%s\r\n"
A
antirez 已提交
2008
            "gcc_version:%d.%d.%d\r\n"
2009
            "process_id:%ld\r\n"
A
antirez 已提交
2010
            "run_id:%s\r\n"
2011 2012 2013
            "tcp_port:%d\r\n"
            "uptime_in_seconds:%ld\r\n"
            "uptime_in_days:%ld\r\n"
2014
            "hz:%d\r\n"
2015 2016 2017 2018
            "lru_clock:%ld\r\n",
            REDIS_VERSION,
            redisGitSHA1(),
            strtol(redisGitDirty(),NULL,10) > 0,
2019
            (unsigned long long) redisBuildId(),
2020
            mode,
2021
            name.sysname, name.release, name.machine,
2022
            server.arch_bits,
2023
            aeGetApiName(),
A
antirez 已提交
2024 2025 2026 2027 2028
#ifdef __GNUC__
            __GNUC__,__GNUC_MINOR__,__GNUC_PATCHLEVEL__,
#else
            0,0,0,
#endif
2029
            (long) getpid(),
A
antirez 已提交
2030
            server.runid,
2031 2032 2033
            server.port,
            uptime,
            uptime/(3600*24),
2034
            server.hz,
2035 2036 2037 2038 2039 2040 2041 2042
            (unsigned long) server.lruclock);
    }

    /* Clients */
    if (allsections || defsections || !strcasecmp(section,"clients")) {
        if (sections++) info = sdscat(info,"\r\n");
        info = sdscatprintf(info,
            "# Clients\r\n"
2043
            "connected_clients:%lu\r\n"
2044 2045 2046 2047 2048 2049 2050 2051 2052 2053
            "client_longest_output_list:%lu\r\n"
            "client_biggest_input_buf:%lu\r\n"
            "blocked_clients:%d\r\n",
            listLength(server.clients)-listLength(server.slaves),
            lol, bib,
            server.bpop_blocked_clients);
    }

    /* Memory */
    if (allsections || defsections || !strcasecmp(section,"memory")) {
2054 2055 2056 2057 2058
        char hmem[64];
        char peak_hmem[64];

        bytesToHuman(hmem,zmalloc_used_memory());
        bytesToHuman(peak_hmem,server.stat_peak_memory);
2059 2060 2061 2062 2063 2064
        if (sections++) info = sdscat(info,"\r\n");
        info = sdscatprintf(info,
            "# Memory\r\n"
            "used_memory:%zu\r\n"
            "used_memory_human:%s\r\n"
            "used_memory_rss:%zu\r\n"
2065 2066
            "used_memory_peak:%zu\r\n"
            "used_memory_peak_human:%s\r\n"
A
antirez 已提交
2067
            "used_memory_lua:%lld\r\n"
2068
            "mem_fragmentation_ratio:%.2f\r\n"
2069
            "mem_allocator:%s\r\n",
2070 2071 2072
            zmalloc_used_memory(),
            hmem,
            zmalloc_get_rss(),
2073 2074
            server.stat_peak_memory,
            peak_hmem,
A
antirez 已提交
2075
            ((long long)lua_gc(server.lua,LUA_GCCOUNT,0))*1024LL,
2076
            zmalloc_get_fragmentation_ratio(),
2077
            ZMALLOC_LIB
2078
            );
2079 2080
    }

2081 2082 2083
    /* Persistence */
    if (allsections || defsections || !strcasecmp(section,"persistence")) {
        if (sections++) info = sdscat(info,"\r\n");
2084
        info = sdscatprintf(info,
2085 2086
            "# Persistence\r\n"
            "loading:%d\r\n"
2087 2088 2089 2090 2091 2092
            "rdb_changes_since_last_save:%lld\r\n"
            "rdb_bgsave_in_progress:%d\r\n"
            "rdb_last_save_time:%ld\r\n"
            "rdb_last_bgsave_status:%s\r\n"
            "rdb_last_bgsave_time_sec:%ld\r\n"
            "rdb_current_bgsave_time_sec:%ld\r\n"
2093
            "aof_enabled:%d\r\n"
2094 2095 2096
            "aof_rewrite_in_progress:%d\r\n"
            "aof_rewrite_scheduled:%d\r\n"
            "aof_last_rewrite_time_sec:%ld\r\n"
2097 2098
            "aof_current_rewrite_time_sec:%ld\r\n"
            "aof_last_bgrewrite_status:%s\r\n",
2099 2100
            server.loading,
            server.dirty,
A
antirez 已提交
2101
            server.rdb_child_pid != -1,
2102
            server.lastsave,
2103
            (server.lastbgsave_status == REDIS_OK) ? "ok" : "err",
2104 2105 2106 2107
            server.rdb_save_time_last,
            (server.rdb_child_pid == -1) ?
                -1 : time(NULL)-server.rdb_save_time_start,
            server.aof_state != REDIS_AOF_OFF,
2108
            server.aof_child_pid != -1,
2109 2110 2111
            server.aof_rewrite_scheduled,
            server.aof_rewrite_time_last,
            (server.aof_child_pid == -1) ?
2112 2113
                -1 : time(NULL)-server.aof_rewrite_time_start,
            (server.aof_lastbgrewrite_status == REDIS_OK) ? "ok" : "err");
2114

2115
        if (server.aof_state != REDIS_AOF_OFF) {
2116 2117 2118
            info = sdscatprintf(info,
                "aof_current_size:%lld\r\n"
                "aof_base_size:%lld\r\n"
2119 2120
                "aof_pending_rewrite:%d\r\n"
                "aof_buffer_length:%zu\r\n"
2121
                "aof_rewrite_buffer_length:%lu\r\n"
2122 2123
                "aof_pending_bio_fsync:%llu\r\n"
                "aof_delayed_fsync:%lu\r\n",
2124 2125 2126
                (long long) server.aof_current_size,
                (long long) server.aof_rewrite_base_size,
                server.aof_rewrite_scheduled,
A
antirez 已提交
2127
                sdslen(server.aof_buf),
2128
                aofRewriteBufferSize(),
2129 2130
                bioPendingJobsOfType(REDIS_BIO_AOF_FSYNC),
                server.aof_delayed_fsync);
2131 2132
        }

2133 2134 2135 2136 2137 2138 2139 2140 2141
        if (server.loading) {
            double perc;
            time_t eta, elapsed;
            off_t remaining_bytes = server.loading_total_bytes-
                                    server.loading_loaded_bytes;

            perc = ((double)server.loading_loaded_bytes /
                   server.loading_total_bytes) * 100;

2142
            elapsed = server.unixtime-server.loading_start_time;
2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162
            if (elapsed == 0) {
                eta = 1; /* A fake 1 second figure if we don't have
                            enough info */
            } else {
                eta = (elapsed*remaining_bytes)/server.loading_loaded_bytes;
            }

            info = sdscatprintf(info,
                "loading_start_time:%ld\r\n"
                "loading_total_bytes:%llu\r\n"
                "loading_loaded_bytes:%llu\r\n"
                "loading_loaded_perc:%.2f\r\n"
                "loading_eta_seconds:%ld\r\n"
                ,(unsigned long) server.loading_start_time,
                (unsigned long long) server.loading_total_bytes,
                (unsigned long long) server.loading_loaded_bytes,
                perc,
                eta
            );
        }
2163
    }
2164 2165 2166 2167

    /* Stats */
    if (allsections || defsections || !strcasecmp(section,"stats")) {
        if (sections++) info = sdscat(info,"\r\n");
2168
        info = sdscatprintf(info,
2169 2170 2171
            "# Stats\r\n"
            "total_connections_received:%lld\r\n"
            "total_commands_processed:%lld\r\n"
2172
            "instantaneous_ops_per_sec:%lld\r\n"
2173
            "rejected_connections:%lld\r\n"
2174 2175 2176
            "sync_full:%lld\r\n"
            "sync_partial_ok:%lld\r\n"
            "sync_partial_err:%lld\r\n"
2177 2178 2179 2180 2181
            "expired_keys:%lld\r\n"
            "evicted_keys:%lld\r\n"
            "keyspace_hits:%lld\r\n"
            "keyspace_misses:%lld\r\n"
            "pubsub_channels:%ld\r\n"
2182
            "pubsub_patterns:%lu\r\n"
2183 2184
            "latest_fork_usec:%lld\r\n"
            "migrate_cached_sockets:%ld\r\n",
2185 2186
            server.stat_numconnections,
            server.stat_numcommands,
2187
            getOperationsPerSecond(),
2188
            server.stat_rejected_conn,
2189 2190 2191
            server.stat_sync_full,
            server.stat_sync_partial_ok,
            server.stat_sync_partial_err,
2192 2193 2194 2195 2196
            server.stat_expiredkeys,
            server.stat_evictedkeys,
            server.stat_keyspace_hits,
            server.stat_keyspace_misses,
            dictSize(server.pubsub_channels),
2197
            listLength(server.pubsub_patterns),
2198 2199
            server.stat_fork_time,
            dictSize(server.migrate_cached_sockets));
2200
    }
A
antirez 已提交
2201

2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215
    /* Replication */
    if (allsections || defsections || !strcasecmp(section,"replication")) {
        if (sections++) info = sdscat(info,"\r\n");
        info = sdscatprintf(info,
            "# Replication\r\n"
            "role:%s\r\n",
            server.masterhost == NULL ? "master" : "slave");
        if (server.masterhost) {
            info = sdscatprintf(info,
                "master_host:%s\r\n"
                "master_port:%d\r\n"
                "master_link_status:%s\r\n"
                "master_last_io_seconds_ago:%d\r\n"
                "master_sync_in_progress:%d\r\n"
2216
                "slave_repl_offset:%lld\r\n"
2217 2218
                ,server.masterhost,
                server.masterport,
A
antirez 已提交
2219
                (server.repl_state == REDIS_REPL_CONNECTED) ?
2220 2221
                    "up" : "down",
                server.master ?
2222
                ((int)(server.unixtime-server.master->lastinteraction)) : -1,
2223 2224
                server.repl_state == REDIS_REPL_TRANSFER,
                server.master ? server.master->reploff : -1
2225 2226
            );

A
antirez 已提交
2227
            if (server.repl_state == REDIS_REPL_TRANSFER) {
2228
                info = sdscatprintf(info,
2229
                    "master_sync_left_bytes:%lld\r\n"
2230
                    "master_sync_last_io_seconds_ago:%d\r\n"
2231 2232
                    , (long long)
                        (server.repl_transfer_size - server.repl_transfer_read),
2233
                    (int)(server.unixtime-server.repl_transfer_lastio)
2234 2235
                );
            }
2236

A
antirez 已提交
2237
            if (server.repl_state != REDIS_REPL_CONNECTED) {
2238 2239
                info = sdscatprintf(info,
                    "master_link_down_since_seconds:%ld\r\n",
2240
                    (long)server.unixtime-server.repl_down_since);
2241
            }
2242
            info = sdscatprintf(info,
2243 2244 2245 2246
                "slave_priority:%d\r\n"
                "slave_read_only:%d\r\n",
                server.slave_priority,
                server.repl_slave_ro);
A
antirez 已提交
2247
        }
2248
        info = sdscatprintf(info,
2249
            "connected_slaves:%lu\r\n",
2250
            listLength(server.slaves));
2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277
        if (listLength(server.slaves)) {
            int slaveid = 0;
            listNode *ln;
            listIter li;

            listRewind(server.slaves,&li);
            while((ln = listNext(&li))) {
                redisClient *slave = listNodeValue(ln);
                char *state = NULL;
                char ip[32];
                int port;

                if (anetPeerToString(slave->fd,ip,&port) == -1) continue;
                switch(slave->replstate) {
                case REDIS_REPL_WAIT_BGSAVE_START:
                case REDIS_REPL_WAIT_BGSAVE_END:
                    state = "wait_bgsave";
                    break;
                case REDIS_REPL_SEND_BULK:
                    state = "send_bulk";
                    break;
                case REDIS_REPL_ONLINE:
                    state = "online";
                    break;
                }
                if (state == NULL) continue;
                info = sdscatprintf(info,"slave%d:%s,%d,%s\r\n",
A
antirez 已提交
2278
                    slaveid,ip,slave->slave_listening_port,state);
2279 2280 2281
                slaveid++;
            }
        }
2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292
        info = sdscatprintf(info,
            "master_repl_offset:%lld\r\n"
            "repl_backlog_active:%d\r\n"
            "repl_backlog_size:%lld\r\n"
            "repl_backlog_first_byte_offset:%lld\r\n"
            "repl_backlog_histlen:%lld\r\n",
            server.master_repl_offset,
            server.repl_backlog != NULL,
            server.repl_backlog_size,
            server.repl_backlog_off,
            server.repl_backlog_histlen);
A
antirez 已提交
2293 2294
    }

2295 2296
    /* CPU */
    if (allsections || defsections || !strcasecmp(section,"cpu")) {
2297 2298
        if (sections++) info = sdscat(info,"\r\n");
        info = sdscatprintf(info,
2299
        "# CPU\r\n"
2300 2301
        "used_cpu_sys:%.2f\r\n"
        "used_cpu_user:%.2f\r\n"
2302 2303
        "used_cpu_sys_children:%.2f\r\n"
        "used_cpu_user_children:%.2f\r\n",
2304
        (float)self_ru.ru_stime.tv_sec+(float)self_ru.ru_stime.tv_usec/1000000,
2305 2306 2307
        (float)self_ru.ru_utime.tv_sec+(float)self_ru.ru_utime.tv_usec/1000000,
        (float)c_ru.ru_stime.tv_sec+(float)c_ru.ru_stime.tv_usec/1000000,
        (float)c_ru.ru_utime.tv_sec+(float)c_ru.ru_utime.tv_usec/1000000);
2308
    }
2309

2310 2311 2312 2313
    /* cmdtime */
    if (allsections || !strcasecmp(section,"commandstats")) {
        if (sections++) info = sdscat(info,"\r\n");
        info = sdscatprintf(info, "# Commandstats\r\n");
2314
        numcommands = sizeof(redisCommandTable)/sizeof(struct redisCommand);
2315
        for (j = 0; j < numcommands; j++) {
2316
            struct redisCommand *c = redisCommandTable+j;
2317

2318 2319 2320 2321 2322
            if (!c->calls) continue;
            info = sdscatprintf(info,
                "cmdstat_%s:calls=%lld,usec=%lld,usec_per_call=%.2f\r\n",
                c->name, c->calls, c->microseconds,
                (c->calls == 0) ? 0 : ((float)c->microseconds/c->calls));
2323
        }
A
antirez 已提交
2324 2325
    }

2326
    /* Cluster */
2327 2328 2329 2330 2331 2332 2333 2334
    if (allsections || defsections || !strcasecmp(section,"cluster")) {
        if (sections++) info = sdscat(info,"\r\n");
        info = sdscatprintf(info,
        "# Cluster\r\n"
        "cluster_enabled:%d\r\n",
        server.cluster_enabled);
    }

2335 2336 2337 2338 2339 2340
    /* Key space */
    if (allsections || defsections || !strcasecmp(section,"keyspace")) {
        if (sections++) info = sdscat(info,"\r\n");
        info = sdscatprintf(info, "# Keyspace\r\n");
        for (j = 0; j < server.dbnum; j++) {
            long long keys, vkeys;
2341

2342 2343 2344 2345 2346 2347
            keys = dictSize(server.db[j].dict);
            vkeys = dictSize(server.db[j].expires);
            if (keys || vkeys) {
                info = sdscatprintf(info, "db%d:keys=%lld,expires=%lld\r\n",
                    j, keys, vkeys);
            }
2348 2349 2350 2351 2352 2353
        }
    }
    return info;
}

void infoCommand(redisClient *c) {
2354 2355 2356 2357 2358 2359 2360
    char *section = c->argc == 2 ? c->argv[1]->ptr : "default";

    if (c->argc > 2) {
        addReply(c,shared.syntaxerr);
        return;
    }
    sds info = genRedisInfoString(section);
2361 2362 2363 2364 2365 2366 2367
    addReplySds(c,sdscatprintf(sdsempty(),"$%lu\r\n",
        (unsigned long)sdslen(info)));
    addReplySds(c,info);
    addReply(c,shared.crlf);
}

void monitorCommand(redisClient *c) {
2368
    /* ignore MONITOR if already slave or in monitor mode */
2369 2370 2371 2372 2373 2374 2375 2376 2377 2378
    if (c->flags & REDIS_SLAVE) return;

    c->flags |= (REDIS_SLAVE|REDIS_MONITOR);
    listAddNodeTail(server.monitors,c);
    addReply(c,shared.ok);
}

/* ============================ Maxmemory directive  ======================== */

/* This function gets called when 'maxmemory' is set on the config file to limit
2379
 * the max memory used by the server, before processing a command.
2380
 *
2381 2382
 * The goal of the function is to free enough memory to keep Redis under the
 * configured memory limit.
2383
 *
2384 2385 2386 2387 2388 2389 2390 2391
 * The function starts calculating how many bytes should be freed to keep
 * Redis under the limit, and enters a loop selecting the best keys to
 * evict accordingly to the configured policy.
 *
 * If all the bytes needed to return back under the limit were freed the
 * function returns REDIS_OK, otherwise REDIS_ERR is returned, and the caller
 * should block the execution of commands that will result in more memory
 * used by the server.
2392
 */
2393 2394 2395 2396
int freeMemoryIfNeeded(void) {
    size_t mem_used, mem_tofree, mem_freed;
    int slaves = listLength(server.slaves);

2397 2398
    /* Remove the size of slaves output buffers and AOF buffer from the
     * count of used memory. */
2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413
    mem_used = zmalloc_used_memory();
    if (slaves) {
        listIter li;
        listNode *ln;

        listRewind(server.slaves,&li);
        while((ln = listNext(&li))) {
            redisClient *slave = listNodeValue(ln);
            unsigned long obuf_bytes = getClientOutputBufferMemoryUsage(slave);
            if (obuf_bytes > mem_used)
                mem_used = 0;
            else
                mem_used -= obuf_bytes;
        }
    }
2414 2415
    if (server.aof_state != REDIS_AOF_OFF) {
        mem_used -= sdslen(server.aof_buf);
2416
        mem_used -= aofRewriteBufferSize();
2417
    }
2418

2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429
    /* Check if we are over the memory limit. */
    if (mem_used <= server.maxmemory) return REDIS_OK;

    if (server.maxmemory_policy == REDIS_MAXMEMORY_NO_EVICTION)
        return REDIS_ERR; /* We need to free memory, but policy forbids. */

    /* Compute how much memory we need to free. */
    mem_tofree = mem_used - server.maxmemory;
    mem_freed = 0;
    while (mem_freed < mem_tofree) {
        int j, k, keys_freed = 0;
2430

2431
        for (j = 0; j < server.dbnum; j++) {
2432
            long bestval = 0; /* just to prevent warning */
2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451
            sds bestkey = NULL;
            struct dictEntry *de;
            redisDb *db = server.db+j;
            dict *dict;

            if (server.maxmemory_policy == REDIS_MAXMEMORY_ALLKEYS_LRU ||
                server.maxmemory_policy == REDIS_MAXMEMORY_ALLKEYS_RANDOM)
            {
                dict = server.db[j].dict;
            } else {
                dict = server.db[j].expires;
            }
            if (dictSize(dict) == 0) continue;

            /* volatile-random and allkeys-random policy */
            if (server.maxmemory_policy == REDIS_MAXMEMORY_ALLKEYS_RANDOM ||
                server.maxmemory_policy == REDIS_MAXMEMORY_VOLATILE_RANDOM)
            {
                de = dictGetRandomKey(dict);
2452
                bestkey = dictGetKey(de);
2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464
            }

            /* volatile-lru and allkeys-lru policy */
            else if (server.maxmemory_policy == REDIS_MAXMEMORY_ALLKEYS_LRU ||
                server.maxmemory_policy == REDIS_MAXMEMORY_VOLATILE_LRU)
            {
                for (k = 0; k < server.maxmemory_samples; k++) {
                    sds thiskey;
                    long thisval;
                    robj *o;

                    de = dictGetRandomKey(dict);
2465
                    thiskey = dictGetKey(de);
2466
                    /* When policy is volatile-lru we need an additional lookup
2467 2468 2469
                     * to locate the real key, as dict is set to db->expires. */
                    if (server.maxmemory_policy == REDIS_MAXMEMORY_VOLATILE_LRU)
                        de = dictFind(db->dict, thiskey);
2470
                    o = dictGetVal(de);
2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487
                    thisval = estimateObjectIdleTime(o);

                    /* Higher idle time is better candidate for deletion */
                    if (bestkey == NULL || thisval > bestval) {
                        bestkey = thiskey;
                        bestval = thisval;
                    }
                }
            }

            /* volatile-ttl */
            else if (server.maxmemory_policy == REDIS_MAXMEMORY_VOLATILE_TTL) {
                for (k = 0; k < server.maxmemory_samples; k++) {
                    sds thiskey;
                    long thisval;

                    de = dictGetRandomKey(dict);
2488 2489
                    thiskey = dictGetKey(de);
                    thisval = (long) dictGetVal(de);
2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501

                    /* Expire sooner (minor expire unix timestamp) is better
                     * candidate for deletion */
                    if (bestkey == NULL || thisval < bestval) {
                        bestkey = thiskey;
                        bestval = thisval;
                    }
                }
            }

            /* Finally remove the selected key. */
            if (bestkey) {
2502 2503
                long long delta;

2504
                robj *keyobj = createStringObject(bestkey,sdslen(bestkey));
2505
                propagateExpire(db,keyobj);
2506 2507 2508 2509 2510 2511 2512 2513 2514
                /* We compute the amount of memory freed by dbDelete() alone.
                 * It is possible that actually the memory needed to propagate
                 * the DEL in AOF and replication link is greater than the one
                 * we are freeing removing the key, but we can't account for
                 * that otherwise we would never exit the loop.
                 *
                 * AOF and Output buffer memory will be freed eventually so
                 * we only care about memory used by the key space. */
                delta = (long long) zmalloc_used_memory();
2515
                dbDelete(db,keyobj);
2516 2517
                delta -= (long long) zmalloc_used_memory();
                mem_freed += delta;
2518
                server.stat_evictedkeys++;
2519 2520
                notifyKeyspaceEvent(REDIS_NOTIFY_EVICTED, "evicted",
                    keyobj, db->id);
2521
                decrRefCount(keyobj);
2522 2523 2524 2525 2526 2527
                keys_freed++;

                /* When the memory to free starts to be big enough, we may
                 * start spending so much time here that is impossible to
                 * deliver data to the slaves fast enough, so we force the
                 * transmission here inside the loop. */
A
antirez 已提交
2528
                if (slaves) flushSlavesOutputBuffers();
2529 2530
            }
        }
2531
        if (!keys_freed) return REDIS_ERR; /* nothing to free... */
2532
    }
2533
    return REDIS_OK;
2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559
}

/* =================================== Main! ================================ */

#ifdef __linux__
int linuxOvercommitMemoryValue(void) {
    FILE *fp = fopen("/proc/sys/vm/overcommit_memory","r");
    char buf[64];

    if (!fp) return -1;
    if (fgets(buf,64,fp) == NULL) {
        fclose(fp);
        return -1;
    }
    fclose(fp);

    return atoi(buf);
}

void linuxOvercommitMemoryWarning(void) {
    if (linuxOvercommitMemoryValue() == 0) {
        redisLog(REDIS_WARNING,"WARNING overcommit_memory is set to 0! Background save may fail under low memory condition. To fix this issue add 'vm.overcommit_memory = 1' to /etc/sysctl.conf and then reboot or run the command 'sysctl vm.overcommit_memory=1' for this to take effect.");
    }
}
#endif /* __linux__ */

2560 2561 2562 2563
void createPidFile(void) {
    /* Try to write the pid file in a best-effort way. */
    FILE *fp = fopen(server.pidfile,"w");
    if (fp) {
2564
        fprintf(fp,"%d\n",(int)getpid());
2565 2566 2567 2568
        fclose(fp);
    }
}

2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586
void daemonize(void) {
    int fd;

    if (fork() != 0) exit(0); /* parent exits */
    setsid(); /* create a new session */

    /* Every output goes to /dev/null. If Redis is daemonized but
     * the 'logfile' is set to 'stdout' in the configuration file
     * it will not log at all. */
    if ((fd = open("/dev/null", O_RDWR, 0)) != -1) {
        dup2(fd, STDIN_FILENO);
        dup2(fd, STDOUT_FILENO);
        dup2(fd, STDERR_FILENO);
        if (fd > STDERR_FILENO) close(fd);
    }
}

void version() {
2587
    printf("Redis server v=%s sha=%s:%d malloc=%s bits=%d build=%llx\n",
A
antirez 已提交
2588 2589 2590 2591
        REDIS_VERSION,
        redisGitSHA1(),
        atoi(redisGitDirty()) > 0,
        ZMALLOC_LIB,
2592
        sizeof(long) == 4 ? 32 : 64,
2593
        (unsigned long long) redisBuildId());
2594 2595 2596 2597
    exit(0);
}

void usage() {
2598
    fprintf(stderr,"Usage: ./redis-server [/path/to/redis.conf] [options]\n");
2599
    fprintf(stderr,"       ./redis-server - (read config from stdin)\n");
2600
    fprintf(stderr,"       ./redis-server -v or --version\n");
2601 2602
    fprintf(stderr,"       ./redis-server -h or --help\n");
    fprintf(stderr,"       ./redis-server --test-memory <megabytes>\n\n");
2603 2604 2605 2606 2607
    fprintf(stderr,"Examples:\n");
    fprintf(stderr,"       ./redis-server (run the server with default conf)\n");
    fprintf(stderr,"       ./redis-server /etc/redis/6379.conf\n");
    fprintf(stderr,"       ./redis-server --port 7777\n");
    fprintf(stderr,"       ./redis-server --port 7777 --slaveof 127.0.0.1 8888\n");
2608 2609 2610
    fprintf(stderr,"       ./redis-server /etc/myredis.conf --loglevel verbose\n\n");
    fprintf(stderr,"Sentinel mode:\n");
    fprintf(stderr,"       ./redis-server /etc/sentinel.conf --sentinel\n");
2611 2612 2613
    exit(1);
}

A
antirez 已提交
2614 2615 2616
void redisAsciiArt(void) {
#include "asciilogo.h"
    char *buf = zmalloc(1024*16);
2617 2618 2619 2620
    char *mode = "stand alone";

    if (server.cluster_enabled) mode = "cluster";
    else if (server.sentinel_mode) mode = "sentinel";
A
antirez 已提交
2621 2622 2623 2624 2625 2626

    snprintf(buf,1024*16,ascii_logo,
        REDIS_VERSION,
        redisGitSHA1(),
        strtol(redisGitDirty(),NULL,10) > 0,
        (sizeof(long) == 8) ? "64" : "32",
2627
        mode, server.port,
A
antirez 已提交
2628 2629 2630 2631 2632 2633
        (long) getpid()
    );
    redisLogRaw(REDIS_NOTICE|REDIS_LOG_RAW,buf);
    zfree(buf);
}

2634
static void sigtermHandler(int sig) {
2635 2636
    REDIS_NOTUSED(sig);

A
antirez 已提交
2637
    redisLogFromHandler(REDIS_WARNING,"Received SIGTERM, scheduling shutdown...");
2638 2639 2640
    server.shutdown_asap = 1;
}

2641
void setupSignalHandlers(void) {
2642 2643
    struct sigaction act;

2644 2645 2646
    /* When the SA_SIGINFO flag is set in sa_flags then sa_sigaction is used.
     * Otherwise, sa_handler is used. */
    sigemptyset(&act.sa_mask);
2647
    act.sa_flags = 0;
2648
    act.sa_handler = sigtermHandler;
2649
    sigaction(SIGTERM, &act, NULL);
2650

2651 2652
#ifdef HAVE_BACKTRACE
    sigemptyset(&act.sa_mask);
2653
    act.sa_flags = SA_NODEFER | SA_RESETHAND | SA_SIGINFO;
2654 2655 2656 2657 2658 2659 2660
    act.sa_sigaction = sigsegvHandler;
    sigaction(SIGSEGV, &act, NULL);
    sigaction(SIGBUS, &act, NULL);
    sigaction(SIGFPE, &act, NULL);
    sigaction(SIGILL, &act, NULL);
#endif
    return;
2661 2662
}

2663 2664
void memtest(size_t megabytes, int passes);

2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692
/* Returns 1 if there is --sentinel among the arguments or if
 * argv[0] is exactly "redis-sentinel". */
int checkForSentinelMode(int argc, char **argv) {
    int j;

    if (strstr(argv[0],"redis-sentinel") != NULL) return 1;
    for (j = 1; j < argc; j++)
        if (!strcmp(argv[j],"--sentinel")) return 1;
    return 0;
}

/* Function called at startup to load RDB or AOF file in memory. */
void loadDataFromDisk(void) {
    long long start = ustime();
    if (server.aof_state == REDIS_AOF_ON) {
        if (loadAppendOnlyFile(server.aof_filename) == REDIS_OK)
            redisLog(REDIS_NOTICE,"DB loaded from append only file: %.3f seconds",(float)(ustime()-start)/1000000);
    } else {
        if (rdbLoad(server.rdb_filename) == REDIS_OK) {
            redisLog(REDIS_NOTICE,"DB loaded from disk: %.3f seconds",
                (float)(ustime()-start)/1000000);
        } else if (errno != ENOENT) {
            redisLog(REDIS_WARNING,"Fatal error loading the DB. Exiting.");
            exit(1);
        }
    }
}

A
antirez 已提交
2693 2694 2695 2696 2697 2698
void redisOutOfMemoryHandler(size_t allocation_size) {
    redisLog(REDIS_WARNING,"Out Of Memory allocating %zu bytes!",
        allocation_size);
    redisPanic("OOM");
}

2699
void redisSetProcTitle(char *title) {
2700
#ifdef USE_SETPROCTITLE
2701 2702 2703 2704
    setproctitle("%s %s:%d",
        title,
        server.bindaddr ? server.bindaddr : "*",
        server.port);
2705 2706 2707
#else
    REDIS_NOTUSED(title);
#endif
2708 2709
}

2710
int main(int argc, char **argv) {
2711
    struct timeval tv;
2712

A
antirez 已提交
2713
    /* We need to initialize our libraries, and the server configuration. */
2714 2715 2716
#ifdef INIT_SETPROCTITLE_REPLACEMENT
    spt_init(argc, argv);
#endif
2717
    zmalloc_enable_thread_safeness();
A
antirez 已提交
2718
    zmalloc_set_oom_handler(redisOutOfMemoryHandler);
2719 2720 2721
    srand(time(NULL)^getpid());
    gettimeofday(&tv,NULL);
    dictSetHashFunctionSeed(tv.tv_sec^tv.tv_usec^getpid());
2722
    server.sentinel_mode = checkForSentinelMode(argc,argv);
2723
    initServerConfig();
2724

2725 2726 2727 2728 2729 2730 2731 2732
    /* We need to init sentinel right now as parsing the configuration file
     * in sentinel mode will have the effect of populating the sentinel
     * data structures with master nodes to monitor. */
    if (server.sentinel_mode) {
        initSentinelConfig();
        initSentinel();
    }

2733 2734 2735 2736 2737 2738
    if (argc >= 2) {
        int j = 1; /* First option to parse in argv[] */
        sds options = sdsempty();
        char *configfile = NULL;

        /* Handle special options --help and --version */
2739 2740
        if (strcmp(argv[1], "-v") == 0 ||
            strcmp(argv[1], "--version") == 0) version();
2741 2742
        if (strcmp(argv[1], "--help") == 0 ||
            strcmp(argv[1], "-h") == 0) usage();
2743 2744
        if (strcmp(argv[1], "--test-memory") == 0) {
            if (argc == 3) {
2745
                memtest(atoi(argv[2]),50);
2746 2747 2748 2749 2750 2751 2752 2753
                exit(0);
            } else {
                fprintf(stderr,"Please specify the amount of memory to test in megabytes.\n");
                fprintf(stderr,"Example: ./redis-server --test-memory 4096\n\n");
                exit(1);
            }
        }

2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773
        /* First argument is the config file name? */
        if (argv[j][0] != '-' || argv[j][1] != '-')
            configfile = argv[j++];
        /* All the other options are parsed and conceptually appended to the
         * configuration file. For instance --port 6380 will generate the
         * string "port 6380\n" to be parsed after the actual file name
         * is parsed, if any. */
        while(j != argc) {
            if (argv[j][0] == '-' && argv[j][1] == '-') {
                /* Option name */
                if (sdslen(options)) options = sdscat(options,"\n");
                options = sdscat(options,argv[j]+2);
                options = sdscat(options," ");
            } else {
                /* Option argument */
                options = sdscatrepr(options,argv[j],strlen(argv[j]));
                options = sdscat(options," ");
            }
            j++;
        }
2774
        resetServerSaveParams();
2775 2776
        loadServerConfig(configfile,options);
        sdsfree(options);
2777
    } else {
2778
        redisLog(REDIS_WARNING, "Warning: no config file specified, using the default config. In order to specify a config file use %s /path/to/%s.conf", argv[0], server.sentinel_mode ? "sentinel" : "redis");
2779 2780 2781 2782
    }
    if (server.daemonize) daemonize();
    initServer();
    if (server.daemonize) createPidFile();
2783
    redisSetProcTitle(argv[0]);
2784
    redisAsciiArt();
2785 2786

    if (!server.sentinel_mode) {
A
antirez 已提交
2787
        /* Things not needed when running in Sentinel mode. */
2788 2789 2790 2791 2792
        redisLog(REDIS_WARNING,"Server started, Redis version " REDIS_VERSION);
    #ifdef __linux__
        linuxOvercommitMemoryWarning();
    #endif
        loadDataFromDisk();
2793 2794 2795 2796 2797 2798 2799 2800
        if (server.cluster_enabled) {
            if (verifyClusterConfigWithData() == REDIS_ERR) {
                redisLog(REDIS_WARNING,
                    "You can't have keys in a DB different than DB 0 when in "
                    "Cluster mode. Exiting.");
                exit(1);
            }
        }
2801 2802 2803 2804
        if (server.ipfd > 0)
            redisLog(REDIS_NOTICE,"The server is now ready to accept connections on port %d", server.port);
        if (server.sofd > 0)
            redisLog(REDIS_NOTICE,"The server is now ready to accept connections at %s", server.unixsocket);
2805
    }
2806

2807 2808 2809 2810 2811
    /* Warning the user about suspicious maxmemory setting. */
    if (server.maxmemory > 0 && server.maxmemory < 1024*1024) {
        redisLog(REDIS_WARNING,"WARNING: You specified a maxmemory value that is less than 1MB (current value is %llu bytes). Are you sure this is what you really want?", server.maxmemory);
    }

2812 2813 2814 2815 2816 2817
    aeSetBeforeSleepProc(server.el,beforeSleep);
    aeMain(server.el);
    aeDeleteEventLoop(server.el);
    return 0;
}

2818
/* The End */