redis.c 131.2 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 "cluster.h"
32
#include "slowlog.h"
33
#include "bio.h"
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50

#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>
51
#include <sys/resource.h>
52
#include <sys/utsname.h>
53
#include <locale.h>
54 55 56 57 58

/* Our shared "common" objects */

struct sharedObjectsStruct shared;

59
/* Global vars that are actually used as constants. The following double
60 61 62 63 64 65 66 67 68 69
 * 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;
70

71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95
/* 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.
96 97 98 99 100 101 102 103
 *
 * 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
104
 * f: force replication of this command, regardless of server.dirty.
105
 * s: command not allowed in scripts.
106
 * R: random command. Command is not deterministic, that is, the same command
107
 *    with the same arguments, with the same key space, may have different
108 109 110
 *    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.
111 112 113 114
 * 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.
115
 * M: Do not automatically propagate the command on MONITOR.
116 117
 * k: Perform an implicit ASKING for this command, so the command will be
 *    accepted in cluster mode if the slot is marked as 'importing'.
118
 */
119
struct redisCommand redisCommandTable[] = {
120
    {"get",getCommand,2,"r",0,NULL,1,1,1,0,0},
121 122 123 124
    {"set",setCommand,-3,"wm",0,NULL,1,1,1,0,0},
    {"setnx",setnxCommand,3,"wm",0,NULL,1,1,1,0,0},
    {"setex",setexCommand,4,"wm",0,NULL,1,1,1,0,0},
    {"psetex",psetexCommand,4,"wm",0,NULL,1,1,1,0,0},
125 126
    {"append",appendCommand,3,"wm",0,NULL,1,1,1,0,0},
    {"strlen",strlenCommand,2,"r",0,NULL,1,1,1,0,0},
127
    {"del",delCommand,-2,"w",0,NULL,1,-1,1,0,0},
128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143
    {"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
144 145 146
    {"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},
147 148 149 150 151 152 153 154 155 156 157 158
    {"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},
159
    {"spop",spopCommand,2,"wRs",0,NULL,1,1,1,0,0},
160
    {"srandmember",srandmemberCommand,-2,"rR",0,NULL,1,1,1,0,0},
161
    {"sinter",sinterCommand,-2,"rS",0,NULL,1,-1,1,0,0},
162
    {"sinterstore",sinterstoreCommand,-3,"wm",0,NULL,1,-1,1,0,0},
163
    {"sunion",sunionCommand,-2,"rS",0,NULL,1,-1,1,0,0},
164
    {"sunionstore",sunionstoreCommand,-3,"wm",0,NULL,1,-1,1,0,0},
165
    {"sdiff",sdiffCommand,-2,"rS",0,NULL,1,-1,1,0,0},
166
    {"sdiffstore",sdiffstoreCommand,-3,"wm",0,NULL,1,-1,1,0,0},
167
    {"smembers",sinterCommand,2,"rS",0,NULL,1,1,1,0,0},
A
antirez 已提交
168
    {"sscan",sscanCommand,-3,"rR",0,NULL,1,1,1,0,0},
169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184
    {"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},
A
antirez 已提交
185
    {"zscan",zscanCommand,-3,"rR",0,NULL,1,1,1,0,0},
186 187 188 189 190 191
    {"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
192
    {"hincrbyfloat",hincrbyfloatCommand,4,"wm",0,NULL,1,1,1,0,0},
193 194
    {"hdel",hdelCommand,-3,"w",0,NULL,1,1,1,0,0},
    {"hlen",hlenCommand,2,"r",0,NULL,1,1,1,0,0},
195 196
    {"hkeys",hkeysCommand,2,"rS",0,NULL,1,1,1,0,0},
    {"hvals",hvalsCommand,2,"rS",0,NULL,1,1,1,0,0},
197 198
    {"hgetall",hgetallCommand,2,"r",0,NULL,1,1,1,0,0},
    {"hexists",hexistsCommand,3,"r",0,NULL,1,1,1,0,0},
A
antirez 已提交
199
    {"hscan",hscanCommand,-3,"rR",0,NULL,1,1,1,0,0},
200 201
    {"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
202
    {"incrbyfloat",incrbyfloatCommand,3,"wm",0,NULL,1,1,1,0,0},
203 204 205
    {"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},
206
    {"randomkey",randomkeyCommand,1,"rR",0,NULL,0,0,0,0,0},
A
antirez 已提交
207
    {"select",selectCommand,2,"rl",0,NULL,0,0,0,0,0},
208
    {"move",moveCommand,3,"w",0,NULL,1,1,1,0,0},
209 210
    {"rename",renameCommand,3,"w",0,NULL,1,2,1,0,0},
    {"renamenx",renamenxCommand,3,"w",0,NULL,1,2,1,0,0},
211 212 213 214
    {"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},
215
    {"keys",keysCommand,2,"rS",0,NULL,0,0,0,0,0},
216
    {"scan",scanCommand,-2,"rR",0,NULL,0,0,0,0,0},
217
    {"dbsize",dbsizeCommand,1,"r",0,NULL,0,0,0,0,0},
218 219
    {"auth",authCommand,2,"rslt",0,NULL,0,0,0,0,0},
    {"ping",pingCommand,1,"rt",0,NULL,0,0,0,0,0},
220
    {"echo",echoCommand,2,"r",0,NULL,0,0,0,0,0},
A
antirez 已提交
221
    {"save",saveCommand,1,"ars",0,NULL,0,0,0,0,0},
222 223
    {"bgsave",bgsaveCommand,1,"ar",0,NULL,0,0,0,0,0},
    {"bgrewriteaof",bgrewriteaofCommand,1,"ar",0,NULL,0,0,0,0,0},
224
    {"shutdown",shutdownCommand,-1,"arlt",0,NULL,0,0,0,0,0},
A
antirez 已提交
225
    {"lastsave",lastsaveCommand,1,"rR",0,NULL,0,0,0,0,0},
226
    {"type",typeCommand,2,"r",0,NULL,1,1,1,0,0},
227
    {"multi",multiCommand,1,"rs",0,NULL,0,0,0,0,0},
228
    {"exec",execCommand,1,"sM",0,NULL,0,0,0,0,0},
229 230
    {"discard",discardCommand,1,"rs",0,NULL,0,0,0,0,0},
    {"sync",syncCommand,1,"ars",0,NULL,0,0,0,0,0},
231
    {"psync",syncCommand,3,"ars",0,NULL,0,0,0,0,0},
A
antirez 已提交
232
    {"replconf",replconfCommand,-1,"arslt",0,NULL,0,0,0,0,0},
233 234
    {"flushdb",flushdbCommand,1,"w",0,NULL,0,0,0,0,0},
    {"flushall",flushallCommand,1,"w",0,NULL,0,0,0,0,0},
235
    {"sort",sortCommand,-2,"wm",0,sortGetKeys,1,1,1,0,0},
236
    {"info",infoCommand,-1,"rlt",0,NULL,0,0,0,0,0},
237
    {"monitor",monitorCommand,1,"ars",0,NULL,0,0,0,0,0},
238 239
    {"ttl",ttlCommand,2,"r",0,NULL,1,1,1,0,0},
    {"pttl",pttlCommand,2,"r",0,NULL,1,1,1,0,0},
240
    {"persist",persistCommand,2,"w",0,NULL,1,1,1,0,0},
241
    {"slaveof",slaveofCommand,3,"ast",0,NULL,0,0,0,0,0},
242
    {"debug",debugCommand,-2,"as",0,NULL,0,0,0,0,0},
243
    {"config",configCommand,-2,"art",0,NULL,0,0,0,0,0},
244 245 246 247
    {"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},
A
antirez 已提交
248
    {"publish",publishCommand,3,"pltr",0,NULL,0,0,0,0,0},
A
antirez 已提交
249
    {"pubsub",pubsubCommand,-2,"pltrR",0,NULL,0,0,0,0,0},
250
    {"watch",watchCommand,-2,"rs",0,NULL,1,-1,1,0,0},
251
    {"unwatch",unwatchCommand,1,"rs",0,NULL,0,0,0,0,0},
252
    {"cluster",clusterCommand,-2,"ar",0,NULL,0,0,0,0,0},
A
antirez 已提交
253
    {"restore",restoreCommand,-4,"awm",0,NULL,1,1,1,0,0},
254
    {"restore-asking",restoreCommand,-4,"awmk",0,NULL,1,1,1,0,0},
A
antirez 已提交
255
    {"migrate",migrateCommand,-6,"aw",0,NULL,0,0,0,0,0},
256
    {"asking",askingCommand,1,"r",0,NULL,0,0,0,0,0},
257 258
    {"readonly",readonlyCommand,1,"r",0,NULL,0,0,0,0,0},
    {"readwrite",readwriteCommand,1,"r",0,NULL,0,0,0,0,0},
259 260
    {"dump",dumpCommand,2,"ar",0,NULL,1,1,1,0,0},
    {"object",objectCommand,-2,"r",0,NULL,2,2,2,0,0},
261
    {"client",clientCommand,-2,"ar",0,NULL,0,0,0,0,0},
262 263
    {"eval",evalCommand,-3,"s",0,evalGetKeys,0,0,0,0,0},
    {"evalsha",evalShaCommand,-3,"s",0,evalGetKeys,0,0,0,0,0},
264
    {"slowlog",slowlogCommand,-2,"r",0,NULL,0,0,0,0,0},
A
antirez 已提交
265
    {"script",scriptCommand,-2,"ras",0,NULL,0,0,0,0,0},
A
antirez 已提交
266 267
    {"time",timeCommand,1,"rR",0,NULL,0,0,0,0,0},
    {"bitop",bitopCommand,-4,"wm",0,NULL,2,-1,1,0,0},
268
    {"bitcount",bitcountCommand,-2,"r",0,NULL,1,1,1,0,0},
A
antirez 已提交
269
    {"bitpos",bitposCommand,-3,"r",0,NULL,1,1,1,0,0},
270
    {"wait",waitCommand,3,"rs",0,NULL,0,0,0,0,0}
271 272
};

A
antirez 已提交
273 274
struct evictionPoolEntry *evictionPoolAlloc(void);

275 276
/*============================ Utility functions ============================ */

277 278 279
/* 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
280 281
    const int syslogLevelMap[] = { LOG_DEBUG, LOG_INFO, LOG_NOTICE, LOG_WARNING };
    const char *c = ".-*#";
282
    FILE *fp;
283
    char buf[64];
A
antirez 已提交
284
    int rawmode = (level & REDIS_LOG_RAW);
285
    int log_to_stdout = server.logfile[0] == '\0';
286

A
antirez 已提交
287
    level &= 0xff; /* clear flags */
288
    if (level < server.verbosity) return;
289

290
    fp = log_to_stdout ? stdout : fopen(server.logfile,"a");
291 292
    if (!fp) return;

A
antirez 已提交
293 294 295
    if (rawmode) {
        fprintf(fp,"%s",msg);
    } else {
296 297 298 299 300 301
        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
302 303
        fprintf(fp,"[%d] %s %c %s\n",(int)getpid(),buf,c[level],msg);
    }
J
Jonah H. Harris 已提交
304 305
    fflush(fp);

306
    if (!log_to_stdout) fclose(fp);
J
Jonah H. Harris 已提交
307
    if (server.syslog_enabled) syslog(syslogLevelMap[level], "%s", msg);
308 309
}

G
guiquanz 已提交
310
/* Like redisLogRaw() but with printf-alike support. This is the function that
311 312 313 314 315 316
 * 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
317
    if ((level&0xff) < server.verbosity) return;
318 319 320 321 322 323 324 325

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

    redisLogRaw(level,msg);
}

A
antirez 已提交
326 327 328 329 330 331 332 333
/* 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;
334
    int log_to_stdout = server.logfile[0] == '\0';
A
antirez 已提交
335 336
    char buf[64];

337 338 339 340
    if ((level&0xff) < server.verbosity || (log_to_stdout && server.daemonize))
        return;
    fd = log_to_stdout ? STDOUT_FILENO : 
                         open(server.logfile, O_APPEND|O_CREAT|O_WRONLY, 0644);
A
antirez 已提交
341 342
    if (fd == -1) return;
    ll2string(buf,sizeof(buf),getpid());
343 344 345
    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
346
    ll2string(buf,sizeof(buf),time(NULL));
347 348 349 350 351
    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:
352
    if (!log_to_stdout) close(fd);
A
antirez 已提交
353 354
}

A
antirez 已提交
355 356 357 358 359 360 361 362 363 364 365
/* 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;
}

366 367 368 369 370
/* Return the UNIX time in milliseconds */
long long mstime(void) {
    return ustime()/1000;
}

371 372 373 374 375 376 377 378 379 380 381 382
/* 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
}

383 384
/*====================== Hash table type implementation  ==================== */

385
/* This is a hash table type that uses the SDS dynamic strings library as
386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412
 * 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
413 414
/* A case insensitive version used for the command lookup table and other
 * places where case insensitive non binary-safe comparison is needed. */
415 416 417 418 419 420 421 422
int dictSdsKeyCaseCompare(void *privdata, const void *key1,
        const void *key2)
{
    DICT_NOTUSED(privdata);

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

423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453
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));
}

454 455 456 457
unsigned int dictSdsCaseHash(const void *key) {
    return dictGenCaseHashFunction((unsigned char*)key, sdslen((char*)key));
}

458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478
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;

479
    if (sdsEncodedObject(o)) {
480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498
        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;
        }
    }
}

499
/* Sets type hash table */
500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515
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 */
516
    NULL                       /* val destructor */
517 518 519 520 521 522 523 524 525 526 527 528
};

/* 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
529 530 531 532 533 534 535 536 537 538
/* 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 */
};

539 540 541 542 543 544 545 546 547 548
/* Db->expires */
dictType keyptrDictType = {
    dictSdsHash,               /* hash function */
    NULL,                      /* key dup */
    NULL,                      /* val dup */
    dictSdsKeyCompare,         /* key compare */
    NULL,                      /* key destructor */
    NULL                       /* val destructor */
};

549 550 551 552 553 554 555 556 557 558
/* 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 */
};

559
/* Hash type hash table (note that small hashes are represented with ziplists) */
560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580
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
581 582 583 584 585 586 587 588 589 590 591
/* 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 */
};

592 593 594 595 596 597 598 599 600 601 602 603
/* Cluster re-addition blacklist. This maps node IDs to the time
 * we can re-add this node. The goal is to avoid readding a removed
 * node for some time. */
dictType clusterNodesBlackListDictType = {
    dictSdsCaseHash,            /* hash function */
    NULL,                       /* key dup */
    NULL,                       /* val dup */
    dictSdsKeyCaseCompare,      /* key compare */
    dictSdsDestructor,          /* key destructor */
    NULL                        /* val destructor */
};

A
antirez 已提交
604 605 606 607 608 609 610 611 612 613
/* Migrate cache dict type. */
dictType migrateCacheDictType = {
    dictSdsHash,                /* hash function */
    NULL,                       /* key dup */
    NULL,                       /* val dup */
    dictSdsKeyCompare,          /* key compare */
    dictSdsDestructor,          /* key destructor */
    NULL                        /* val destructor */
};

614 615 616 617
/* Replication cached script dict (server.repl_scriptcache_dict).
 * Keys are sds SHA1 strings, while values are not used at all in the current
 * implementation. */
dictType replScriptCacheDictType = {
618
    dictSdsCaseHash,            /* hash function */
619 620
    NULL,                       /* key dup */
    NULL,                       /* val dup */
621
    dictSdsKeyCaseCompare,      /* key compare */
622 623 624 625
    dictSdsDestructor,          /* key destructor */
    NULL                        /* val destructor */
};

626 627 628 629 630 631 632 633 634 635 636
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 */
637 638 639 640 641
void tryResizeHashTables(int dbid) {
    if (htNeedsResize(server.db[dbid].dict))
        dictResize(server.db[dbid].dict);
    if (htNeedsResize(server.db[dbid].expires))
        dictResize(server.db[dbid].expires);
642 643 644 645 646
}

/* 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
647 648 649 650 651 652 653 654 655 656 657 658 659 660
 * of CPU time at every call of this function to perform some rehahsing.
 *
 * The function returns 1 if some rehashing was performed, otherwise 0
 * is returned. */
int incrementallyRehash(int dbid) {
    /* Keys dictionary */
    if (dictIsRehashing(server.db[dbid].dict)) {
        dictRehashMilliseconds(server.db[dbid].dict,1);
        return 1; /* already used our millisecond for this loop... */
    }
    /* Expires */
    if (dictIsRehashing(server.db[dbid].expires)) {
        dictRehashMilliseconds(server.db[dbid].expires,1);
        return 1; /* already used our millisecond for this loop... */
661
    }
662
    return 0;
663 664 665 666 667 668 669 670 671
}

/* 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
672
    if (server.rdb_child_pid == -1 && server.aof_child_pid == -1)
673 674 675 676 677 678 679
        dictEnableResize();
    else
        dictDisableResize();
}

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

680 681 682 683 684 685 686 687 688 689 690
/* Helper function for the activeExpireCycle() function.
 * This function will try to expire the key that is stored in the hash table
 * entry 'de' of the 'expires' hash table of a Redis database.
 *
 * If the key is found to be expired, it is removed from the database and
 * 1 is returned. Otherwise no operation is performed and 0 is returned.
 *
 * When a key is expired, server.stat_expiredkeys is incremented.
 *
 * The parameter 'now' is the current time in milliseconds as is passed
 * to the function to avoid too many gettimeofday() syscalls. */
A
antirez 已提交
691
int activeExpireCycleTryExpire(redisDb *db, dictEntry *de, long long now) {
692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708
    long long t = dictGetSignedIntegerVal(de);
    if (now > t) {
        sds key = dictGetKey(de);
        robj *keyobj = createStringObject(key,sdslen(key));

        propagateExpire(db,keyobj);
        dbDelete(db,keyobj);
        notifyKeyspaceEvent(REDIS_NOTIFY_EXPIRED,
            "expired",keyobj,db->id);
        decrRefCount(keyobj);
        server.stat_expiredkeys++;
        return 1;
    } else {
        return 0;
    }
}

709 710 711
/* 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
712 713
 * keys that can be removed from the keyspace.
 *
714
 * No more than REDIS_DBCRON_DBS_PER_CALL databases are tested at every
715 716
 * iteration.
 *
717 718
 * This kind of call is used when Redis detects that timelimit_exit is
 * true, so there is more work to do, and we do it more incrementally from
A
antirez 已提交
719 720 721 722 723 724 725 726 727 728 729
 * the beforeSleep() function of the event loop.
 *
 * Expire cycle type:
 *
 * If type is ACTIVE_EXPIRE_CYCLE_FAST the function will try to run a
 * "fast" expire cycle that takes no longer than EXPIRE_FAST_CYCLE_DURATION
 * microseconds, and is not repeated again before the same amount of time.
 *
 * If type is ACTIVE_EXPIRE_CYCLE_SLOW, that normal expire cycle is
 * executed, where the time limit is a percentage of the REDIS_HZ period
 * as specified by the REDIS_EXPIRELOOKUPS_TIME_PERC define. */
730

A
antirez 已提交
731
void activeExpireCycle(int type) {
732 733 734 735
    /* This function has some global state in order to continue the work
     * incrementally across calls. */
    static unsigned int current_db = 0; /* Last DB tested. */
    static int timelimit_exit = 0;      /* Time limit hit in previous call? */
A
antirez 已提交
736
    static long long last_fast_cycle = 0; /* When last fast cycle ran. */
737

738
    unsigned int j, iteration = 0;
739
    unsigned int dbs_per_call = REDIS_DBCRON_DBS_PER_CALL;
740
    long long start = ustime(), timelimit;
741

A
antirez 已提交
742
    if (type == ACTIVE_EXPIRE_CYCLE_FAST) {
743 744 745 746
        /* Don't start a fast cycle if the previous cycle did not exited
         * for time limt. Also don't repeat a fast cycle for the same period
         * as the fast cycle total duration itself. */
        if (!timelimit_exit) return;
747
        if (start < last_fast_cycle + ACTIVE_EXPIRE_CYCLE_FAST_DURATION*2) return;
748 749
        last_fast_cycle = start;
    }
750

751 752 753 754 755 756 757 758 759
    /* We usually should test REDIS_DBCRON_DBS_PER_CALL per iteration, with
     * two exceptions:
     *
     * 1) Don't test more DBs than we have.
     * 2) If last time we hit the time limit, we want to scan all DBs
     * in this iteration, as there is work to do in some DB and we don't want
     * expired keys to use memory for too much time. */
    if (dbs_per_call > server.dbnum || timelimit_exit)
        dbs_per_call = server.dbnum;
760

A
antirez 已提交
761
    /* We can use at max ACTIVE_EXPIRE_CYCLE_SLOW_TIME_PERC percentage of CPU time
762
     * per iteration. Since this function gets called with a frequency of
763
     * server.hz times per second, the following is the max amount of
764
     * microseconds we can spend in this function. */
A
antirez 已提交
765
    timelimit = 1000000*ACTIVE_EXPIRE_CYCLE_SLOW_TIME_PERC/server.hz/100;
766
    timelimit_exit = 0;
767
    if (timelimit <= 0) timelimit = 1;
768

A
antirez 已提交
769 770
    if (type == ACTIVE_EXPIRE_CYCLE_FAST)
        timelimit = ACTIVE_EXPIRE_CYCLE_FAST_DURATION; /* in microseconds. */
771

772
    for (j = 0; j < dbs_per_call; j++) {
773
        int expired;
774 775 776 777 778 779
        redisDb *db = server.db+(current_db % server.dbnum);

        /* Increment the DB now so we are sure if we run out of time
         * in the current DB we'll restart from the next. This allows to
         * distribute the time evenly across DBs. */
        current_db++;
780 781 782 783

        /* Continue to expire if at the end of the cycle more than 25%
         * of the keys were expired. */
        do {
784
            unsigned long num, slots;
785 786
            long long now, ttl_sum;
            int ttl_samples;
787 788

            /* If there is nothing to expire try next DB ASAP. */
789 790 791 792
            if ((num = dictSize(db->expires)) == 0) {
                db->avg_ttl = 0;
                break;
            }
793 794
            slots = dictSlots(db->expires);
            now = mstime();
795

796 797 798 799 800 801
            /* 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;

802 803
            /* The main collection cycle. Sample random keys among keys
             * with an expire set, checking for expired ones. */
804
            expired = 0;
805 806 807
            ttl_sum = 0;
            ttl_samples = 0;

A
antirez 已提交
808 809
            if (num > ACTIVE_EXPIRE_CYCLE_LOOKUPS_PER_LOOP)
                num = ACTIVE_EXPIRE_CYCLE_LOOKUPS_PER_LOOP;
810

811 812
            while (num--) {
                dictEntry *de;
813
                long long ttl;
814 815

                if ((de = dictGetRandomKey(db->expires)) == NULL) break;
816
                ttl = dictGetSignedIntegerVal(de)-now;
817
                if (activeExpireCycleTryExpire(db,de,now)) expired++;
818 819 820
                if (ttl < 0) ttl = 0;
                ttl_sum += ttl;
                ttl_samples++;
821
            }
822 823 824 825 826 827 828 829 830 831

            /* Update the average TTL stats for this database. */
            if (ttl_samples) {
                long long avg_ttl = ttl_sum/ttl_samples;

                if (db->avg_ttl == 0) db->avg_ttl = avg_ttl;
                /* Smooth the value averaging with the previous one. */
                db->avg_ttl = (db->avg_ttl+avg_ttl)/2;
            }

832 833 834 835
            /* 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++;
836
            if ((iteration & 0xf) == 0 && /* check once every 16 iterations. */
837 838 839 840
                (ustime()-start) > timelimit)
            {
                timelimit_exit = 1;
            }
841
            if (timelimit_exit) return;
A
antirez 已提交
842 843 844
            /* We don't repeat the cycle if there are less than 25% of keys
             * found expired in the current DB. */
        } while (expired > ACTIVE_EXPIRE_CYCLE_LOOKUPS_PER_LOOP/4);
845 846 847
    }
}

848 849
unsigned int getLRUClock(void) {
    return (mstime()/REDIS_LRU_CLOCK_RESOLUTION) & REDIS_LRU_CLOCK_MAX;
850
}
851

852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875
/* 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
876 877
/* Check for timeouts. Returns non-zero if the client was terminated */
int clientsCronHandleTimeout(redisClient *c) {
A
antirez 已提交
878
    time_t now = server.unixtime;
879 880 881 882 883 884 885 886 887 888 889

    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
890
        return 1;
891
    } else if (c->flags & REDIS_BLOCKED) {
A
antirez 已提交
892 893 894 895 896 897
        /* Blocked OPS timeout is handled with milliseconds resolution.
         * However note that the actual resolution is limited by
         * server.hz. */
        mstime_t now_ms = mstime();

        if (c->bpop.timeout != 0 && c->bpop.timeout < now_ms) {
898 899
            replyToBlockedClientTimedOut(c);
            unblockClient(c);
900 901
        }
    }
A
antirez 已提交
902
    return 0;
903 904
}

905
/* The client query buffer is an sds.c string that can end with a lot of
A
antirez 已提交
906 907
 * free space not used, this function reclaims space if needed.
 *
G
guiquanz 已提交
908
 * The function always returns 0 as it never terminates the client. */
A
antirez 已提交
909
int clientsCronResizeQueryBuffer(redisClient *c) {
910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927
    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
928
    return 0;
929 930
}

931
void clientsCron(void) {
932 933
    /* 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
934 935 936
     * 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. */
937
    int numclients = listLength(server.clients);
938
    int iterations = numclients/(server.hz*10);
939

940 941
    if (iterations < 50)
        iterations = (numclients < 50) ? numclients : 50;
942 943 944 945 946 947 948 949 950 951
    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
952 953 954 955 956
        /* 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;
957 958 959
    }
}

960 961 962 963
/* This function handles 'background' operations we are required to do
 * incrementally in Redis databases, such as active key expiring, resizing,
 * rehashing. */
void databasesCron(void) {
964 965
    /* Expire keys by random sampling. Not required for slaves
     * as master will synthesize DELs for us. */
966
    if (server.active_expire_enabled && server.masterhost == NULL)
A
antirez 已提交
967
        activeExpireCycle(ACTIVE_EXPIRE_CYCLE_SLOW);
968

969 970 971
    /* Perform hash tables rehashing if needed, but only if there are no
     * other processes saving the DB on disk. Otherwise rehashing is bad
     * as will cause a lot of copy-on-write of memory pages. */
972
    if (server.rdb_child_pid == -1 && server.aof_child_pid == -1) {
973 974 975
        /* We use global counters so if we stop the computation at a given
         * DB we'll be able to start from the successive in the next
         * cron loop iteration. */
976 977
        static unsigned int resize_db = 0;
        static unsigned int rehash_db = 0;
978
        unsigned int dbs_per_call = REDIS_DBCRON_DBS_PER_CALL;
979
        unsigned int j;
980

981 982 983
        /* Don't test more DBs than we have. */
        if (dbs_per_call > server.dbnum) dbs_per_call = server.dbnum;

984
        /* Resize */
985
        for (j = 0; j < dbs_per_call; j++) {
986 987 988 989 990 991
            tryResizeHashTables(resize_db % server.dbnum);
            resize_db++;
        }

        /* Rehash */
        if (server.activerehashing) {
992
            for (j = 0; j < dbs_per_call; j++) {
993 994 995 996 997 998 999 1000 1001
                int work_done = incrementallyRehash(rehash_db % server.dbnum);
                rehash_db++;
                if (work_done) {
                    /* If the function did some work, stop here, we'll do
                     * more at the next cron loop. */
                    break;
                }
            }
        }
1002 1003 1004
    }
}

1005 1006 1007 1008 1009 1010 1011 1012 1013
/* 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
 * a lot faster than calling time(NULL) */
void updateCachedTime(void) {
    server.unixtime = time(NULL);
    server.mstime = mstime();
}

1014
/* This is our timer interrupt, called server.hz times per second.
1015 1016 1017 1018 1019
 * 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
1020
 * - Software watchdog.
1021 1022 1023
 * - Update some statistic.
 * - Incremental rehashing of the DBs hash tables.
 * - Triggering BGSAVE / AOF rewrite, and handling of terminated children.
G
guiquanz 已提交
1024
 * - Clients timeout of different kinds.
1025 1026 1027
 * - Replication reconnection.
 * - Many more...
 *
1028
 * Everything directly called here will be called server.hz times per second,
1029 1030 1031 1032
 * so in order to throttle execution of things we want to do less frequently
 * a macro is used: run_with_period(milliseconds) { .... }
 */

1033
int serverCron(struct aeEventLoop *eventLoop, long long id, void *clientData) {
1034
    int j;
1035 1036 1037 1038
    REDIS_NOTUSED(eventLoop);
    REDIS_NOTUSED(id);
    REDIS_NOTUSED(clientData);

A
antirez 已提交
1039 1040 1041 1042
    /* 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);

1043 1044
    /* Update the time cache. */
    updateCachedTime();
A
antirez 已提交
1045

1046
    run_with_period(100) trackOperationsPerSecond();
1047

1048 1049
    /* We have just REDIS_LRU_BITS bits per object for LRU information.
     * So we use an (eventually wrapping) LRU clock.
1050
     *
1051 1052 1053 1054 1055
     * Note that even if the counter wraps it's not a big problem,
     * everything will still work but some object will appear younger
     * to Redis. However for this to happen a given object should never be
     * touched for all the time needed to the counter to wrap, which is
     * not likely.
1056 1057
     *
     * Note that you can change the resolution altering the
1058 1059
     * REDIS_LRU_CLOCK_RESOLUTION define. */
    server.lruclock = getLRUClock();
1060

1061 1062 1063 1064
    /* 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();

A
antirez 已提交
1065 1066 1067
    /* Sample the RSS here since this is a relatively slow call. */
    server.resident_set_size = zmalloc_get_rss();

1068 1069 1070
    /* 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) {
1071
        if (prepareForShutdown(0) == REDIS_OK) exit(0);
1072
        redisLog(REDIS_WARNING,"SIGTERM received but errors trying to shut down the server, check the logs for more information");
1073
        server.shutdown_asap = 0;
1074 1075 1076
    }

    /* Show some info about non-empty databases */
1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087
    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); */
            }
1088 1089 1090 1091
        }
    }

    /* Show information about connected clients */
1092 1093 1094
    if (!server.sentinel_mode) {
        run_with_period(5000) {
            redisLog(REDIS_VERBOSE,
1095
                "%lu clients connected (%lu slaves), %zu bytes in use",
1096 1097 1098 1099
                listLength(server.clients)-listLength(server.slaves),
                listLength(server.slaves),
                zmalloc_used_memory());
        }
1100 1101
    }

1102 1103
    /* We need to do a few operations on clients asynchronously. */
    clientsCron();
1104

1105 1106 1107
    /* Handle background operations on Redis databases. */
    databasesCron();

1108 1109
    /* Start a scheduled AOF rewrite if this was requested by the user while
     * a BGSAVE was in progress. */
A
antirez 已提交
1110
    if (server.rdb_child_pid == -1 && server.aof_child_pid == -1 &&
1111
        server.aof_rewrite_scheduled)
1112 1113 1114 1115
    {
        rewriteAppendOnlyFileBackground();
    }

A
antirez 已提交
1116
    /* Check if a background saving or AOF rewrite in progress terminated. */
A
antirez 已提交
1117
    if (server.rdb_child_pid != -1 || server.aof_child_pid != -1) {
1118 1119 1120 1121
        int statloc;
        pid_t pid;

        if ((pid = wait3(&statloc,WNOHANG,NULL)) != 0) {
1122 1123 1124 1125 1126
            int exitcode = WEXITSTATUS(statloc);
            int bysignal = 0;
            
            if (WIFSIGNALED(statloc)) bysignal = WTERMSIG(statloc);

A
antirez 已提交
1127
            if (pid == server.rdb_child_pid) {
1128
                backgroundSaveDoneHandler(exitcode,bysignal);
1129
            } else if (pid == server.aof_child_pid) {
1130
                backgroundRewriteDoneHandler(exitcode,bysignal);
1131 1132 1133 1134
            } else {
                redisLog(REDIS_WARNING,
                    "Warning, detected child with unmatched pid: %ld",
                    (long)pid);
1135 1136 1137
            }
            updateDictResizePolicy();
        }
A
antirez 已提交
1138
    } else {
1139 1140
        /* If there is not a background saving/rewrite in progress check if
         * we have to save/rewrite now */
1141 1142 1143
         for (j = 0; j < server.saveparamslen; j++) {
            struct saveparam *sp = server.saveparams+j;

1144 1145 1146 1147
            /* Save if we reached the given amount of changes,
             * the given amount of seconds, and if the latest bgsave was
             * successful or if, in case of an error, at least
             * REDIS_BGSAVE_RETRY_DELAY seconds already elapsed. */
1148
            if (server.dirty >= sp->changes &&
1149 1150 1151 1152 1153
                server.unixtime-server.lastsave > sp->seconds &&
                (server.unixtime-server.lastbgsave_try >
                 REDIS_BGSAVE_RETRY_DELAY ||
                 server.lastbgsave_status == REDIS_OK))
            {
1154
                redisLog(REDIS_NOTICE,"%d changes in %d seconds. Saving...",
1155
                    sp->changes, (int)sp->seconds);
A
antirez 已提交
1156
                rdbSaveBackground(server.rdb_filename);
1157 1158 1159
                break;
            }
         }
1160 1161

         /* Trigger an AOF rewrite if needed */
A
antirez 已提交
1162
         if (server.rdb_child_pid == -1 &&
A
antirez 已提交
1163
             server.aof_child_pid == -1 &&
1164 1165
             server.aof_rewrite_perc &&
             server.aof_current_size > server.aof_rewrite_min_size)
1166
         {
1167 1168 1169 1170
            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) {
1171
                redisLog(REDIS_NOTICE,"Starting automatic rewriting of AOF on %lld%% growth",growth);
1172 1173 1174
                rewriteAppendOnlyFileBackground();
            }
         }
1175 1176
    }

1177

1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188
    /* AOF postponed flush: Try at every cron cycle if the slow fsync
     * completed. */
    if (server.aof_flush_postponed_start) flushAppendOnlyFile(0);

    /* AOF write errors: in this case we have a buffer to flush as well and
     * clear the AOF error in case of success to make the DB writable again,
     * however to try every second is enough in case of 'hz' is set to
     * an higher frequency. */
    run_with_period(1000) {
        if (server.aof_last_write_status == REDIS_ERR)
            flushAppendOnlyFile(0);
1189
    }
1190

1191 1192 1193
    /* Close clients that need to be closed asynchronous */
    freeClientsInAsyncFreeQueue();

1194 1195 1196
    /* Clear the paused clients flag if needed. */
    clientsArePaused(); /* Don't check return value, just use the side effect. */

1197 1198
    /* Replication cron function -- used to reconnect to master and
     * to detect transfer failures. */
1199
    run_with_period(1000) replicationCron();
1200

A
antirez 已提交
1201
    /* Run the Redis Cluster cron. */
1202
    run_with_period(100) {
1203 1204
        if (server.cluster_enabled) clusterCron();
    }
A
antirez 已提交
1205

A
antirez 已提交
1206
    /* Run the Sentinel timer if we are in sentinel mode. */
1207 1208 1209 1210
    run_with_period(100) {
        if (server.sentinel_mode) sentinelTimer();
    }

A
antirez 已提交
1211 1212 1213 1214 1215
    /* Cleanup expired MIGRATE cached sockets. */
    run_with_period(1000) {
        migrateCloseTimedoutSockets();
    }

1216
    server.cronloops++;
1217
    return 1000/server.hz;
1218 1219 1220 1221 1222 1223 1224 1225
}

/* 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);

1226 1227 1228 1229
    /* Run a fast expire cycle (the called function will return
     * ASAP if a fast cycle is not needed). */
    if (server.active_expire_enabled && server.masterhost == NULL)
        activeExpireCycle(ACTIVE_EXPIRE_CYCLE_FAST);
1230

1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250
    /* Send all the slaves an ACK request if at least one client blocked
     * during the previous event loop iteration. */
    if (server.get_ack_from_slaves) {
        robj *argv[3];

        argv[0] = createStringObject("REPLCONF",8);
        argv[1] = createStringObject("GETACK",6);
        argv[2] = createStringObject("*",1); /* Not used argument. */
        replicationFeedSlaves(server.slaves, server.slaveseldb, argv, 3);
        decrRefCount(argv[0]);
        decrRefCount(argv[1]);
        decrRefCount(argv[2]);
        server.get_ack_from_slaves = 0;
    }

    /* Unblock all the clients blocked for synchronous replication
     * in WAIT. */
    if (listLength(server.clients_waiting_acks))
        processClientsWaitingReplicas();

1251
    /* Try to process pending commands for clients that were just unblocked. */
1252 1253
    if (listLength(server.unblocked_clients))
        processUnblockedClients();
1254

1255
    /* Write the AOF buffer on disk */
1256
    flushAppendOnlyFile(0);
1257 1258 1259

    /* Call the Redis Cluster before sleep function. */
    if (server.cluster_enabled) clusterBeforeSleep();
1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278
}

/* =========================== 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"));
A
antirez 已提交
1279
    shared.emptyscan = createObject(REDIS_STRING,sdsnew("*2\r\n$1\r\n0\r\n*0\r\n"));
1280
    shared.wrongtypeerr = createObject(REDIS_STRING,sdsnew(
1281
        "-WRONGTYPE Operation against a key holding the wrong kind of value\r\n"));
1282 1283 1284 1285 1286 1287 1288 1289
    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
1290 1291
    shared.noscripterr = createObject(REDIS_STRING,sdsnew(
        "-NOSCRIPT No matching script. Please use EVAL.\r\n"));
1292 1293
    shared.loadingerr = createObject(REDIS_STRING,sdsnew(
        "-LOADING Redis is loading the dataset in memory\r\n"));
1294
    shared.slowscripterr = createObject(REDIS_STRING,sdsnew(
1295
        "-BUSY Redis is busy running a script. You can only call SCRIPT KILL or SHUTDOWN NOSAVE.\r\n"));
1296 1297
    shared.masterdownerr = createObject(REDIS_STRING,sdsnew(
        "-MASTERDOWN Link with MASTER is down and slave-serve-stale-data is set to 'no'.\r\n"));
1298
    shared.bgsaveerr = createObject(REDIS_STRING,sdsnew(
1299 1300 1301
        "-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"));
1302 1303
    shared.noautherr = createObject(REDIS_STRING,sdsnew(
        "-NOAUTH Authentication required.\r\n"));
1304 1305
    shared.oomerr = createObject(REDIS_STRING,sdsnew(
        "-OOM command not allowed when used memory > 'maxmemory'.\r\n"));
A
antirez 已提交
1306 1307
    shared.execaborterr = createObject(REDIS_STRING,sdsnew(
        "-EXECABORT Transaction discarded because of previous errors.\r\n"));
1308 1309
    shared.noreplicaserr = createObject(REDIS_STRING,sdsnew(
        "-NOREPLICAS Not enough good slaves to write.\r\n"));
1310 1311 1312
    shared.space = createObject(REDIS_STRING,sdsnew(" "));
    shared.colon = createObject(REDIS_STRING,sdsnew(":"));
    shared.plus = createObject(REDIS_STRING,sdsnew("+"));
1313 1314

    for (j = 0; j < REDIS_SHARED_SELECT_CMDS; j++) {
1315 1316 1317 1318
        char dictid_str[64];
        int dictid_len;

        dictid_len = ll2string(dictid_str,sizeof(dictid_str),j);
1319
        shared.select[j] = createObject(REDIS_STRING,
1320 1321 1322
            sdscatprintf(sdsempty(),
                "*2\r\n$6\r\nSELECT\r\n$%d\r\n%s\r\n",
                dictid_len, dictid_str));
1323
    }
1324 1325 1326 1327 1328 1329
    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("$12\r\npunsubscribe\r\n",19);
1330
    shared.del = createStringObject("DEL",3);
1331 1332
    shared.rpop = createStringObject("RPOP",4);
    shared.lpop = createStringObject("LPOP",4);
1333
    shared.lpush = createStringObject("LPUSH",5);
1334 1335 1336 1337
    for (j = 0; j < REDIS_SHARED_INTEGERS; j++) {
        shared.integers[j] = createObject(REDIS_STRING,(void*)(long)j);
        shared.integers[j]->encoding = REDIS_ENCODING_INT;
    }
1338 1339 1340 1341 1342 1343
    for (j = 0; j < REDIS_SHARED_BULKHDR_LEN; j++) {
        shared.mbulkhdr[j] = createObject(REDIS_STRING,
            sdscatprintf(sdsempty(),"*%d\r\n",j));
        shared.bulkhdr[j] = createObject(REDIS_STRING,
            sdscatprintf(sdsempty(),"$%d\r\n",j));
    }
1344 1345 1346
}

void initServerConfig() {
1347 1348
    int j;

A
antirez 已提交
1349
    getRandomHexChars(server.runid,REDIS_RUN_ID_SIZE);
1350
    server.configfile = NULL;
1351
    server.hz = REDIS_DEFAULT_HZ;
A
antirez 已提交
1352
    server.runid[REDIS_RUN_ID_SIZE] = '\0';
1353
    server.arch_bits = (sizeof(long) == 8) ? 64 : 32;
1354
    server.port = REDIS_SERVERPORT;
1355
    server.tcp_backlog = REDIS_TCP_BACKLOG;
A
antirez 已提交
1356
    server.bindaddr_count = 0;
1357
    server.unixsocket = NULL;
1358
    server.unixsocketperm = REDIS_DEFAULT_UNIX_SOCKET_PERM;
1359
    server.ipfd_count = 0;
1360 1361
    server.sofd = -1;
    server.dbnum = REDIS_DEFAULT_DBNUM;
1362
    server.verbosity = REDIS_DEFAULT_VERBOSITY;
1363
    server.maxidletime = REDIS_MAXIDLETIME;
1364
    server.tcpkeepalive = REDIS_DEFAULT_TCP_KEEPALIVE;
A
antirez 已提交
1365
    server.active_expire_enabled = 1;
1366
    server.client_max_querybuf_len = REDIS_MAX_QUERYBUF_LEN;
1367
    server.saveparams = NULL;
1368
    server.loading = 0;
1369 1370
    server.logfile = zstrdup(REDIS_DEFAULT_LOGFILE);
    server.syslog_enabled = REDIS_DEFAULT_SYSLOG_ENABLED;
1371
    server.syslog_ident = zstrdup(REDIS_DEFAULT_SYSLOG_IDENT);
J
Jonah H. Harris 已提交
1372
    server.syslog_facility = LOG_LOCAL0;
1373
    server.daemonize = REDIS_DEFAULT_DAEMONIZE;
1374
    server.aof_state = REDIS_AOF_OFF;
1375 1376
    server.aof_fsync = REDIS_DEFAULT_AOF_FSYNC;
    server.aof_no_fsync_on_rewrite = REDIS_DEFAULT_AOF_NO_FSYNC_ON_REWRITE;
1377 1378 1379 1380
    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
1381
    server.aof_last_fsync = time(NULL);
1382 1383
    server.aof_rewrite_time_last = -1;
    server.aof_rewrite_time_start = -1;
1384
    server.aof_lastbgrewrite_status = REDIS_OK;
1385
    server.aof_delayed_fsync = 0;
A
antirez 已提交
1386 1387
    server.aof_fd = -1;
    server.aof_selected_db = -1; /* Make sure the first time will not match */
1388
    server.aof_flush_postponed_start = 0;
1389
    server.aof_rewrite_incremental_fsync = REDIS_DEFAULT_AOF_REWRITE_INCREMENTAL_FSYNC;
1390
    server.pidfile = zstrdup(REDIS_DEFAULT_PID_FILE);
1391
    server.rdb_filename = zstrdup(REDIS_DEFAULT_RDB_FILENAME);
1392
    server.aof_filename = zstrdup(REDIS_DEFAULT_AOF_FILENAME);
1393
    server.requirepass = NULL;
1394 1395 1396 1397
    server.rdb_compression = REDIS_DEFAULT_RDB_COMPRESSION;
    server.rdb_checksum = REDIS_DEFAULT_RDB_CHECKSUM;
    server.stop_writes_on_bgsave_err = REDIS_DEFAULT_STOP_WRITES_ON_BGSAVE_ERROR;
    server.activerehashing = REDIS_DEFAULT_ACTIVE_REHASHING;
A
antirez 已提交
1398
    server.notify_keyspace_events = 0;
1399
    server.maxclients = REDIS_MAX_CLIENTS;
1400
    server.bpop_blocked_clients = 0;
1401 1402 1403
    server.maxmemory = REDIS_DEFAULT_MAXMEMORY;
    server.maxmemory_policy = REDIS_DEFAULT_MAXMEMORY_POLICY;
    server.maxmemory_samples = REDIS_DEFAULT_MAXMEMORY_SAMPLES;
1404 1405
    server.hash_max_ziplist_entries = REDIS_HASH_MAX_ZIPLIST_ENTRIES;
    server.hash_max_ziplist_value = REDIS_HASH_MAX_ZIPLIST_VALUE;
1406 1407
    server.list_max_ziplist_entries = REDIS_LIST_MAX_ZIPLIST_ENTRIES;
    server.list_max_ziplist_value = REDIS_LIST_MAX_ZIPLIST_VALUE;
1408
    server.set_max_intset_entries = REDIS_SET_MAX_INTSET_ENTRIES;
1409 1410
    server.zset_max_ziplist_entries = REDIS_ZSET_MAX_ZIPLIST_ENTRIES;
    server.zset_max_ziplist_value = REDIS_ZSET_MAX_ZIPLIST_VALUE;
1411
    server.shutdown_asap = 0;
A
1fe4cd5  
antirez 已提交
1412 1413
    server.repl_ping_slave_period = REDIS_REPL_PING_SLAVE_PERIOD;
    server.repl_timeout = REDIS_REPL_TIMEOUT;
1414 1415
    server.repl_min_slaves_to_write = REDIS_DEFAULT_MIN_SLAVES_TO_WRITE;
    server.repl_min_slaves_max_lag = REDIS_DEFAULT_MIN_SLAVES_MAX_LAG;
A
antirez 已提交
1416
    server.cluster_enabled = 0;
1417
    server.cluster_node_timeout = REDIS_CLUSTER_DEFAULT_NODE_TIMEOUT;
1418
    server.cluster_migration_barrier = REDIS_CLUSTER_DEFAULT_MIGRATION_BARRIER;
1419
    server.cluster_configfile = zstrdup(REDIS_DEFAULT_CLUSTER_CONFIG_FILE);
1420
    server.lua_caller = NULL;
A
antirez 已提交
1421
    server.lua_time_limit = REDIS_LUA_TIME_LIMIT;
1422
    server.lua_client = NULL;
1423
    server.lua_timedout = 0;
A
antirez 已提交
1424
    server.migrate_cached_sockets = dictCreate(&migrateCacheDictType,NULL);
1425
    server.loading_process_events_interval_bytes = (1024*1024*2);
1426

1427
    server.lruclock = getLRUClock();
1428 1429 1430 1431 1432 1433 1434 1435 1436 1437
    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;
1438 1439
    server.cached_master = NULL;
    server.repl_master_initial_offset = -1;
A
antirez 已提交
1440
    server.repl_state = REDIS_REPL_NONE;
1441
    server.repl_syncio_timeout = REDIS_REPL_SYNCIO_TIMEOUT;
1442 1443
    server.repl_serve_stale_data = REDIS_DEFAULT_SLAVE_SERVE_STALE_DATA;
    server.repl_slave_ro = REDIS_DEFAULT_SLAVE_READ_ONLY;
1444
    server.repl_down_since = 0; /* Never connected, repl is down since EVER. */
1445
    server.repl_disable_tcp_nodelay = REDIS_DEFAULT_REPL_DISABLE_TCP_NODELAY;
1446
    server.slave_priority = REDIS_DEFAULT_SLAVE_PRIORITY;
1447 1448 1449 1450 1451 1452 1453 1454 1455 1456
    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);
1457

1458
    /* Client output buffer limits */
1459 1460
    for (j = 0; j < REDIS_CLIENT_LIMIT_NUM_CLASSES; j++)
        server.client_obuf_limits[j] = clientBufferLimitsDefaults[j];
1461

1462 1463 1464 1465 1466
    /* 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;
1467

G
guiquanz 已提交
1468
    /* Command table -- we initiialize it here as it is part of the
1469 1470 1471
     * initial configuration, since command names may be changed via
     * redis.conf using the rename-command directive. */
    server.commands = dictCreate(&commandTableDictType,NULL);
1472
    server.orig_commands = dictCreate(&commandTableDictType,NULL);
1473 1474 1475
    populateCommandTable();
    server.delCommand = lookupCommandByCString("del");
    server.multiCommand = lookupCommandByCString("multi");
1476
    server.lpushCommand = lookupCommandByCString("lpush");
1477 1478
    server.lpopCommand = lookupCommandByCString("lpop");
    server.rpopCommand = lookupCommandByCString("rpop");
1479 1480 1481 1482
    
    /* 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
1483

A
antirez 已提交
1484
    /* Debugging */
A
antirez 已提交
1485 1486 1487 1488
    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
1489
    server.watchdog_period = 0;
1490 1491
}

1492
/* This function will try to raise the max number of open files accordingly to
1493
 * the configured max number of clients. It also reserves a number of file
1494
 * descriptors (REDIS_MIN_RESERVED_FDS) for extra operations of
1495
 * persistence, listening sockets, log files and so forth.
1496 1497 1498 1499 1500
 *
 * 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) {
1501
    rlim_t maxfiles = server.maxclients+REDIS_MIN_RESERVED_FDS;
1502 1503 1504 1505 1506
    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));
1507
        server.maxclients = 1024-REDIS_MIN_RESERVED_FDS;
1508 1509 1510 1511 1512 1513
    } 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) {
1514
            rlim_t f;
A
antirez 已提交
1515 1516 1517 1518
            int setrlimit_error = 0;

            /* Try to set the file limit to match 'maxfiles' or at least
             * to the higher value supported less than maxfiles. */
1519 1520
            f = maxfiles;
            while(f > oldlimit) {
A
antirez 已提交
1521 1522
                int decr_step = 16;

1523 1524 1525
                limit.rlim_cur = f;
                limit.rlim_max = f;
                if (setrlimit(RLIMIT_NOFILE,&limit) != -1) break;
A
antirez 已提交
1526 1527 1528 1529 1530 1531
                setrlimit_error = errno;

                /* We failed to set file limit to 'f'. Try with a
                 * smaller limit decrementing by a few FDs per iteration. */
                if (f < decr_step) break;
                f -= decr_step;
1532
            }
A
antirez 已提交
1533 1534 1535

            /* Assume that the limit we get initially is still valid if
             * our last try was even lower. */
1536
            if (f < oldlimit) f = oldlimit;
A
antirez 已提交
1537

1538
            if (f != maxfiles) {
1539
                int old_maxclients = server.maxclients;
1540
                server.maxclients = f-REDIS_MIN_RESERVED_FDS;
1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552
                if (server.maxclients < 1) {
                    redisLog(REDIS_WARNING,"Your current 'ulimit -n' "
                        "of %llu is not enough for Redis to start. "
                        "Please increase your open file limit to at least "
                        "%llu. Exiting.", oldlimit, maxfiles);
                    exit(1);
                }
                redisLog(REDIS_WARNING,"You requested maxclients of %d "
                    "requiring at least %llu max file descriptors.",
                    old_maxclients, maxfiles);
                redisLog(REDIS_WARNING,"Redis can't set maximum open files "
                    "to %llu because of OS error: %s.",
A
antirez 已提交
1553
                    maxfiles, strerror(setrlimit_error));
1554 1555 1556 1557 1558
                redisLog(REDIS_WARNING,"Current maximum open files is %llu. "
                    "maxclients has been reduced to %d to compensate for "
                    "low ulimit. "
                    "If you need higher maxclients increase 'ulimit -n'.",
                    oldlimit, server.maxclients);
1559
            } else {
1560 1561 1562
                redisLog(REDIS_NOTICE,"Increased maximum number of open files "
                    "to %llu (it was originally set to %llu).",
                    maxfiles, oldlimit);
1563 1564 1565 1566 1567
            }
        }
    }
}

1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595
/* Initialize a set of file descriptors to listen to the specified 'port'
 * binding the addresses specified in the Redis server configuration.
 *
 * The listening file descriptors are stored in the integer array 'fds'
 * and their number is set in '*count'.
 *
 * The addresses to bind are specified in the global server.bindaddr array
 * and their number is server.bindaddr_count. If the server configuration
 * contains no specific addresses to bind, this function will try to
 * bind * (all addresses) for both the IPv4 and IPv6 protocols.
 *
 * On success the function returns REDIS_OK.
 *
 * On error the function returns REDIS_ERR. For the function to be on
 * error, at least one of the server.bindaddr addresses was
 * impossible to bind, or no bind addresses were specified in the server
 * configuration but the function is not able to bind * for at least
 * one of the IPv4 or IPv6 protocols. */
int listenToPort(int port, int *fds, int *count) {
    int j;

    /* Force binding of 0.0.0.0 if no bind address is specified, always
     * entering the loop if j == 0. */
    if (server.bindaddr_count == 0) server.bindaddr[0] = NULL;
    for (j = 0; j < server.bindaddr_count || j == 0; j++) {
        if (server.bindaddr[j] == NULL) {
            /* Bind * for both IPv6 and IPv4, we enter here only if
             * server.bindaddr_count == 0. */
1596 1597
            fds[*count] = anetTcp6Server(server.neterr,port,NULL,
                server.tcp_backlog);
1598
            if (fds[*count] != ANET_ERR) (*count)++;
1599 1600
            fds[*count] = anetTcpServer(server.neterr,port,NULL,
                server.tcp_backlog);
1601 1602 1603 1604 1605 1606 1607
            if (fds[*count] != ANET_ERR) (*count)++;
            /* Exit the loop if we were able to bind * on IPv4 or IPv6,
             * otherwise fds[*count] will be ANET_ERR and we'll print an
             * error and return to the caller with an error. */
            if (*count) break;
        } else if (strchr(server.bindaddr[j],':')) {
            /* Bind IPv6 address. */
1608 1609
            fds[*count] = anetTcp6Server(server.neterr,port,server.bindaddr[j],
                server.tcp_backlog);
1610 1611
        } else {
            /* Bind IPv4 address. */
1612 1613
            fds[*count] = anetTcpServer(server.neterr,port,server.bindaddr[j],
                server.tcp_backlog);
1614 1615 1616 1617 1618
        }
        if (fds[*count] == ANET_ERR) {
            redisLog(REDIS_WARNING,
                "Creating Server TCP listening socket %s:%d: %s",
                server.bindaddr[j] ? server.bindaddr[j] : "*",
1619
                port, server.neterr);
1620 1621 1622 1623 1624 1625 1626
            return REDIS_ERR;
        }
        (*count)++;
    }
    return REDIS_OK;
}

1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647
/* Resets the stats that we expose via INFO or other means that we want
 * to reset via CONFIG RESETSTAT. The function is also used in order to
 * initialize these fields in initServer() at server startup. */
void resetServerStats(void) {
    server.stat_numcommands = 0;
    server.stat_numconnections = 0;
    server.stat_expiredkeys = 0;
    server.stat_evictedkeys = 0;
    server.stat_keyspace_misses = 0;
    server.stat_keyspace_hits = 0;
    server.stat_fork_time = 0;
    server.stat_rejected_conn = 0;
    server.stat_sync_full = 0;
    server.stat_sync_partial_ok = 0;
    server.stat_sync_partial_err = 0;
    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;
}

1648 1649 1650 1651 1652
void initServer() {
    int j;

    signal(SIGHUP, SIG_IGN);
    signal(SIGPIPE, SIG_IGN);
1653
    setupSignalHandlers();
1654

J
Jonah H. Harris 已提交
1655 1656 1657 1658 1659
    if (server.syslog_enabled) {
        openlog(server.syslog_ident, LOG_PID | LOG_NDELAY | LOG_NOWAIT,
            server.syslog_facility);
    }

1660
    server.current_client = NULL;
1661
    server.clients = listCreate();
1662
    server.clients_to_close = listCreate();
1663 1664
    server.slaves = listCreate();
    server.monitors = listCreate();
1665
    server.slaveseldb = -1; /* Force to emit the first SELECT command. */
1666
    server.unblocked_clients = listCreate();
1667
    server.ready_keys = listCreate();
1668 1669
    server.clients_waiting_acks = listCreate();
    server.get_ack_from_slaves = 0;
1670
    server.clients_paused = 0;
1671

1672
    createSharedObjects();
1673
    adjustOpenFilesLimit();
A
antirez 已提交
1674
    server.el = aeCreateEventLoop(server.maxclients+REDIS_EVENTLOOP_FDSET_INCR);
1675
    server.db = zmalloc(sizeof(redisDb)*server.dbnum);
1676

1677
    /* Open the TCP listening socket for the user commands. */
1678 1679
    if (server.port != 0 &&
        listenToPort(server.port,server.ipfd,&server.ipfd_count) == REDIS_ERR)
1680
        exit(1);
1681 1682

    /* Open the listening Unix domain socket. */
1683 1684
    if (server.unixsocket != NULL) {
        unlink(server.unixsocket); /* don't care if this fails */
1685 1686
        server.sofd = anetUnixServer(server.neterr,server.unixsocket,
            server.unixsocketperm, server.tcp_backlog);
1687 1688 1689 1690
        if (server.sofd == ANET_ERR) {
            redisLog(REDIS_WARNING, "Opening socket: %s", server.neterr);
            exit(1);
        }
1691
    }
1692 1693 1694

    /* Abort if there are no listening sockets at all. */
    if (server.ipfd_count == 0 && server.sofd < 0) {
1695
        redisLog(REDIS_WARNING, "Configured to not listen anywhere, exiting.");
1696 1697
        exit(1);
    }
1698 1699

    /* Create the Redis databases, and initialize other internal state. */
1700 1701 1702 1703
    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);
1704
        server.db[j].ready_keys = dictCreate(&setDictType,NULL);
1705
        server.db[j].watched_keys = dictCreate(&keylistDictType,NULL);
A
antirez 已提交
1706
        server.db[j].eviction_pool = evictionPoolAlloc();
1707
        server.db[j].id = j;
1708
        server.db[j].avg_ttl = 0;
1709 1710 1711 1712 1713 1714
    }
    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
1715
    server.rdb_child_pid = -1;
A
antirez 已提交
1716
    server.aof_child_pid = -1;
1717
    aofRewriteBufferReset();
A
antirez 已提交
1718
    server.aof_buf = sdsempty();
1719 1720
    server.lastsave = time(NULL); /* At startup we consider the DB saved. */
    server.lastbgsave_try = 0;    /* At startup we never tried to BGSAVE. */
1721 1722
    server.rdb_save_time_last = -1;
    server.rdb_save_time_start = -1;
1723
    server.dirty = 0;
1724 1725
    resetServerStats();
    /* A few stats we don't want to reset: server startup time, and peak mem. */
1726
    server.stat_starttime = time(NULL);
1727
    server.stat_peak_memory = 0;
A
antirez 已提交
1728
    server.resident_set_size = 0;
1729
    server.lastbgsave_status = REDIS_OK;
1730 1731
    server.aof_last_write_status = REDIS_OK;
    server.aof_last_write_errno = 0;
1732
    server.repl_good_slaves_count = 0;
1733
    updateCachedTime();
1734 1735 1736

    /* Create the serverCron() time event, that's our main way to process
     * background operations. */
stamhe's avatar
stamhe 已提交
1737
    if(aeCreateTimeEvent(server.el, 1, serverCron, NULL, NULL) == AE_ERR) {
1738
        redisPanic("Can't create the serverCron time event.");
stamhe's avatar
stamhe 已提交
1739 1740
        exit(1);
    }
1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751

    /* Create an event handler for accepting new connections in TCP and Unix
     * domain sockets. */
    for (j = 0; j < server.ipfd_count; j++) {
        if (aeCreateFileEvent(server.el, server.ipfd[j], AE_READABLE,
            acceptTcpHandler,NULL) == AE_ERR)
            {
                redisPanic(
                    "Unrecoverable error creating server.ipfd file event.");
            }
    }
1752
    if (server.sofd > 0 && aeCreateFileEvent(server.el,server.sofd,AE_READABLE,
A
antirez 已提交
1753
        acceptUnixHandler,NULL) == AE_ERR) redisPanic("Unrecoverable error creating server.sofd file event.");
1754

1755
    /* Open the AOF file if needed. */
1756
    if (server.aof_state == REDIS_AOF_ON) {
A
antirez 已提交
1757
        server.aof_fd = open(server.aof_filename,
1758
                               O_WRONLY|O_APPEND|O_CREAT,0644);
A
antirez 已提交
1759
        if (server.aof_fd == -1) {
1760 1761 1762 1763 1764 1765
            redisLog(REDIS_WARNING, "Can't open the append-only file: %s",
                strerror(errno));
            exit(1);
        }
    }

1766 1767
    /* 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
1768 1769
     * at 3 GB using maxmemory with 'noeviction' policy'. This avoids
     * useless crashes of the Redis instance for out of memory. */
1770
    if (server.arch_bits == 32 && server.maxmemory == 0) {
1771 1772
        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 */
1773 1774 1775
        server.maxmemory_policy = REDIS_MAXMEMORY_NO_EVICTION;
    }

A
antirez 已提交
1776
    if (server.cluster_enabled) clusterInit();
1777
    replicationScriptCacheInit();
1778
    scriptingInit();
1779
    slowlogInit();
1780
    bioInit();
1781 1782
}

1783 1784 1785 1786
/* Populates the Redis Command Table starting from the hard coded list
 * we have on top of redis.c file. */
void populateCommandTable(void) {
    int j;
1787
    int numcommands = sizeof(redisCommandTable)/sizeof(struct redisCommand);
1788 1789

    for (j = 0; j < numcommands; j++) {
1790
        struct redisCommand *c = redisCommandTable+j;
1791
        char *f = c->sflags;
1792
        int retval1, retval2;
1793

1794 1795 1796 1797 1798 1799 1800
        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;
1801 1802
            case 's': c->flags |= REDIS_CMD_NOSCRIPT; break;
            case 'R': c->flags |= REDIS_CMD_RANDOM; break;
1803
            case 'S': c->flags |= REDIS_CMD_SORT_FOR_SCRIPT; break;
1804 1805
            case 'l': c->flags |= REDIS_CMD_LOADING; break;
            case 't': c->flags |= REDIS_CMD_STALE; break;
1806
            case 'M': c->flags |= REDIS_CMD_SKIP_MONITOR; break;
1807
            case 'k': c->flags |= REDIS_CMD_ASKING; break;
1808 1809 1810 1811 1812
            default: redisPanic("Unsupported command flag"); break;
            }
            f++;
        }

1813 1814 1815 1816 1817
        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);
1818
    }
1819 1820
}

1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832
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;
    }
}

1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869
/* ========================== 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);
}

1870 1871
/* ====================== Commands lookup and execution ===================== */

1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882
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;
1883 1884
}

1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898
/* 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
1899
/* Propagate the specified command (in the context of the specified database id)
1900
 * to AOF and Slaves.
A
antirez 已提交
1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911
 *
 * 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);
1912
    if (flags & REDIS_PROPAGATE_REPL)
A
antirez 已提交
1913 1914 1915
        replicationFeedSlaves(server.slaves,dbid,argv,argc);
}

1916 1917
/* Used inside commands to schedule the propagation of additional commands
 * after the current command is propagated to AOF / Replication. */
1918 1919 1920
void alsoPropagate(struct redisCommand *cmd, int dbid, robj **argv, int argc,
                   int target)
{
1921
    redisOpArrayAppend(&server.also_propagate,cmd,dbid,argv,argc,target);
1922 1923
}

A
antirez 已提交
1924 1925 1926 1927 1928 1929 1930 1931
/* It is possible to call the function forceCommandPropagation() inside a
 * Redis command implementaiton in order to to force the propagation of a
 * specific command execution into AOF / Replication. */
void forceCommandPropagation(redisClient *c, int flags) {
    if (flags & REDIS_PROPAGATE_REPL) c->flags |= REDIS_FORCE_REPL;
    if (flags & REDIS_PROPAGATE_AOF) c->flags |= REDIS_FORCE_AOF;
}

1932
/* Call() is the core of Redis execution of a command */
1933
void call(redisClient *c, int flags) {
1934
    long long dirty, start, duration;
A
antirez 已提交
1935
    int client_old_flags = c->flags;
1936

A
antirez 已提交
1937
    /* Sent the command to clients in MONITOR mode, only if the commands are
G
guiquanz 已提交
1938
     * not generated from reading an AOF. */
1939 1940 1941 1942
    if (listLength(server.monitors) &&
        !server.loading &&
        !(c->cmd->flags & REDIS_CMD_SKIP_MONITOR))
    {
1943
        replicationFeedMonitors(c,server.monitors,c->db->id,c->argv,c->argc);
1944
    }
A
antirez 已提交
1945 1946

    /* Call the command. */
A
antirez 已提交
1947
    c->flags &= ~(REDIS_FORCE_AOF|REDIS_FORCE_REPL);
1948
    redisOpArrayInit(&server.also_propagate);
1949
    dirty = server.dirty;
1950
    start = ustime();
1951
    c->cmd->proc(c);
1952
    duration = ustime()-start;
1953
    dirty = server.dirty-dirty;
1954 1955 1956 1957 1958 1959

    /* 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
1960 1961 1962 1963 1964 1965 1966 1967 1968 1969
    /* If the caller is Lua, we want to force the EVAL caller to propagate
     * the script if the command flag or client flag are forcing the
     * propagation. */
    if (c->flags & REDIS_LUA_CLIENT && server.lua_caller) {
        if (c->flags & REDIS_FORCE_REPL)
            server.lua_caller->flags |= REDIS_FORCE_REPL;
        if (c->flags & REDIS_FORCE_AOF)
            server.lua_caller->flags |= REDIS_FORCE_AOF;
    }

A
antirez 已提交
1970 1971
    /* Log the command into the Slow log if needed, and populate the
     * per-command statistics that we show in INFO commandstats. */
1972
    if (flags & REDIS_CALL_SLOWLOG && c->cmd->proc != execCommand)
1973 1974 1975 1976 1977
        slowlogPushEntryIfNeeded(c->argv,c->argc,duration);
    if (flags & REDIS_CALL_STATS) {
        c->cmd->microseconds += duration;
        c->cmd->calls++;
    }
A
antirez 已提交
1978 1979

    /* Propagate the command into the AOF and replication link */
1980
    if (flags & REDIS_CALL_PROPAGATE) {
A
antirez 已提交
1981 1982
        int flags = REDIS_PROPAGATE_NONE;

A
antirez 已提交
1983 1984
        if (c->flags & REDIS_FORCE_REPL) flags |= REDIS_PROPAGATE_REPL;
        if (c->flags & REDIS_FORCE_AOF) flags |= REDIS_PROPAGATE_AOF;
A
antirez 已提交
1985 1986 1987 1988
        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);
1989
    }
1990

A
antirez 已提交
1991 1992 1993 1994 1995
    /* Restore the old FORCE_AOF/REPL flags, since call can be executed
     * recursively. */
    c->flags &= ~(REDIS_FORCE_AOF|REDIS_FORCE_REPL);
    c->flags |= client_old_flags & (REDIS_FORCE_AOF|REDIS_FORCE_REPL);

1996 1997
    /* Handle the alsoPropagate() API to handle commands that want to propagate
     * multiple separated commands. */
1998
    if (server.also_propagate.numops) {
1999
        int j;
2000
        redisOp *rop;
2001

2002 2003 2004 2005 2006
        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);
2007
    }
2008 2009 2010 2011
    server.stat_numcommands++;
}

/* If this function gets called we already read a whole
2012
 * command, arguments are in the client argv/argc fields.
2013 2014 2015 2016
 * 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
2017 2018
 * other operations can be performed by the caller. Otherwise
 * if 0 is returned the client was destroyed (i.e. after QUIT). */
2019
int processCommand(redisClient *c) {
P
Pieter Noordhuis 已提交
2020 2021 2022 2023
    /* 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. */
2024
    if (!strcasecmp(c->argv[0]->ptr,"quit")) {
P
Pieter Noordhuis 已提交
2025
        addReply(c,shared.ok);
2026
        c->flags |= REDIS_CLOSE_AFTER_REPLY;
2027
        return REDIS_ERR;
2028 2029 2030
    }

    /* Now lookup the command and check ASAP about trivial error conditions
2031
     * such as wrong arity, bad command name and so forth. */
2032
    c->cmd = c->lastcmd = lookupCommand(c->argv[0]->ptr);
2033
    if (!c->cmd) {
A
antirez 已提交
2034
        flagTransaction(c);
2035 2036
        addReplyErrorFormat(c,"unknown command '%s'",
            (char*)c->argv[0]->ptr);
2037
        return REDIS_OK;
2038 2039
    } else if ((c->cmd->arity > 0 && c->cmd->arity != c->argc) ||
               (c->argc < -c->cmd->arity)) {
A
antirez 已提交
2040
        flagTransaction(c);
2041
        addReplyErrorFormat(c,"wrong number of arguments for '%s' command",
2042
            c->cmd->name);
2043
        return REDIS_OK;
2044 2045 2046
    }

    /* Check if the user is authenticated */
2047 2048
    if (server.requirepass && !c->authenticated && c->cmd->proc != authCommand)
    {
A
antirez 已提交
2049
        flagTransaction(c);
2050
        addReply(c,shared.noautherr);
2051
        return REDIS_OK;
2052 2053
    }

2054 2055 2056 2057
    /* 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
2058
    if (server.cluster_enabled &&
2059 2060 2061
        !(c->flags & REDIS_MASTER) &&
        !(c->cmd->getkeys_proc == NULL && c->cmd->firstkey == 0))
    {
A
antirez 已提交
2062 2063
        int hashslot;

2064
        if (server.cluster->state != REDIS_CLUSTER_OK) {
2065
            flagTransaction(c);
2066
            addReplySds(c,sdsnew("-CLUSTERDOWN The cluster is down. Use CLUSTER INFO for more information\r\n"));
A
antirez 已提交
2067 2068
            return REDIS_OK;
        } else {
2069 2070
            int error_code;
            clusterNode *n = getNodeByQuery(c,c->cmd,c->argv,c->argc,&hashslot,&error_code);
A
antirez 已提交
2071
            if (n == NULL) {
2072
                flagTransaction(c);
2073 2074 2075 2076 2077 2078 2079 2080 2081 2082
                if (error_code == REDIS_CLUSTER_REDIR_CROSS_SLOT) {
                    addReplySds(c,sdsnew("-CROSSSLOT Keys in request don't hash to the same slot\r\n"));
                } else if (error_code == REDIS_CLUSTER_REDIR_UNSTABLE) {
                    /* The request spawns mutliple keys in the same slot,
                     * but the slot is not "stable" currently as there is
                     * a migration or import in progress. */
                    addReplySds(c,sdsnew("-TRYAGAIN Multiple keys request during rehashing of slot\r\n"));
                } else {
                    redisPanic("getNodeByQuery() unknown error.");
                }
A
antirez 已提交
2083
                return REDIS_OK;
2084
            } else if (n != server.cluster->myself) {
2085
                flagTransaction(c);
A
antirez 已提交
2086
                addReplySds(c,sdscatprintf(sdsempty(),
2087 2088
                    "-%s %d %s:%d\r\n",
                    (error_code == REDIS_CLUSTER_REDIR_ASK) ? "ASK" : "MOVED",
2089
                    hashslot,n->ip,n->port));
A
antirez 已提交
2090 2091 2092 2093 2094
                return REDIS_OK;
            }
        }
    }

2095 2096 2097 2098 2099
    /* 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. */
2100 2101 2102
    if (server.maxmemory) {
        int retval = freeMemoryIfNeeded();
        if ((c->cmd->flags & REDIS_CMD_DENYOOM) && retval == REDIS_ERR) {
A
antirez 已提交
2103
            flagTransaction(c);
2104
            addReply(c, shared.oomerr);
2105 2106
            return REDIS_OK;
        }
2107 2108
    }

2109 2110
    /* Don't accept write commands if there are problems persisting on disk
     * and if this is a master instance. */
2111 2112 2113 2114
    if (((server.stop_writes_on_bgsave_err &&
          server.saveparamslen > 0 &&
          server.lastbgsave_status == REDIS_ERR) ||
          server.aof_last_write_status == REDIS_ERR) &&
2115
        server.masterhost == NULL &&
2116 2117
        (c->cmd->flags & REDIS_CMD_WRITE ||
         c->cmd->proc == pingCommand))
2118
    {
A
antirez 已提交
2119
        flagTransaction(c);
2120 2121 2122 2123 2124 2125 2126
        if (server.aof_last_write_status == REDIS_OK)
            addReply(c, shared.bgsaveerr);
        else
            addReplySds(c,
                sdscatprintf(sdsempty(),
                "-MISCONF Errors writing to the AOF file: %s\r\n",
                strerror(server.aof_last_write_errno)));
2127 2128 2129
        return REDIS_OK;
    }

2130
    /* Don't accept write commands if there are not enough good slaves and
2131
     * user configured the min-slaves-to-write option. */
2132 2133 2134
    if (server.repl_min_slaves_to_write &&
        server.repl_min_slaves_max_lag &&
        c->cmd->flags & REDIS_CMD_WRITE &&
2135 2136 2137 2138 2139 2140 2141
        server.repl_good_slaves_count < server.repl_min_slaves_to_write)
    {
        flagTransaction(c);
        addReply(c, shared.noreplicaserr);
        return REDIS_OK;
    }

2142
    /* Don't accept write commands if this is a read only slave. But
2143 2144 2145 2146 2147 2148 2149 2150 2151
     * 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;
    }

2152 2153 2154
    /* Only allow SUBSCRIBE and UNSUBSCRIBE in the context of Pub/Sub */
    if ((dictSize(c->pubsub_channels) > 0 || listLength(c->pubsub_patterns) > 0)
        &&
2155 2156 2157 2158
        c->cmd->proc != subscribeCommand &&
        c->cmd->proc != unsubscribeCommand &&
        c->cmd->proc != psubscribeCommand &&
        c->cmd->proc != punsubscribeCommand) {
2159
        addReplyError(c,"only (P)SUBSCRIBE / (P)UNSUBSCRIBE / QUIT allowed in this context");
2160
        return REDIS_OK;
2161 2162
    }

2163 2164
    /* 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
2165
    if (server.masterhost && server.repl_state != REDIS_REPL_CONNECTED &&
2166
        server.repl_serve_stale_data == 0 &&
2167
        !(c->cmd->flags & REDIS_CMD_STALE))
2168
    {
A
antirez 已提交
2169
        flagTransaction(c);
2170
        addReply(c, shared.masterdownerr);
2171 2172 2173
        return REDIS_OK;
    }

2174 2175 2176
    /* Loading DB? Return an error if the command has not the
     * REDIS_CMD_LOADING flag. */
    if (server.loading && !(c->cmd->flags & REDIS_CMD_LOADING)) {
2177 2178 2179 2180
        addReply(c, shared.loadingerr);
        return REDIS_OK;
    }

A
antirez 已提交
2181
    /* Lua script too slow? Only allow a limited number of commands. */
2182
    if (server.lua_timedout &&
2183
          c->cmd->proc != authCommand &&
A
antirez 已提交
2184
          c->cmd->proc != replconfCommand &&
2185
        !(c->cmd->proc == shutdownCommand &&
2186 2187 2188 2189 2190 2191
          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
2192
        flagTransaction(c);
2193 2194 2195 2196
        addReply(c, shared.slowscripterr);
        return REDIS_OK;
    }

2197 2198
    /* Exec the command */
    if (c->flags & REDIS_MULTI &&
2199 2200
        c->cmd->proc != execCommand && c->cmd->proc != discardCommand &&
        c->cmd->proc != multiCommand && c->cmd->proc != watchCommand)
2201
    {
2202
        queueMultiCommand(c);
2203 2204
        addReply(c,shared.queued);
    } else {
2205
        call(c,REDIS_CALL_FULL);
2206
        c->woff = server.master_repl_offset;
2207 2208
        if (listLength(server.ready_keys))
            handleClientsBlockedOnLists();
2209
    }
2210
    return REDIS_OK;
2211 2212 2213 2214
}

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

2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229
/* Close listening sockets. Also unlink the unix domain socket if
 * unlink_unix_socket is non-zero. */
void closeListeningSockets(int unlink_unix_socket) {
    int j;

    for (j = 0; j < server.ipfd_count; j++) close(server.ipfd[j]);
    if (server.sofd != -1) close(server.sofd);
    if (server.cluster_enabled)
        for (j = 0; j < server.cfd_count; j++) close(server.cfd[j]);
    if (unlink_unix_socket && server.unixsocket) {
        redisLog(REDIS_NOTICE,"Removing the unix socket file.");
        unlink(server.unixsocket); /* don't care if this fails */
    }
}

2230 2231 2232 2233
int prepareForShutdown(int flags) {
    int save = flags & REDIS_SHUTDOWN_SAVE;
    int nosave = flags & REDIS_SHUTDOWN_NOSAVE;

2234
    redisLog(REDIS_WARNING,"User requested shutdown...");
2235 2236 2237
    /* 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
2238
    if (server.rdb_child_pid != -1) {
2239
        redisLog(REDIS_WARNING,"There is a child saving an .rdb. Killing it!");
2240
        kill(server.rdb_child_pid,SIGUSR1);
A
antirez 已提交
2241
        rdbRemoveTempFile(server.rdb_child_pid);
2242
    }
2243
    if (server.aof_state != REDIS_AOF_OFF) {
2244 2245
        /* 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
2246
        if (server.aof_child_pid != -1) {
2247 2248
            redisLog(REDIS_WARNING,
                "There is a child rewriting the AOF. Killing it!");
2249
            kill(server.aof_child_pid,SIGUSR1);
2250
        }
2251
        /* Append only file: fsync() the AOF and exit */
2252
        redisLog(REDIS_NOTICE,"Calling fsync() on the AOF file.");
A
antirez 已提交
2253
        aof_fsync(server.aof_fd);
2254
    }
2255
    if ((server.saveparamslen > 0 && !nosave) || save) {
2256
        redisLog(REDIS_NOTICE,"Saving the final RDB snapshot before exiting.");
2257
        /* Snapshotting. Perform a SYNC SAVE and exit */
A
antirez 已提交
2258
        if (rdbSave(server.rdb_filename) != REDIS_OK) {
2259 2260 2261 2262 2263
            /* 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... */
2264
            redisLog(REDIS_WARNING,"Error trying to save the DB, can't exit.");
2265 2266 2267
            return REDIS_ERR;
        }
    }
2268 2269 2270 2271
    if (server.daemonize) {
        redisLog(REDIS_NOTICE,"Removing the pid file.");
        unlink(server.pidfile);
    }
2272
    /* Close the listening sockets. Apparently this allows faster restarts. */
2273
    closeListeningSockets(1);
2274 2275
    redisLog(REDIS_WARNING,"%s is now ready to exit, bye bye...",
        server.sentinel_mode ? "Sentinel" : "Redis");
2276 2277 2278 2279 2280
    return REDIS_OK;
}

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

2281
/* Return zero if strings are the same, non-zero if they are not.
2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322
 * 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. */
}

2323
void authCommand(redisClient *c) {
2324 2325
    if (!server.requirepass) {
        addReplyError(c,"Client sent AUTH, but no password is set");
2326
    } else if (!time_independent_strcmp(c->argv[1]->ptr, server.requirepass)) {
2327 2328 2329 2330
      c->authenticated = 1;
      addReply(c,shared.ok);
    } else {
      c->authenticated = 0;
2331
      addReplyError(c,"invalid password");
2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342
    }
}

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

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

A
antirez 已提交
2343 2344 2345
void timeCommand(redisClient *c) {
    struct timeval tv;

G
guiquanz 已提交
2346
    /* gettimeofday() can only fail if &tv is a bad address so we
A
antirez 已提交
2347 2348 2349 2350 2351 2352 2353
     * don't check for errors. */
    gettimeofday(&tv,NULL);
    addReplyMultiBulkLen(c,2);
    addReplyBulkLongLong(c,tv.tv_sec);
    addReplyBulkLongLong(c,tv.tv_usec);
}

2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377
/* 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. */
2378 2379
sds genRedisInfoString(char *section) {
    sds info = sdsempty();
2380
    time_t uptime = server.unixtime-server.stat_starttime;
A
antirez 已提交
2381
    int j, numcommands;
2382
    struct rusage self_ru, c_ru;
2383
    unsigned long lol, bib;
2384 2385
    int allsections = 0, defsections = 0;
    int sections = 0;
2386

2387 2388
    if (section) {
        allsections = strcasecmp(section,"all") == 0;
2389
        defsections = strcasecmp(section,"default") == 0;
2390
    }
2391 2392 2393

    getrusage(RUSAGE_SELF, &self_ru);
    getrusage(RUSAGE_CHILDREN, &c_ru);
2394
    getClientsMaxBuffers(&lol,&bib);
2395 2396 2397

    /* Server */
    if (allsections || defsections || !strcasecmp(section,"server")) {
2398 2399
        static int call_uname = 1;
        static struct utsname name;
2400
        char *mode;
2401

2402 2403 2404 2405
        if (server.cluster_enabled) mode = "cluster";
        else if (server.sentinel_mode) mode = "sentinel";
        else mode = "standalone";
    
2406
        if (sections++) info = sdscat(info,"\r\n");
2407 2408 2409 2410 2411 2412 2413

        if (call_uname) {
            /* Uname can be slow and is always the same output. Cache it. */
            uname(&name);
            call_uname = 0;
        }

2414
        info = sdscatprintf(info,
2415 2416 2417 2418
            "# Server\r\n"
            "redis_version:%s\r\n"
            "redis_git_sha1:%s\r\n"
            "redis_git_dirty:%d\r\n"
2419
            "redis_build_id:%llx\r\n"
2420
            "redis_mode:%s\r\n"
2421
            "os:%s %s %s\r\n"
2422
            "arch_bits:%d\r\n"
2423
            "multiplexing_api:%s\r\n"
A
antirez 已提交
2424
            "gcc_version:%d.%d.%d\r\n"
2425
            "process_id:%ld\r\n"
A
antirez 已提交
2426
            "run_id:%s\r\n"
2427
            "tcp_port:%d\r\n"
Y
YAMAMOTO Takashi 已提交
2428 2429
            "uptime_in_seconds:%jd\r\n"
            "uptime_in_days:%jd\r\n"
2430
            "hz:%d\r\n"
2431 2432
            "lru_clock:%ld\r\n"
            "config_file:%s\r\n",
2433 2434 2435
            REDIS_VERSION,
            redisGitSHA1(),
            strtol(redisGitDirty(),NULL,10) > 0,
2436
            (unsigned long long) redisBuildId(),
2437
            mode,
2438
            name.sysname, name.release, name.machine,
2439
            server.arch_bits,
2440
            aeGetApiName(),
A
antirez 已提交
2441 2442 2443 2444 2445
#ifdef __GNUC__
            __GNUC__,__GNUC_MINOR__,__GNUC_PATCHLEVEL__,
#else
            0,0,0,
#endif
2446
            (long) getpid(),
A
antirez 已提交
2447
            server.runid,
2448
            server.port,
Y
YAMAMOTO Takashi 已提交
2449 2450
            (intmax_t)uptime,
            (intmax_t)(uptime/(3600*24)),
2451
            server.hz,
2452 2453
            (unsigned long) server.lruclock,
            server.configfile ? server.configfile : "");
2454 2455 2456 2457 2458 2459 2460
    }

    /* Clients */
    if (allsections || defsections || !strcasecmp(section,"clients")) {
        if (sections++) info = sdscat(info,"\r\n");
        info = sdscatprintf(info,
            "# Clients\r\n"
2461
            "connected_clients:%lu\r\n"
2462 2463 2464 2465 2466 2467 2468 2469 2470 2471
            "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")) {
2472 2473
        char hmem[64];
        char peak_hmem[64];
2474
        size_t zmalloc_used = zmalloc_used_memory();
2475

2476 2477 2478 2479 2480
        /* Peak memory is updated from time to time by serverCron() so it
         * may happen that the instantaneous value is slightly bigger than
         * the peak value. This may confuse users, so we update the peak
         * if found smaller than the current memory usage. */
        if (zmalloc_used > server.stat_peak_memory)
2481 2482 2483
            server.stat_peak_memory = zmalloc_used;

        bytesToHuman(hmem,zmalloc_used);
2484
        bytesToHuman(peak_hmem,server.stat_peak_memory);
2485 2486 2487 2488 2489 2490
        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"
2491 2492
            "used_memory_peak:%zu\r\n"
            "used_memory_peak_human:%s\r\n"
A
antirez 已提交
2493
            "used_memory_lua:%lld\r\n"
2494
            "mem_fragmentation_ratio:%.2f\r\n"
2495
            "mem_allocator:%s\r\n",
2496
            zmalloc_used,
2497
            hmem,
A
antirez 已提交
2498
            server.resident_set_size,
2499 2500
            server.stat_peak_memory,
            peak_hmem,
A
antirez 已提交
2501
            ((long long)lua_gc(server.lua,LUA_GCCOUNT,0))*1024LL,
A
antirez 已提交
2502
            zmalloc_get_fragmentation_ratio(server.resident_set_size),
2503
            ZMALLOC_LIB
2504
            );
2505 2506
    }

2507 2508 2509
    /* Persistence */
    if (allsections || defsections || !strcasecmp(section,"persistence")) {
        if (sections++) info = sdscat(info,"\r\n");
2510
        info = sdscatprintf(info,
2511 2512
            "# Persistence\r\n"
            "loading:%d\r\n"
2513 2514
            "rdb_changes_since_last_save:%lld\r\n"
            "rdb_bgsave_in_progress:%d\r\n"
Y
YAMAMOTO Takashi 已提交
2515
            "rdb_last_save_time:%jd\r\n"
2516
            "rdb_last_bgsave_status:%s\r\n"
Y
YAMAMOTO Takashi 已提交
2517 2518
            "rdb_last_bgsave_time_sec:%jd\r\n"
            "rdb_current_bgsave_time_sec:%jd\r\n"
2519
            "aof_enabled:%d\r\n"
2520 2521
            "aof_rewrite_in_progress:%d\r\n"
            "aof_rewrite_scheduled:%d\r\n"
Y
YAMAMOTO Takashi 已提交
2522 2523
            "aof_last_rewrite_time_sec:%jd\r\n"
            "aof_current_rewrite_time_sec:%jd\r\n"
2524 2525
            "aof_last_bgrewrite_status:%s\r\n"
            "aof_last_write_status:%s\r\n",
2526 2527
            server.loading,
            server.dirty,
A
antirez 已提交
2528
            server.rdb_child_pid != -1,
Y
YAMAMOTO Takashi 已提交
2529
            (intmax_t)server.lastsave,
2530
            (server.lastbgsave_status == REDIS_OK) ? "ok" : "err",
Y
YAMAMOTO Takashi 已提交
2531 2532 2533
            (intmax_t)server.rdb_save_time_last,
            (intmax_t)((server.rdb_child_pid == -1) ?
                -1 : time(NULL)-server.rdb_save_time_start),
2534
            server.aof_state != REDIS_AOF_OFF,
2535
            server.aof_child_pid != -1,
2536
            server.aof_rewrite_scheduled,
Y
YAMAMOTO Takashi 已提交
2537 2538 2539
            (intmax_t)server.aof_rewrite_time_last,
            (intmax_t)((server.aof_child_pid == -1) ?
                -1 : time(NULL)-server.aof_rewrite_time_start),
2540 2541
            (server.aof_lastbgrewrite_status == REDIS_OK) ? "ok" : "err",
            (server.aof_last_write_status == REDIS_OK) ? "ok" : "err");
2542

2543
        if (server.aof_state != REDIS_AOF_OFF) {
2544 2545 2546
            info = sdscatprintf(info,
                "aof_current_size:%lld\r\n"
                "aof_base_size:%lld\r\n"
2547 2548
                "aof_pending_rewrite:%d\r\n"
                "aof_buffer_length:%zu\r\n"
2549
                "aof_rewrite_buffer_length:%lu\r\n"
2550 2551
                "aof_pending_bio_fsync:%llu\r\n"
                "aof_delayed_fsync:%lu\r\n",
2552 2553 2554
                (long long) server.aof_current_size,
                (long long) server.aof_rewrite_base_size,
                server.aof_rewrite_scheduled,
A
antirez 已提交
2555
                sdslen(server.aof_buf),
2556
                aofRewriteBufferSize(),
2557 2558
                bioPendingJobsOfType(REDIS_BIO_AOF_FSYNC),
                server.aof_delayed_fsync);
2559 2560
        }

2561 2562 2563 2564 2565 2566 2567 2568 2569
        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;

2570
            elapsed = server.unixtime-server.loading_start_time;
2571 2572 2573 2574 2575 2576 2577 2578
            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,
Y
YAMAMOTO Takashi 已提交
2579
                "loading_start_time:%jd\r\n"
2580 2581 2582
                "loading_total_bytes:%llu\r\n"
                "loading_loaded_bytes:%llu\r\n"
                "loading_loaded_perc:%.2f\r\n"
Y
YAMAMOTO Takashi 已提交
2583 2584
                "loading_eta_seconds:%jd\r\n",
                (intmax_t) server.loading_start_time,
2585 2586 2587
                (unsigned long long) server.loading_total_bytes,
                (unsigned long long) server.loading_loaded_bytes,
                perc,
Y
YAMAMOTO Takashi 已提交
2588
                (intmax_t)eta
2589 2590
            );
        }
2591
    }
2592 2593 2594 2595

    /* Stats */
    if (allsections || defsections || !strcasecmp(section,"stats")) {
        if (sections++) info = sdscat(info,"\r\n");
2596
        info = sdscatprintf(info,
2597 2598 2599
            "# Stats\r\n"
            "total_connections_received:%lld\r\n"
            "total_commands_processed:%lld\r\n"
2600
            "instantaneous_ops_per_sec:%lld\r\n"
2601
            "rejected_connections:%lld\r\n"
2602 2603 2604
            "sync_full:%lld\r\n"
            "sync_partial_ok:%lld\r\n"
            "sync_partial_err:%lld\r\n"
2605 2606 2607 2608 2609
            "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"
2610
            "pubsub_patterns:%lu\r\n"
2611 2612
            "latest_fork_usec:%lld\r\n"
            "migrate_cached_sockets:%ld\r\n",
2613 2614
            server.stat_numconnections,
            server.stat_numcommands,
2615
            getOperationsPerSecond(),
2616
            server.stat_rejected_conn,
2617 2618 2619
            server.stat_sync_full,
            server.stat_sync_partial_ok,
            server.stat_sync_partial_err,
2620 2621 2622 2623 2624
            server.stat_expiredkeys,
            server.stat_evictedkeys,
            server.stat_keyspace_hits,
            server.stat_keyspace_misses,
            dictSize(server.pubsub_channels),
2625
            listLength(server.pubsub_patterns),
2626 2627
            server.stat_fork_time,
            dictSize(server.migrate_cached_sockets));
2628
    }
A
antirez 已提交
2629

2630 2631 2632 2633 2634 2635 2636 2637
    /* 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) {
2638 2639 2640 2641 2642 2643 2644
            long long slave_repl_offset = 1;

            if (server.master)
                slave_repl_offset = server.master->reploff;
            else if (server.cached_master)
                slave_repl_offset = server.cached_master->reploff;

2645 2646 2647 2648 2649 2650
            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"
2651
                "slave_repl_offset:%lld\r\n"
2652 2653
                ,server.masterhost,
                server.masterport,
A
antirez 已提交
2654
                (server.repl_state == REDIS_REPL_CONNECTED) ?
2655 2656
                    "up" : "down",
                server.master ?
2657
                ((int)(server.unixtime-server.master->lastinteraction)) : -1,
2658
                server.repl_state == REDIS_REPL_TRANSFER,
2659
                slave_repl_offset
2660 2661
            );

A
antirez 已提交
2662
            if (server.repl_state == REDIS_REPL_TRANSFER) {
2663
                info = sdscatprintf(info,
2664
                    "master_sync_left_bytes:%lld\r\n"
2665
                    "master_sync_last_io_seconds_ago:%d\r\n"
2666 2667
                    , (long long)
                        (server.repl_transfer_size - server.repl_transfer_read),
2668
                    (int)(server.unixtime-server.repl_transfer_lastio)
2669 2670
                );
            }
2671

A
antirez 已提交
2672
            if (server.repl_state != REDIS_REPL_CONNECTED) {
2673
                info = sdscatprintf(info,
Y
YAMAMOTO Takashi 已提交
2674 2675
                    "master_link_down_since_seconds:%jd\r\n",
                    (intmax_t)server.unixtime-server.repl_down_since);
2676
            }
2677
            info = sdscatprintf(info,
2678 2679 2680 2681
                "slave_priority:%d\r\n"
                "slave_read_only:%d\r\n",
                server.slave_priority,
                server.repl_slave_ro);
A
antirez 已提交
2682
        }
2683

2684
        info = sdscatprintf(info,
2685
            "connected_slaves:%lu\r\n",
2686
            listLength(server.slaves));
2687 2688 2689 2690 2691 2692 2693 2694 2695 2696

        /* If min-slaves-to-write is active, write the number of slaves
         * currently considered 'good'. */
        if (server.repl_min_slaves_to_write &&
            server.repl_min_slaves_max_lag) {
            info = sdscatprintf(info,
                "min_slaves_good_slaves:%d\r\n",
                server.repl_good_slaves_count);
        }

2697 2698 2699 2700 2701 2702 2703 2704 2705
        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;
2706
                char ip[REDIS_IP_STR_LEN];
2707
                int port;
2708
                long lag = 0;
2709

2710
                if (anetPeerToString(slave->fd,ip,sizeof(ip),&port) == -1) continue;
2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723
                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;
2724 2725 2726 2727 2728
                if (slave->replstate == REDIS_REPL_ONLINE)
                    lag = time(NULL) - slave->repl_ack_time;

                info = sdscatprintf(info,
                    "slave%d:ip=%s,port=%d,state=%s,"
2729
                    "offset=%lld,lag=%ld\r\n",
2730
                    slaveid,ip,slave->slave_listening_port,state,
2731
                    slave->repl_ack_off, lag);
2732 2733 2734
                slaveid++;
            }
        }
2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745
        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
2746 2747
    }

2748 2749
    /* CPU */
    if (allsections || defsections || !strcasecmp(section,"cpu")) {
2750 2751
        if (sections++) info = sdscat(info,"\r\n");
        info = sdscatprintf(info,
2752
        "# CPU\r\n"
2753 2754
        "used_cpu_sys:%.2f\r\n"
        "used_cpu_user:%.2f\r\n"
2755 2756
        "used_cpu_sys_children:%.2f\r\n"
        "used_cpu_user_children:%.2f\r\n",
2757
        (float)self_ru.ru_stime.tv_sec+(float)self_ru.ru_stime.tv_usec/1000000,
2758 2759 2760
        (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);
2761
    }
2762

2763 2764 2765 2766
    /* cmdtime */
    if (allsections || !strcasecmp(section,"commandstats")) {
        if (sections++) info = sdscat(info,"\r\n");
        info = sdscatprintf(info, "# Commandstats\r\n");
2767
        numcommands = sizeof(redisCommandTable)/sizeof(struct redisCommand);
2768
        for (j = 0; j < numcommands; j++) {
2769
            struct redisCommand *c = redisCommandTable+j;
2770

2771 2772 2773 2774 2775
            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));
2776
        }
A
antirez 已提交
2777 2778
    }

2779
    /* Cluster */
2780 2781 2782 2783 2784 2785 2786 2787
    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);
    }

2788 2789 2790 2791 2792 2793
    /* 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;
2794

2795 2796 2797
            keys = dictSize(server.db[j].dict);
            vkeys = dictSize(server.db[j].expires);
            if (keys || vkeys) {
2798 2799 2800
                info = sdscatprintf(info,
                    "db%d:keys=%lld,expires=%lld,avg_ttl=%lld\r\n",
                    j, keys, vkeys, server.db[j].avg_ttl);
2801
            }
2802 2803 2804 2805 2806 2807
        }
    }
    return info;
}

void infoCommand(redisClient *c) {
2808 2809 2810 2811 2812 2813 2814
    char *section = c->argc == 2 ? c->argv[1]->ptr : "default";

    if (c->argc > 2) {
        addReply(c,shared.syntaxerr);
        return;
    }
    sds info = genRedisInfoString(section);
2815 2816 2817 2818 2819 2820 2821
    addReplySds(c,sdscatprintf(sdsempty(),"$%lu\r\n",
        (unsigned long)sdslen(info)));
    addReplySds(c,info);
    addReply(c,shared.crlf);
}

void monitorCommand(redisClient *c) {
2822
    /* ignore MONITOR if already slave or in monitor mode */
2823 2824 2825 2826 2827 2828 2829 2830 2831
    if (c->flags & REDIS_SLAVE) return;

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

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

A
antirez 已提交
2832 2833 2834
/* freeMemoryIfNeeded() gets called when 'maxmemory' is set on the config
 * file to limit the max memory used by the server, before processing a
 * command.
2835
 *
2836 2837
 * The goal of the function is to free enough memory to keep Redis under the
 * configured memory limit.
2838
 *
2839 2840 2841 2842 2843 2844 2845 2846
 * 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.
A
antirez 已提交
2847 2848 2849 2850 2851 2852
 *
 * ------------------------------------------------------------------------
 *
 * LRU approximation algorithm
 *
 * Redis uses an approximation of the LRU algorithm that runs in constant
J
Jan-Erik Rediger 已提交
2853
 * memory. Every time there is a key to expire, we sample N keys (with
A
antirez 已提交
2854 2855 2856 2857
 * N very small, usually in around 5) to populate a pool of best keys to
 * evict of M keys (the pool size is defined by REDIS_EVICTION_POOL_SIZE).
 *
 * The N keys sampled are added in the pool of good keys to expire (the one
J
Jan-Erik Rediger 已提交
2858
 * with an old access time) if they are better than one of the current keys
A
antirez 已提交
2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890
 * in the pool.
 *
 * After the pool is populated, the best key we have in the pool is expired.
 * However note that we don't remove keys from the pool when they are deleted
 * so the pool may contain keys that no longer exist.
 *
 * When we try to evict a key, and all the entries in the pool don't exist
 * we populate it again. This time we'll be sure that the pool has at least
 * one key that can be evicted, if there is at least one key that can be
 * evicted in the whole database. */

/* Create a new eviction pool. */
struct evictionPoolEntry *evictionPoolAlloc(void) {
    struct evictionPoolEntry *ep;
    int j;

    ep = zmalloc(sizeof(*ep)*REDIS_EVICTION_POOL_SIZE);
    for (j = 0; j < REDIS_EVICTION_POOL_SIZE; j++) {
        ep[j].idle = 0;
        ep[j].key = NULL;
    }
    return ep;
}

/* This is an helper function for freeMemoryIfNeeded(), it is used in order
 * to populate the evictionPool with a few entries every time we want to
 * expire a key. Keys with idle time smaller than one of the current
 * keys are added. Keys are always added if there are free entries.
 *
 * We insert keys on place in ascending order, so keys with the smaller
 * idle time are on the left, and keys with the higher idle time on the
 * right. */
2891 2892

#define EVICTION_SAMPLES_ARRAY_SIZE 16
A
antirez 已提交
2893
void evictionPoolPopulate(dict *sampledict, dict *keydict, struct evictionPoolEntry *pool) {
2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911
    int j, k, count;
    dictEntry *_samples[EVICTION_SAMPLES_ARRAY_SIZE];
    dictEntry **samples;

    /* Try to use a static buffer: this function is a big hit...
     * Note: it was actually measured that this helps. */
    if (server.maxmemory_samples <= EVICTION_SAMPLES_ARRAY_SIZE) {
        samples = _samples;
    } else {
        samples = zmalloc(sizeof(samples[0])*server.maxmemory_samples);
    }

#if 1 /* Use bulk get by default. */
    count = dictGetRandomKeys(sampledict,samples,server.maxmemory_samples);
#else
    count = server.maxmemory_samples;
    for (j = 0; j < count; j++) samples[j] = dictGetRandomKey(sampledict);
#endif
A
antirez 已提交
2912

2913
    for (j = 0; j < count; j++) {
A
antirez 已提交
2914 2915 2916
        unsigned long long idle;
        sds key;
        robj *o;
A
antirez 已提交
2917
        dictEntry *de;
A
antirez 已提交
2918

2919
        de = samples[j];
A
antirez 已提交
2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935
        key = dictGetKey(de);
        /* If the dictionary we are sampling from is not the main
         * dictionary (but the expires one) we need to lookup the key
         * again in the key dictionary to obtain the value object. */
        if (sampledict != keydict) de = dictFind(keydict, key);
        o = dictGetVal(de);
        idle = estimateObjectIdleTime(o);

        /* Insert the element inside the pool.
         * First, find the first empty bucket or the first populated
         * bucket that has an idle time smaller than our idle time. */
        k = 0;
        while (k < REDIS_EVICTION_POOL_SIZE &&
               pool[k].key &&
               pool[k].idle < idle) k++;
        if (k == 0 && pool[REDIS_EVICTION_POOL_SIZE-1].key != NULL) {
J
Jan-Erik Rediger 已提交
2936
            /* Can't insert if the element is < the worst element we have
A
antirez 已提交
2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960
             * and there are no empty buckets. */
            continue;
        } else if (k < REDIS_EVICTION_POOL_SIZE && pool[k].key == NULL) {
            /* Inserting into empty position. No setup needed before insert. */
        } else {
            /* Inserting in the middle. Now k points to the first element
             * greater than the element to insert.  */
            if (pool[REDIS_EVICTION_POOL_SIZE-1].key == NULL) {
                /* Free space on the right? Insert at k shifting
                 * all the elements from k to end to the right. */
                memmove(pool+k+1,pool+k,
                    sizeof(pool[0])*(REDIS_EVICTION_POOL_SIZE-k-1));
            } else {
                /* No free space on right? Insert at k-1 */
                k--;
                /* Shift all elements on the left of k (included) to the
                 * left, so we discard the element with smaller idle time. */
                sdsfree(pool[0].key);
                memmove(pool,pool+1,sizeof(pool[0])*k);
            }
        }
        pool[k].key = sdsdup(key);
        pool[k].idle = idle;
    }
2961
    if (samples != _samples) zfree(samples);
A
antirez 已提交
2962 2963
}

2964 2965 2966 2967
int freeMemoryIfNeeded(void) {
    size_t mem_used, mem_tofree, mem_freed;
    int slaves = listLength(server.slaves);

2968 2969
    /* Remove the size of slaves output buffers and AOF buffer from the
     * count of used memory. */
2970 2971 2972 2973 2974 2975 2976 2977 2978 2979 2980 2981 2982 2983 2984
    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;
        }
    }
2985 2986
    if (server.aof_state != REDIS_AOF_OFF) {
        mem_used -= sdslen(server.aof_buf);
2987
        mem_used -= aofRewriteBufferSize();
2988
    }
2989

2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000
    /* 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;
3001

3002
        for (j = 0; j < server.dbnum; j++) {
3003
            long bestval = 0; /* just to prevent warning */
3004
            sds bestkey = NULL;
A
antirez 已提交
3005
            dictEntry *de;
3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022
            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);
3023
                bestkey = dictGetKey(de);
3024 3025 3026 3027 3028 3029
            }

            /* volatile-lru and allkeys-lru policy */
            else if (server.maxmemory_policy == REDIS_MAXMEMORY_ALLKEYS_LRU ||
                server.maxmemory_policy == REDIS_MAXMEMORY_VOLATILE_LRU)
            {
A
antirez 已提交
3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057
                struct evictionPoolEntry *pool = db->eviction_pool;

                while(bestkey == NULL) {
                    evictionPoolPopulate(dict, db->dict, db->eviction_pool);
                    /* Go backward from best to worst element to evict. */
                    for (k = REDIS_EVICTION_POOL_SIZE-1; k >= 0; k--) {
                        if (pool[k].key == NULL) continue;
                        de = dictFind(dict,pool[k].key);

                        /* Remove the entry from the pool. */
                        sdsfree(pool[k].key);
                        /* Shift all elements on its right to left. */
                        memmove(pool+k,pool+k+1,
                            sizeof(pool[0])*(REDIS_EVICTION_POOL_SIZE-k));
                        /* Clear the element on the right which is empty
                         * since we shifted one position to the left.  */
                        pool[REDIS_EVICTION_POOL_SIZE-1].key = NULL;
                        pool[REDIS_EVICTION_POOL_SIZE-1].idle = 0;

                        /* If the key exists, is our pick. Otherwise it is
                         * a ghost and we need to try the next element. */
                        if (de) {
                            bestkey = dictGetKey(de);
                            break;
                        } else {
                            /* Ghost... */
                            continue;
                        }
3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068
                    }
                }
            }

            /* 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);
3069 3070
                    thiskey = dictGetKey(de);
                    thisval = (long) dictGetVal(de);
3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082

                    /* 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) {
3083 3084
                long long delta;

3085
                robj *keyobj = createStringObject(bestkey,sdslen(bestkey));
3086
                propagateExpire(db,keyobj);
3087 3088 3089 3090 3091 3092 3093 3094 3095
                /* 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();
3096
                dbDelete(db,keyobj);
3097 3098
                delta -= (long long) zmalloc_used_memory();
                mem_freed += delta;
3099
                server.stat_evictedkeys++;
3100 3101
                notifyKeyspaceEvent(REDIS_NOTIFY_EVICTED, "evicted",
                    keyobj, db->id);
3102
                decrRefCount(keyobj);
3103 3104 3105 3106 3107 3108
                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
3109
                if (slaves) flushSlavesOutputBuffers();
3110 3111
            }
        }
3112
        if (!keys_freed) return REDIS_ERR; /* nothing to free... */
3113
    }
3114
    return REDIS_OK;
3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140
}

/* =================================== 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__ */

3141 3142 3143 3144
void createPidFile(void) {
    /* Try to write the pid file in a best-effort way. */
    FILE *fp = fopen(server.pidfile,"w");
    if (fp) {
3145
        fprintf(fp,"%d\n",(int)getpid());
3146 3147 3148 3149
        fclose(fp);
    }
}

3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167
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() {
3168
    printf("Redis server v=%s sha=%s:%d malloc=%s bits=%d build=%llx\n",
A
antirez 已提交
3169 3170 3171 3172
        REDIS_VERSION,
        redisGitSHA1(),
        atoi(redisGitDirty()) > 0,
        ZMALLOC_LIB,
3173
        sizeof(long) == 4 ? 32 : 64,
3174
        (unsigned long long) redisBuildId());
3175 3176 3177 3178
    exit(0);
}

void usage() {
3179
    fprintf(stderr,"Usage: ./redis-server [/path/to/redis.conf] [options]\n");
3180
    fprintf(stderr,"       ./redis-server - (read config from stdin)\n");
3181
    fprintf(stderr,"       ./redis-server -v or --version\n");
3182 3183
    fprintf(stderr,"       ./redis-server -h or --help\n");
    fprintf(stderr,"       ./redis-server --test-memory <megabytes>\n\n");
3184 3185 3186 3187 3188
    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");
3189 3190 3191
    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");
3192 3193 3194
    exit(1);
}

A
antirez 已提交
3195 3196 3197
void redisAsciiArt(void) {
#include "asciilogo.h"
    char *buf = zmalloc(1024*16);
3198 3199 3200 3201
    char *mode = "stand alone";

    if (server.cluster_enabled) mode = "cluster";
    else if (server.sentinel_mode) mode = "sentinel";
A
antirez 已提交
3202 3203 3204 3205 3206 3207

    snprintf(buf,1024*16,ascii_logo,
        REDIS_VERSION,
        redisGitSHA1(),
        strtol(redisGitDirty(),NULL,10) > 0,
        (sizeof(long) == 8) ? "64" : "32",
3208
        mode, server.port,
A
antirez 已提交
3209 3210 3211 3212 3213 3214
        (long) getpid()
    );
    redisLogRaw(REDIS_NOTICE|REDIS_LOG_RAW,buf);
    zfree(buf);
}

3215
static void sigtermHandler(int sig) {
3216 3217
    REDIS_NOTUSED(sig);

A
antirez 已提交
3218
    redisLogFromHandler(REDIS_WARNING,"Received SIGTERM, scheduling shutdown...");
3219 3220 3221
    server.shutdown_asap = 1;
}

3222
void setupSignalHandlers(void) {
3223 3224
    struct sigaction act;

3225 3226 3227
    /* When the SA_SIGINFO flag is set in sa_flags then sa_sigaction is used.
     * Otherwise, sa_handler is used. */
    sigemptyset(&act.sa_mask);
3228
    act.sa_flags = 0;
3229
    act.sa_handler = sigtermHandler;
3230
    sigaction(SIGTERM, &act, NULL);
3231

3232 3233
#ifdef HAVE_BACKTRACE
    sigemptyset(&act.sa_mask);
3234
    act.sa_flags = SA_NODEFER | SA_RESETHAND | SA_SIGINFO;
3235 3236 3237 3238 3239 3240 3241
    act.sa_sigaction = sigsegvHandler;
    sigaction(SIGSEGV, &act, NULL);
    sigaction(SIGBUS, &act, NULL);
    sigaction(SIGFPE, &act, NULL);
    sigaction(SIGILL, &act, NULL);
#endif
    return;
3242 3243
}

3244 3245
void memtest(size_t megabytes, int passes);

3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265 3266 3267
/* 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) {
3268
            redisLog(REDIS_WARNING,"Fatal error loading the DB: %s. Exiting.",strerror(errno));
3269 3270 3271 3272 3273
            exit(1);
        }
    }
}

A
antirez 已提交
3274 3275 3276
void redisOutOfMemoryHandler(size_t allocation_size) {
    redisLog(REDIS_WARNING,"Out Of Memory allocating %zu bytes!",
        allocation_size);
3277
    redisPanic("Redis aborting for OUT OF MEMORY");
A
antirez 已提交
3278 3279
}

3280
void redisSetProcTitle(char *title) {
3281
#ifdef USE_SETPROCTITLE
3282 3283 3284 3285 3286
    char *server_mode = "";
    if (server.cluster_enabled) server_mode = " [cluster]";
    else if (server.sentinel_mode) server_mode = " [sentinel]";

    setproctitle("%s %s:%d%s",
3287
        title,
A
antirez 已提交
3288
        server.bindaddr_count ? server.bindaddr[0] : "*",
3289 3290
        server.port,
        server_mode);
3291 3292 3293
#else
    REDIS_NOTUSED(title);
#endif
3294 3295
}

3296
int main(int argc, char **argv) {
3297
    struct timeval tv;
3298

A
antirez 已提交
3299
    /* We need to initialize our libraries, and the server configuration. */
3300 3301 3302
#ifdef INIT_SETPROCTITLE_REPLACEMENT
    spt_init(argc, argv);
#endif
3303
    setlocale(LC_COLLATE,"");
3304
    zmalloc_enable_thread_safeness();
A
antirez 已提交
3305
    zmalloc_set_oom_handler(redisOutOfMemoryHandler);
3306 3307 3308
    srand(time(NULL)^getpid());
    gettimeofday(&tv,NULL);
    dictSetHashFunctionSeed(tv.tv_sec^tv.tv_usec^getpid());
3309
    server.sentinel_mode = checkForSentinelMode(argc,argv);
3310
    initServerConfig();
3311

3312 3313 3314 3315 3316 3317 3318 3319
    /* 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();
    }

3320 3321 3322 3323 3324 3325
    if (argc >= 2) {
        int j = 1; /* First option to parse in argv[] */
        sds options = sdsempty();
        char *configfile = NULL;

        /* Handle special options --help and --version */
3326 3327
        if (strcmp(argv[1], "-v") == 0 ||
            strcmp(argv[1], "--version") == 0) version();
3328 3329
        if (strcmp(argv[1], "--help") == 0 ||
            strcmp(argv[1], "-h") == 0) usage();
3330 3331
        if (strcmp(argv[1], "--test-memory") == 0) {
            if (argc == 3) {
3332
                memtest(atoi(argv[2]),50);
3333 3334 3335 3336 3337 3338 3339 3340
                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);
            }
        }

3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360
        /* 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++;
        }
3361
        if (configfile) server.configfile = getAbsolutePath(configfile);
3362
        resetServerSaveParams();
3363 3364
        loadServerConfig(configfile,options);
        sdsfree(options);
3365
    } else {
3366
        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");
3367 3368 3369 3370
    }
    if (server.daemonize) daemonize();
    initServer();
    if (server.daemonize) createPidFile();
3371
    redisSetProcTitle(argv[0]);
3372
    redisAsciiArt();
3373 3374

    if (!server.sentinel_mode) {
A
antirez 已提交
3375
        /* Things not needed when running in Sentinel mode. */
3376 3377 3378 3379 3380
        redisLog(REDIS_WARNING,"Server started, Redis version " REDIS_VERSION);
    #ifdef __linux__
        linuxOvercommitMemoryWarning();
    #endif
        loadDataFromDisk();
3381 3382 3383 3384 3385 3386 3387 3388
        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);
            }
        }
3389
        if (server.ipfd_count > 0)
3390 3391 3392
            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);
3393
    } else {
A
antirez 已提交
3394
        sentinelIsRunning();
3395
    }
3396

3397 3398 3399 3400 3401
    /* 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);
    }

3402 3403 3404 3405 3406 3407
    aeSetBeforeSleepProc(server.el,beforeSleep);
    aeMain(server.el);
    aeDeleteEventLoop(server.el);
    return 0;
}

3408
/* The End */