redis.c 141.6 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
#include "latency.h"
35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51

#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>
52
#include <sys/resource.h>
53
#include <sys/utsname.h>
54
#include <locale.h>
55 56 57 58 59

/* Our shared "common" objects */

struct sharedObjectsStruct shared;

60
/* Global vars that are actually used as constants. The following double
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 */
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'.
118 119 120 121
 * F: Fast command: O(1) or O(log(N)) command that should never delay
 *    its execution as long as the kernel scheduler is giving us time.
 *    Note that commands that may trigger a DEL as a side effect (like SET)
 *    are not fast commands.
122
 */
123
struct redisCommand redisCommandTable[] = {
124
    {"get",getCommand,2,"rF",0,NULL,1,1,1,0,0},
125
    {"set",setCommand,-3,"wm",0,NULL,1,1,1,0,0},
126
    {"setnx",setnxCommand,3,"wmF",0,NULL,1,1,1,0,0},
127 128
    {"setex",setexCommand,4,"wm",0,NULL,1,1,1,0,0},
    {"psetex",psetexCommand,4,"wm",0,NULL,1,1,1,0,0},
129
    {"append",appendCommand,3,"wm",0,NULL,1,1,1,0,0},
130
    {"strlen",strlenCommand,2,"rF",0,NULL,1,1,1,0,0},
131
    {"del",delCommand,-2,"w",0,NULL,1,-1,1,0,0},
132
    {"exists",existsCommand,2,"rF",0,NULL,1,1,1,0,0},
133
    {"setbit",setbitCommand,4,"wm",0,NULL,1,1,1,0,0},
134
    {"getbit",getbitCommand,3,"rF",0,NULL,1,1,1,0,0},
135 136 137
    {"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},
138 139
    {"incr",incrCommand,2,"wmF",0,NULL,1,1,1,0,0},
    {"decr",decrCommand,2,"wmF",0,NULL,1,1,1,0,0},
140
    {"mget",mgetCommand,-2,"r",0,NULL,1,-1,1,0,0},
141 142 143 144
    {"rpush",rpushCommand,-3,"wmF",0,NULL,1,1,1,0,0},
    {"lpush",lpushCommand,-3,"wmF",0,NULL,1,1,1,0,0},
    {"rpushx",rpushxCommand,3,"wmF",0,NULL,1,1,1,0,0},
    {"lpushx",lpushxCommand,3,"wmF",0,NULL,1,1,1,0,0},
145
    {"linsert",linsertCommand,5,"wm",0,NULL,1,1,1,0,0},
146 147
    {"rpop",rpopCommand,2,"wF",0,NULL,1,1,1,0,0},
    {"lpop",lpopCommand,2,"wF",0,NULL,1,1,1,0,0},
A
antirez 已提交
148 149 150
    {"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},
151
    {"llen",llenCommand,2,"rF",0,NULL,1,1,1,0,0},
152 153 154 155 156 157
    {"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},
158 159 160 161 162 163
    {"sadd",saddCommand,-3,"wmF",0,NULL,1,1,1,0,0},
    {"srem",sremCommand,-3,"wF",0,NULL,1,1,1,0,0},
    {"smove",smoveCommand,4,"wF",0,NULL,1,2,1,0,0},
    {"sismember",sismemberCommand,3,"rF",0,NULL,1,1,1,0,0},
    {"scard",scardCommand,2,"rF",0,NULL,1,1,1,0,0},
    {"spop",spopCommand,2,"wRsF",0,NULL,1,1,1,0,0},
164
    {"srandmember",srandmemberCommand,-2,"rR",0,NULL,1,1,1,0,0},
165
    {"sinter",sinterCommand,-2,"rS",0,NULL,1,-1,1,0,0},
166
    {"sinterstore",sinterstoreCommand,-3,"wm",0,NULL,1,-1,1,0,0},
167
    {"sunion",sunionCommand,-2,"rS",0,NULL,1,-1,1,0,0},
168
    {"sunionstore",sunionstoreCommand,-3,"wm",0,NULL,1,-1,1,0,0},
169
    {"sdiff",sdiffCommand,-2,"rS",0,NULL,1,-1,1,0,0},
170
    {"sdiffstore",sdiffstoreCommand,-3,"wm",0,NULL,1,-1,1,0,0},
171
    {"smembers",sinterCommand,2,"rS",0,NULL,1,1,1,0,0},
A
antirez 已提交
172
    {"sscan",sscanCommand,-3,"rR",0,NULL,1,1,1,0,0},
173 174 175
    {"zadd",zaddCommand,-4,"wmF",0,NULL,1,1,1,0,0},
    {"zincrby",zincrbyCommand,4,"wmF",0,NULL,1,1,1,0,0},
    {"zrem",zremCommand,-3,"wF",0,NULL,1,1,1,0,0},
176 177
    {"zremrangebyscore",zremrangebyscoreCommand,4,"w",0,NULL,1,1,1,0,0},
    {"zremrangebyrank",zremrangebyrankCommand,4,"w",0,NULL,1,1,1,0,0},
A
antirez 已提交
178
    {"zremrangebylex",zremrangebylexCommand,4,"w",0,NULL,1,1,1,0,0},
179 180 181 182 183
    {"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},
184 185
    {"zrangebylex",zrangebylexCommand,-4,"r",0,NULL,1,1,1,0,0},
    {"zrevrangebylex",zrevrangebylexCommand,-4,"r",0,NULL,1,1,1,0,0},
186 187
    {"zcount",zcountCommand,4,"rF",0,NULL,1,1,1,0,0},
    {"zlexcount",zlexcountCommand,4,"rF",0,NULL,1,1,1,0,0},
188
    {"zrevrange",zrevrangeCommand,-4,"r",0,NULL,1,1,1,0,0},
189 190 191 192
    {"zcard",zcardCommand,2,"rF",0,NULL,1,1,1,0,0},
    {"zscore",zscoreCommand,3,"rF",0,NULL,1,1,1,0,0},
    {"zrank",zrankCommand,3,"rF",0,NULL,1,1,1,0,0},
    {"zrevrank",zrevrankCommand,3,"rF",0,NULL,1,1,1,0,0},
A
antirez 已提交
193
    {"zscan",zscanCommand,-3,"rR",0,NULL,1,1,1,0,0},
194 195 196
    {"hset",hsetCommand,4,"wmF",0,NULL,1,1,1,0,0},
    {"hsetnx",hsetnxCommand,4,"wmF",0,NULL,1,1,1,0,0},
    {"hget",hgetCommand,3,"rF",0,NULL,1,1,1,0,0},
197 198
    {"hmset",hmsetCommand,-4,"wm",0,NULL,1,1,1,0,0},
    {"hmget",hmgetCommand,-3,"r",0,NULL,1,1,1,0,0},
199 200 201 202
    {"hincrby",hincrbyCommand,4,"wmF",0,NULL,1,1,1,0,0},
    {"hincrbyfloat",hincrbyfloatCommand,4,"wmF",0,NULL,1,1,1,0,0},
    {"hdel",hdelCommand,-3,"wF",0,NULL,1,1,1,0,0},
    {"hlen",hlenCommand,2,"rF",0,NULL,1,1,1,0,0},
203 204
    {"hkeys",hkeysCommand,2,"rS",0,NULL,1,1,1,0,0},
    {"hvals",hvalsCommand,2,"rS",0,NULL,1,1,1,0,0},
205
    {"hgetall",hgetallCommand,2,"r",0,NULL,1,1,1,0,0},
206
    {"hexists",hexistsCommand,3,"rF",0,NULL,1,1,1,0,0},
A
antirez 已提交
207
    {"hscan",hscanCommand,-3,"rR",0,NULL,1,1,1,0,0},
208 209 210
    {"incrby",incrbyCommand,3,"wmF",0,NULL,1,1,1,0,0},
    {"decrby",decrbyCommand,3,"wmF",0,NULL,1,1,1,0,0},
    {"incrbyfloat",incrbyfloatCommand,3,"wmF",0,NULL,1,1,1,0,0},
211 212 213
    {"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},
214
    {"randomkey",randomkeyCommand,1,"rR",0,NULL,0,0,0,0,0},
215 216
    {"select",selectCommand,2,"rlF",0,NULL,0,0,0,0,0},
    {"move",moveCommand,3,"wF",0,NULL,1,1,1,0,0},
217
    {"rename",renameCommand,3,"w",0,NULL,1,2,1,0,0},
218 219 220 221 222
    {"renamenx",renamenxCommand,3,"wF",0,NULL,1,2,1,0,0},
    {"expire",expireCommand,3,"wF",0,NULL,1,1,1,0,0},
    {"expireat",expireatCommand,3,"wF",0,NULL,1,1,1,0,0},
    {"pexpire",pexpireCommand,3,"wF",0,NULL,1,1,1,0,0},
    {"pexpireat",pexpireatCommand,3,"wF",0,NULL,1,1,1,0,0},
223
    {"keys",keysCommand,2,"rS",0,NULL,0,0,0,0,0},
224
    {"scan",scanCommand,-2,"rR",0,NULL,0,0,0,0,0},
225 226
    {"dbsize",dbsizeCommand,1,"rF",0,NULL,0,0,0,0,0},
    {"auth",authCommand,2,"rsltF",0,NULL,0,0,0,0,0},
227
    {"ping",pingCommand,-1,"rtF",0,NULL,0,0,0,0,0},
228
    {"echo",echoCommand,2,"rF",0,NULL,0,0,0,0,0},
A
antirez 已提交
229
    {"save",saveCommand,1,"ars",0,NULL,0,0,0,0,0},
230 231
    {"bgsave",bgsaveCommand,1,"ar",0,NULL,0,0,0,0,0},
    {"bgrewriteaof",bgrewriteaofCommand,1,"ar",0,NULL,0,0,0,0,0},
232
    {"shutdown",shutdownCommand,-1,"arlt",0,NULL,0,0,0,0,0},
233 234 235
    {"lastsave",lastsaveCommand,1,"rRF",0,NULL,0,0,0,0,0},
    {"type",typeCommand,2,"rF",0,NULL,1,1,1,0,0},
    {"multi",multiCommand,1,"rsF",0,NULL,0,0,0,0,0},
236
    {"exec",execCommand,1,"sM",0,NULL,0,0,0,0,0},
237
    {"discard",discardCommand,1,"rsF",0,NULL,0,0,0,0,0},
238
    {"sync",syncCommand,1,"ars",0,NULL,0,0,0,0,0},
239
    {"psync",syncCommand,3,"ars",0,NULL,0,0,0,0,0},
A
antirez 已提交
240
    {"replconf",replconfCommand,-1,"arslt",0,NULL,0,0,0,0,0},
241 242
    {"flushdb",flushdbCommand,1,"w",0,NULL,0,0,0,0,0},
    {"flushall",flushallCommand,1,"w",0,NULL,0,0,0,0,0},
243
    {"sort",sortCommand,-2,"wm",0,sortGetKeys,1,1,1,0,0},
244
    {"info",infoCommand,-1,"rlt",0,NULL,0,0,0,0,0},
245
    {"monitor",monitorCommand,1,"ars",0,NULL,0,0,0,0,0},
246 247 248
    {"ttl",ttlCommand,2,"rF",0,NULL,1,1,1,0,0},
    {"pttl",pttlCommand,2,"rF",0,NULL,1,1,1,0,0},
    {"persist",persistCommand,2,"wF",0,NULL,1,1,1,0,0},
249
    {"slaveof",slaveofCommand,3,"ast",0,NULL,0,0,0,0,0},
A
antirez 已提交
250
    {"role",roleCommand,1,"last",0,NULL,0,0,0,0,0},
251
    {"debug",debugCommand,-2,"as",0,NULL,0,0,0,0,0},
252
    {"config",configCommand,-2,"art",0,NULL,0,0,0,0,0},
253 254 255 256
    {"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},
257
    {"publish",publishCommand,3,"pltrF",0,NULL,0,0,0,0,0},
A
antirez 已提交
258
    {"pubsub",pubsubCommand,-2,"pltrR",0,NULL,0,0,0,0,0},
259 260
    {"watch",watchCommand,-2,"rsF",0,NULL,1,-1,1,0,0},
    {"unwatch",unwatchCommand,1,"rsF",0,NULL,0,0,0,0,0},
261
    {"cluster",clusterCommand,-2,"ar",0,NULL,0,0,0,0,0},
A
antirez 已提交
262
    {"restore",restoreCommand,-4,"awm",0,NULL,1,1,1,0,0},
263
    {"restore-asking",restoreCommand,-4,"awmk",0,NULL,1,1,1,0,0},
A
antirez 已提交
264
    {"migrate",migrateCommand,-6,"aw",0,NULL,0,0,0,0,0},
265
    {"asking",askingCommand,1,"r",0,NULL,0,0,0,0,0},
266 267
    {"readonly",readonlyCommand,1,"rF",0,NULL,0,0,0,0,0},
    {"readwrite",readwriteCommand,1,"rF",0,NULL,0,0,0,0,0},
268
    {"dump",dumpCommand,2,"ar",0,NULL,1,1,1,0,0},
M
michael-grunder 已提交
269
    {"object",objectCommand,3,"r",0,NULL,2,2,2,0,0},
M
Matt Stancliff 已提交
270
    {"client",clientCommand,-2,"ars",0,NULL,0,0,0,0,0},
271 272
    {"eval",evalCommand,-3,"s",0,evalGetKeys,0,0,0,0,0},
    {"evalsha",evalShaCommand,-3,"s",0,evalGetKeys,0,0,0,0,0},
273
    {"slowlog",slowlogCommand,-2,"r",0,NULL,0,0,0,0,0},
A
antirez 已提交
274
    {"script",scriptCommand,-2,"ras",0,NULL,0,0,0,0,0},
275
    {"time",timeCommand,1,"rRF",0,NULL,0,0,0,0,0},
A
antirez 已提交
276
    {"bitop",bitopCommand,-4,"wm",0,NULL,2,-1,1,0,0},
277
    {"bitcount",bitcountCommand,-2,"r",0,NULL,1,1,1,0,0},
A
antirez 已提交
278
    {"bitpos",bitposCommand,-3,"r",0,NULL,1,1,1,0,0},
A
antirez 已提交
279
    {"wait",waitCommand,3,"rs",0,NULL,0,0,0,0,0},
A
antirez 已提交
280
    {"command",commandCommand,0,"rlt",0,NULL,0,0,0,0,0},
281
    {"pfselftest",pfselftestCommand,1,"r",0,NULL,0,0,0,0,0},
282
    {"pfadd",pfaddCommand,-2,"wmF",0,NULL,1,1,1,0,0},
283
    {"pfcount",pfcountCommand,-2,"r",0,NULL,1,1,1,0,0},
A
antirez 已提交
284
    {"pfmerge",pfmergeCommand,-2,"wm",0,NULL,1,-1,1,0,0},
A
antirez 已提交
285 286
    {"pfdebug",pfdebugCommand,-3,"w",0,NULL,0,0,0,0,0},
    {"latency",latencyCommand,-2,"arslt",0,NULL,0,0,0,0,0}
287 288
};

A
antirez 已提交
289 290
struct evictionPoolEntry *evictionPoolAlloc(void);

291 292
/*============================ Utility functions ============================ */

293 294 295
/* 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
296 297
    const int syslogLevelMap[] = { LOG_DEBUG, LOG_INFO, LOG_NOTICE, LOG_WARNING };
    const char *c = ".-*#";
298
    FILE *fp;
299
    char buf[64];
A
antirez 已提交
300
    int rawmode = (level & REDIS_LOG_RAW);
301
    int log_to_stdout = server.logfile[0] == '\0';
302

A
antirez 已提交
303
    level &= 0xff; /* clear flags */
304
    if (level < server.verbosity) return;
305

306
    fp = log_to_stdout ? stdout : fopen(server.logfile,"a");
307 308
    if (!fp) return;

A
antirez 已提交
309 310 311
    if (rawmode) {
        fprintf(fp,"%s",msg);
    } else {
312 313
        int off;
        struct timeval tv;
A
antirez 已提交
314 315
        int role_char;
        pid_t pid = getpid();
316 317 318 319

        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
320 321 322 323 324 325 326 327 328
        if (server.sentinel_mode) {
            role_char = 'X'; /* Sentinel. */
        } else if (pid != server.pid) {
            role_char = 'C'; /* RDB / AOF writing child. */
        } else {
            role_char = (server.masterhost ? 'S':'M'); /* Slave or Master. */
        }
        fprintf(fp,"%d:%c %s %c %s\n",
            (int)getpid(),role_char, buf,c[level],msg);
A
antirez 已提交
329
    }
J
Jonah H. Harris 已提交
330 331
    fflush(fp);

332
    if (!log_to_stdout) fclose(fp);
J
Jonah H. Harris 已提交
333
    if (server.syslog_enabled) syslog(syslogLevelMap[level], "%s", msg);
334 335
}

G
guiquanz 已提交
336
/* Like redisLogRaw() but with printf-alike support. This is the function that
337 338 339 340 341 342
 * 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
343
    if ((level&0xff) < server.verbosity) return;
344 345 346 347 348 349 350 351

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

    redisLogRaw(level,msg);
}

A
antirez 已提交
352 353 354 355 356 357 358 359
/* 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;
360
    int log_to_stdout = server.logfile[0] == '\0';
A
antirez 已提交
361 362
    char buf[64];

363 364
    if ((level&0xff) < server.verbosity || (log_to_stdout && server.daemonize))
        return;
365
    fd = log_to_stdout ? STDOUT_FILENO :
366
                         open(server.logfile, O_APPEND|O_CREAT|O_WRONLY, 0644);
A
antirez 已提交
367 368
    if (fd == -1) return;
    ll2string(buf,sizeof(buf),getpid());
369
    if (write(fd,buf,strlen(buf)) == -1) goto err;
370
    if (write(fd,":signal-handler (",17) == -1) goto err;
A
antirez 已提交
371
    ll2string(buf,sizeof(buf),time(NULL));
372 373 374 375 376
    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:
377
    if (!log_to_stdout) close(fd);
A
antirez 已提交
378 379
}

A
antirez 已提交
380 381 382 383 384 385 386 387 388 389 390
/* 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;
}

391 392 393 394 395
/* Return the UNIX time in milliseconds */
long long mstime(void) {
    return ustime()/1000;
}

396 397 398 399 400 401 402 403 404 405 406 407
/* 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
}

408 409
/*====================== Hash table type implementation  ==================== */

410
/* This is a hash table type that uses the SDS dynamic strings library as
T
T.J. Schuck 已提交
411
 * keys and redis objects as values (objects can hold SDS strings,
412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437
 * 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
438 439
/* A case insensitive version used for the command lookup table and other
 * places where case insensitive non binary-safe comparison is needed. */
440 441 442 443 444 445 446 447
int dictSdsKeyCaseCompare(void *privdata, const void *key1,
        const void *key2)
{
    DICT_NOTUSED(privdata);

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

448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478
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));
}

479 480 481 482
unsigned int dictSdsCaseHash(const void *key) {
    return dictGenCaseHashFunction((unsigned char*)key, sdslen((char*)key));
}

483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503
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;

504
    if (sdsEncodedObject(o)) {
505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523
        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;
        }
    }
}

524
/* Sets type hash table */
525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540
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 */
541
    NULL                       /* val destructor */
542 543 544 545 546 547 548 549 550 551 552 553
};

/* 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
554 555 556 557 558 559 560 561 562 563
/* 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 */
};

564 565 566 567 568 569 570 571 572 573
/* Db->expires */
dictType keyptrDictType = {
    dictSdsHash,               /* hash function */
    NULL,                      /* key dup */
    NULL,                      /* val dup */
    dictSdsKeyCompare,         /* key compare */
    NULL,                      /* key destructor */
    NULL                       /* val destructor */
};

574 575 576 577 578 579 580 581 582 583
/* 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 */
};

584
/* Hash type hash table (note that small hashes are represented with ziplists) */
585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605
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
606 607 608 609 610 611 612 613 614 615 616
/* 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 */
};

617 618 619 620 621 622 623 624 625 626 627 628
/* 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
629 630 631 632 633 634 635 636 637 638
/* Migrate cache dict type. */
dictType migrateCacheDictType = {
    dictSdsHash,                /* hash function */
    NULL,                       /* key dup */
    NULL,                       /* val dup */
    dictSdsKeyCompare,          /* key compare */
    dictSdsDestructor,          /* key destructor */
    NULL                        /* val destructor */
};

639 640 641 642
/* 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 = {
643
    dictSdsCaseHash,            /* hash function */
644 645
    NULL,                       /* key dup */
    NULL,                       /* val dup */
646
    dictSdsKeyCaseCompare,      /* key compare */
647 648 649 650
    dictSdsDestructor,          /* key destructor */
    NULL                        /* val destructor */
};

651 652 653 654 655 656 657 658 659 660 661
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 */
662 663 664 665 666
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);
667 668 669 670 671
}

/* 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
672 673 674 675 676 677 678 679 680 681 682 683 684 685
 * 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... */
686
    }
687
    return 0;
688 689 690 691 692 693 694 695 696
}

/* 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
697
    if (server.rdb_child_pid == -1 && server.aof_child_pid == -1)
698 699 700 701 702 703 704
        dictEnableResize();
    else
        dictDisableResize();
}

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

705 706 707 708 709 710 711 712 713 714 715
/* 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
716
int activeExpireCycleTryExpire(redisDb *db, dictEntry *de, long long now) {
717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733
    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;
    }
}

734 735 736
/* 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
737 738
 * keys that can be removed from the keyspace.
 *
739
 * No more than REDIS_DBCRON_DBS_PER_CALL databases are tested at every
740 741
 * iteration.
 *
742 743
 * 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
744 745 746 747 748 749 750 751 752 753 754
 * 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. */
755

A
antirez 已提交
756
void activeExpireCycle(int type) {
757 758 759 760
    /* 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
761
    static long long last_fast_cycle = 0; /* When last fast cycle ran. */
762

763 764
    int j, iteration = 0;
    int dbs_per_call = REDIS_DBCRON_DBS_PER_CALL;
765
    long long start = ustime(), timelimit;
766

A
antirez 已提交
767
    if (type == ACTIVE_EXPIRE_CYCLE_FAST) {
768 769 770 771
        /* 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;
772
        if (start < last_fast_cycle + ACTIVE_EXPIRE_CYCLE_FAST_DURATION*2) return;
773 774
        last_fast_cycle = start;
    }
775

776 777 778 779 780 781 782 783 784
    /* 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;
785

A
antirez 已提交
786
    /* We can use at max ACTIVE_EXPIRE_CYCLE_SLOW_TIME_PERC percentage of CPU time
787
     * per iteration. Since this function gets called with a frequency of
788
     * server.hz times per second, the following is the max amount of
789
     * microseconds we can spend in this function. */
A
antirez 已提交
790
    timelimit = 1000000*ACTIVE_EXPIRE_CYCLE_SLOW_TIME_PERC/server.hz/100;
791
    timelimit_exit = 0;
792
    if (timelimit <= 0) timelimit = 1;
793

A
antirez 已提交
794 795
    if (type == ACTIVE_EXPIRE_CYCLE_FAST)
        timelimit = ACTIVE_EXPIRE_CYCLE_FAST_DURATION; /* in microseconds. */
796

797
    for (j = 0; j < dbs_per_call; j++) {
798
        int expired;
799 800 801 802 803 804
        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++;
805 806 807 808

        /* Continue to expire if at the end of the cycle more than 25%
         * of the keys were expired. */
        do {
809
            unsigned long num, slots;
810 811
            long long now, ttl_sum;
            int ttl_samples;
812 813

            /* If there is nothing to expire try next DB ASAP. */
814 815 816 817
            if ((num = dictSize(db->expires)) == 0) {
                db->avg_ttl = 0;
                break;
            }
818 819
            slots = dictSlots(db->expires);
            now = mstime();
820

821 822 823 824 825 826
            /* 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;

827 828
            /* The main collection cycle. Sample random keys among keys
             * with an expire set, checking for expired ones. */
829
            expired = 0;
830 831 832
            ttl_sum = 0;
            ttl_samples = 0;

A
antirez 已提交
833 834
            if (num > ACTIVE_EXPIRE_CYCLE_LOOKUPS_PER_LOOP)
                num = ACTIVE_EXPIRE_CYCLE_LOOKUPS_PER_LOOP;
835

836 837
            while (num--) {
                dictEntry *de;
838
                long long ttl;
839 840

                if ((de = dictGetRandomKey(db->expires)) == NULL) break;
841
                ttl = dictGetSignedIntegerVal(de)-now;
842
                if (activeExpireCycleTryExpire(db,de,now)) expired++;
843 844 845
                if (ttl < 0) ttl = 0;
                ttl_sum += ttl;
                ttl_samples++;
846
            }
847 848 849 850 851 852 853 854 855 856

            /* 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;
            }

857 858 859 860
            /* 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++;
861 862 863 864 865
            if ((iteration & 0xf) == 0) { /* check once every 16 iterations. */
                long long elapsed = ustime()-start;

                latencyAddSampleIfNeeded("expire-cycle",elapsed/1000);
                if (elapsed > timelimit) timelimit_exit = 1;
866
            }
867
            if (timelimit_exit) return;
A
antirez 已提交
868 869 870
            /* 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);
871 872 873
    }
}

874 875
unsigned int getLRUClock(void) {
    return (mstime()/REDIS_LRU_CLOCK_RESOLUTION) & REDIS_LRU_CLOCK_MAX;
876
}
877

878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901
/* 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
902 903
/* Check for timeouts. Returns non-zero if the client was terminated */
int clientsCronHandleTimeout(redisClient *c) {
A
antirez 已提交
904
    time_t now = server.unixtime;
905 906 907 908 909

    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 */
910
        !(c->flags & REDIS_PUBSUB) &&   /* no timeout for Pub/Sub clients */
911 912 913 914
        (now - c->lastinteraction > server.maxidletime))
    {
        redisLog(REDIS_VERBOSE,"Closing idle client");
        freeClient(c);
A
antirez 已提交
915
        return 1;
916
    } else if (c->flags & REDIS_BLOCKED) {
A
antirez 已提交
917 918 919 920 921 922
        /* 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) {
923 924
            replyToBlockedClientTimedOut(c);
            unblockClient(c);
925 926
        }
    }
A
antirez 已提交
927
    return 0;
928 929
}

930
/* The client query buffer is an sds.c string that can end with a lot of
A
antirez 已提交
931 932
 * free space not used, this function reclaims space if needed.
 *
G
guiquanz 已提交
933
 * The function always returns 0 as it never terminates the client. */
A
antirez 已提交
934
int clientsCronResizeQueryBuffer(redisClient *c) {
935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952
    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
953
    return 0;
954 955
}

956
void clientsCron(void) {
957 958
    /* 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
959 960 961
     * 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. */
962
    int numclients = listLength(server.clients);
963
    int iterations = numclients/(server.hz*10);
964

965 966
    if (iterations < 50)
        iterations = (numclients < 50) ? numclients : 50;
967 968 969 970 971 972 973 974 975 976
    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
977 978 979 980 981
        /* 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;
982 983 984
    }
}

985 986 987 988
/* This function handles 'background' operations we are required to do
 * incrementally in Redis databases, such as active key expiring, resizing,
 * rehashing. */
void databasesCron(void) {
989 990
    /* Expire keys by random sampling. Not required for slaves
     * as master will synthesize DELs for us. */
991
    if (server.active_expire_enabled && server.masterhost == NULL)
A
antirez 已提交
992
        activeExpireCycle(ACTIVE_EXPIRE_CYCLE_SLOW);
993

994 995 996
    /* 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. */
997
    if (server.rdb_child_pid == -1 && server.aof_child_pid == -1) {
998 999 1000
        /* 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. */
1001 1002
        static unsigned int resize_db = 0;
        static unsigned int rehash_db = 0;
1003 1004
        int dbs_per_call = REDIS_DBCRON_DBS_PER_CALL;
        int j;
1005

1006 1007 1008
        /* Don't test more DBs than we have. */
        if (dbs_per_call > server.dbnum) dbs_per_call = server.dbnum;

1009
        /* Resize */
1010
        for (j = 0; j < dbs_per_call; j++) {
1011 1012 1013 1014 1015 1016
            tryResizeHashTables(resize_db % server.dbnum);
            resize_db++;
        }

        /* Rehash */
        if (server.activerehashing) {
1017
            for (j = 0; j < dbs_per_call; j++) {
1018 1019 1020 1021 1022 1023 1024 1025 1026
                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;
                }
            }
        }
1027 1028 1029
    }
}

1030 1031 1032 1033 1034 1035 1036 1037 1038
/* 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();
}

1039
/* This is our timer interrupt, called server.hz times per second.
1040 1041 1042 1043 1044
 * 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
1045
 * - Software watchdog.
1046 1047 1048
 * - Update some statistic.
 * - Incremental rehashing of the DBs hash tables.
 * - Triggering BGSAVE / AOF rewrite, and handling of terminated children.
G
guiquanz 已提交
1049
 * - Clients timeout of different kinds.
1050 1051 1052
 * - Replication reconnection.
 * - Many more...
 *
1053
 * Everything directly called here will be called server.hz times per second,
1054 1055 1056 1057
 * so in order to throttle execution of things we want to do less frequently
 * a macro is used: run_with_period(milliseconds) { .... }
 */

1058
int serverCron(struct aeEventLoop *eventLoop, long long id, void *clientData) {
1059
    int j;
1060 1061 1062 1063
    REDIS_NOTUSED(eventLoop);
    REDIS_NOTUSED(id);
    REDIS_NOTUSED(clientData);

A
antirez 已提交
1064 1065 1066 1067
    /* 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);

1068 1069
    /* Update the time cache. */
    updateCachedTime();
A
antirez 已提交
1070

1071
    run_with_period(100) trackOperationsPerSecond();
1072

1073 1074
    /* We have just REDIS_LRU_BITS bits per object for LRU information.
     * So we use an (eventually wrapping) LRU clock.
1075
     *
1076 1077 1078 1079 1080
     * 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.
1081 1082
     *
     * Note that you can change the resolution altering the
1083 1084
     * REDIS_LRU_CLOCK_RESOLUTION define. */
    server.lruclock = getLRUClock();
1085

1086 1087 1088 1089
    /* 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
1090 1091 1092
    /* Sample the RSS here since this is a relatively slow call. */
    server.resident_set_size = zmalloc_get_rss();

1093 1094 1095
    /* 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) {
1096
        if (prepareForShutdown(0) == REDIS_OK) exit(0);
1097
        redisLog(REDIS_WARNING,"SIGTERM received but errors trying to shut down the server, check the logs for more information");
1098
        server.shutdown_asap = 0;
1099 1100 1101
    }

    /* Show some info about non-empty databases */
1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112
    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); */
            }
1113 1114 1115 1116
        }
    }

    /* Show information about connected clients */
1117 1118 1119
    if (!server.sentinel_mode) {
        run_with_period(5000) {
            redisLog(REDIS_VERBOSE,
1120
                "%lu clients connected (%lu slaves), %zu bytes in use",
1121 1122 1123 1124
                listLength(server.clients)-listLength(server.slaves),
                listLength(server.slaves),
                zmalloc_used_memory());
        }
1125 1126
    }

1127 1128
    /* We need to do a few operations on clients asynchronously. */
    clientsCron();
1129

1130 1131 1132
    /* Handle background operations on Redis databases. */
    databasesCron();

1133 1134
    /* Start a scheduled AOF rewrite if this was requested by the user while
     * a BGSAVE was in progress. */
A
antirez 已提交
1135
    if (server.rdb_child_pid == -1 && server.aof_child_pid == -1 &&
1136
        server.aof_rewrite_scheduled)
1137 1138 1139 1140
    {
        rewriteAppendOnlyFileBackground();
    }

A
antirez 已提交
1141
    /* Check if a background saving or AOF rewrite in progress terminated. */
A
antirez 已提交
1142
    if (server.rdb_child_pid != -1 || server.aof_child_pid != -1) {
1143 1144 1145 1146
        int statloc;
        pid_t pid;

        if ((pid = wait3(&statloc,WNOHANG,NULL)) != 0) {
1147 1148
            int exitcode = WEXITSTATUS(statloc);
            int bysignal = 0;
1149

1150 1151
            if (WIFSIGNALED(statloc)) bysignal = WTERMSIG(statloc);

A
antirez 已提交
1152
            if (pid == server.rdb_child_pid) {
1153
                backgroundSaveDoneHandler(exitcode,bysignal);
1154
            } else if (pid == server.aof_child_pid) {
1155
                backgroundRewriteDoneHandler(exitcode,bysignal);
1156 1157 1158 1159
            } else {
                redisLog(REDIS_WARNING,
                    "Warning, detected child with unmatched pid: %ld",
                    (long)pid);
1160 1161 1162
            }
            updateDictResizePolicy();
        }
A
antirez 已提交
1163
    } else {
1164 1165
        /* If there is not a background saving/rewrite in progress check if
         * we have to save/rewrite now */
1166 1167 1168
         for (j = 0; j < server.saveparamslen; j++) {
            struct saveparam *sp = server.saveparams+j;

1169 1170 1171 1172
            /* 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. */
1173
            if (server.dirty >= sp->changes &&
1174 1175 1176 1177 1178
                server.unixtime-server.lastsave > sp->seconds &&
                (server.unixtime-server.lastbgsave_try >
                 REDIS_BGSAVE_RETRY_DELAY ||
                 server.lastbgsave_status == REDIS_OK))
            {
1179
                redisLog(REDIS_NOTICE,"%d changes in %d seconds. Saving...",
1180
                    sp->changes, (int)sp->seconds);
A
antirez 已提交
1181
                rdbSaveBackground(server.rdb_filename);
1182 1183 1184
                break;
            }
         }
1185 1186

         /* Trigger an AOF rewrite if needed */
A
antirez 已提交
1187
         if (server.rdb_child_pid == -1 &&
A
antirez 已提交
1188
             server.aof_child_pid == -1 &&
1189 1190
             server.aof_rewrite_perc &&
             server.aof_current_size > server.aof_rewrite_min_size)
1191
         {
1192 1193 1194 1195
            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) {
1196
                redisLog(REDIS_NOTICE,"Starting automatic rewriting of AOF on %lld%% growth",growth);
1197 1198 1199
                rewriteAppendOnlyFileBackground();
            }
         }
1200 1201
    }

1202

1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213
    /* 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);
1214
    }
1215

1216 1217 1218
    /* Close clients that need to be closed asynchronous */
    freeClientsInAsyncFreeQueue();

1219 1220 1221
    /* Clear the paused clients flag if needed. */
    clientsArePaused(); /* Don't check return value, just use the side effect. */

1222 1223
    /* Replication cron function -- used to reconnect to master and
     * to detect transfer failures. */
1224
    run_with_period(1000) replicationCron();
1225

A
antirez 已提交
1226
    /* Run the Redis Cluster cron. */
1227
    run_with_period(100) {
1228 1229
        if (server.cluster_enabled) clusterCron();
    }
A
antirez 已提交
1230

A
antirez 已提交
1231
    /* Run the Sentinel timer if we are in sentinel mode. */
1232 1233 1234 1235
    run_with_period(100) {
        if (server.sentinel_mode) sentinelTimer();
    }

A
antirez 已提交
1236 1237 1238 1239 1240
    /* Cleanup expired MIGRATE cached sockets. */
    run_with_period(1000) {
        migrateCloseTimedoutSockets();
    }

1241
    server.cronloops++;
1242
    return 1000/server.hz;
1243 1244 1245 1246 1247 1248 1249 1250
}

/* 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);

1251 1252 1253 1254
    /* 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);
1255

1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275
    /* 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();

1276
    /* Try to process pending commands for clients that were just unblocked. */
1277 1278
    if (listLength(server.unblocked_clients))
        processUnblockedClients();
1279

1280
    /* Write the AOF buffer on disk */
1281
    flushAppendOnlyFile(0);
1282 1283 1284

    /* Call the Redis Cluster before sleep function. */
    if (server.cluster_enabled) clusterBeforeSleep();
1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303
}

/* =========================== 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
1304
    shared.emptyscan = createObject(REDIS_STRING,sdsnew("*2\r\n$1\r\n0\r\n*0\r\n"));
1305
    shared.wrongtypeerr = createObject(REDIS_STRING,sdsnew(
1306
        "-WRONGTYPE Operation against a key holding the wrong kind of value\r\n"));
1307 1308 1309 1310 1311 1312 1313 1314
    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
1315 1316
    shared.noscripterr = createObject(REDIS_STRING,sdsnew(
        "-NOSCRIPT No matching script. Please use EVAL.\r\n"));
1317 1318
    shared.loadingerr = createObject(REDIS_STRING,sdsnew(
        "-LOADING Redis is loading the dataset in memory\r\n"));
1319
    shared.slowscripterr = createObject(REDIS_STRING,sdsnew(
1320
        "-BUSY Redis is busy running a script. You can only call SCRIPT KILL or SHUTDOWN NOSAVE.\r\n"));
1321 1322
    shared.masterdownerr = createObject(REDIS_STRING,sdsnew(
        "-MASTERDOWN Link with MASTER is down and slave-serve-stale-data is set to 'no'.\r\n"));
1323
    shared.bgsaveerr = createObject(REDIS_STRING,sdsnew(
1324 1325 1326
        "-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"));
1327 1328
    shared.noautherr = createObject(REDIS_STRING,sdsnew(
        "-NOAUTH Authentication required.\r\n"));
1329 1330
    shared.oomerr = createObject(REDIS_STRING,sdsnew(
        "-OOM command not allowed when used memory > 'maxmemory'.\r\n"));
A
antirez 已提交
1331 1332
    shared.execaborterr = createObject(REDIS_STRING,sdsnew(
        "-EXECABORT Transaction discarded because of previous errors.\r\n"));
1333 1334
    shared.noreplicaserr = createObject(REDIS_STRING,sdsnew(
        "-NOREPLICAS Not enough good slaves to write.\r\n"));
1335 1336
    shared.busykeyerr = createObject(REDIS_STRING,sdsnew(
        "-BUSYKEY Target key name already exists.\r\n"));
1337 1338 1339
    shared.space = createObject(REDIS_STRING,sdsnew(" "));
    shared.colon = createObject(REDIS_STRING,sdsnew(":"));
    shared.plus = createObject(REDIS_STRING,sdsnew("+"));
1340 1341

    for (j = 0; j < REDIS_SHARED_SELECT_CMDS; j++) {
1342 1343 1344 1345
        char dictid_str[64];
        int dictid_len;

        dictid_len = ll2string(dictid_str,sizeof(dictid_str),j);
1346
        shared.select[j] = createObject(REDIS_STRING,
1347 1348 1349
            sdscatprintf(sdsempty(),
                "*2\r\n$6\r\nSELECT\r\n$%d\r\n%s\r\n",
                dictid_len, dictid_str));
1350
    }
1351 1352 1353 1354 1355 1356
    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("$12\r\npunsubscribe\r\n",19);
1357
    shared.del = createStringObject("DEL",3);
1358 1359
    shared.rpop = createStringObject("RPOP",4);
    shared.lpop = createStringObject("LPOP",4);
1360
    shared.lpush = createStringObject("LPUSH",5);
1361 1362 1363 1364
    for (j = 0; j < REDIS_SHARED_INTEGERS; j++) {
        shared.integers[j] = createObject(REDIS_STRING,(void*)(long)j);
        shared.integers[j]->encoding = REDIS_ENCODING_INT;
    }
1365 1366 1367 1368 1369 1370
    for (j = 0; j < REDIS_SHARED_BULKHDR_LEN; j++) {
        shared.mbulkhdr[j] = createObject(REDIS_STRING,
            sdscatprintf(sdsempty(),"*%d\r\n",j));
        shared.bulkhdr[j] = createObject(REDIS_STRING,
            sdscatprintf(sdsempty(),"$%d\r\n",j));
    }
1371 1372 1373 1374 1375 1376
    /* The following two shared objects, minstring and maxstrings, are not
     * actually used for their value but as a special object meaning
     * respectively the minimum possible string and the maximum possible
     * string in string comparisons for the ZRANGEBYLEX command. */
    shared.minstring = createStringObject("minstring",9);
    shared.maxstring = createStringObject("maxstring",9);
1377 1378
}

1379
void initServerConfig(void) {
1380 1381
    int j;

A
antirez 已提交
1382
    getRandomHexChars(server.runid,REDIS_RUN_ID_SIZE);
1383
    server.configfile = NULL;
1384
    server.hz = REDIS_DEFAULT_HZ;
A
antirez 已提交
1385
    server.runid[REDIS_RUN_ID_SIZE] = '\0';
1386
    server.arch_bits = (sizeof(long) == 8) ? 64 : 32;
1387
    server.port = REDIS_SERVERPORT;
1388
    server.tcp_backlog = REDIS_TCP_BACKLOG;
A
antirez 已提交
1389
    server.bindaddr_count = 0;
1390
    server.unixsocket = NULL;
1391
    server.unixsocketperm = REDIS_DEFAULT_UNIX_SOCKET_PERM;
1392
    server.ipfd_count = 0;
1393 1394
    server.sofd = -1;
    server.dbnum = REDIS_DEFAULT_DBNUM;
1395
    server.verbosity = REDIS_DEFAULT_VERBOSITY;
1396
    server.maxidletime = REDIS_MAXIDLETIME;
1397
    server.tcpkeepalive = REDIS_DEFAULT_TCP_KEEPALIVE;
A
antirez 已提交
1398
    server.active_expire_enabled = 1;
1399
    server.client_max_querybuf_len = REDIS_MAX_QUERYBUF_LEN;
1400
    server.saveparams = NULL;
1401
    server.loading = 0;
1402 1403
    server.logfile = zstrdup(REDIS_DEFAULT_LOGFILE);
    server.syslog_enabled = REDIS_DEFAULT_SYSLOG_ENABLED;
1404
    server.syslog_ident = zstrdup(REDIS_DEFAULT_SYSLOG_IDENT);
J
Jonah H. Harris 已提交
1405
    server.syslog_facility = LOG_LOCAL0;
1406
    server.daemonize = REDIS_DEFAULT_DAEMONIZE;
1407
    server.aof_state = REDIS_AOF_OFF;
1408 1409
    server.aof_fsync = REDIS_DEFAULT_AOF_FSYNC;
    server.aof_no_fsync_on_rewrite = REDIS_DEFAULT_AOF_NO_FSYNC_ON_REWRITE;
1410 1411 1412 1413
    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
1414
    server.aof_last_fsync = time(NULL);
1415 1416
    server.aof_rewrite_time_last = -1;
    server.aof_rewrite_time_start = -1;
1417
    server.aof_lastbgrewrite_status = REDIS_OK;
1418
    server.aof_delayed_fsync = 0;
A
antirez 已提交
1419 1420
    server.aof_fd = -1;
    server.aof_selected_db = -1; /* Make sure the first time will not match */
1421
    server.aof_flush_postponed_start = 0;
1422
    server.aof_rewrite_incremental_fsync = REDIS_DEFAULT_AOF_REWRITE_INCREMENTAL_FSYNC;
A
antirez 已提交
1423
    server.aof_load_truncated = REDIS_DEFAULT_AOF_LOAD_TRUNCATED;
1424
    server.pidfile = zstrdup(REDIS_DEFAULT_PID_FILE);
1425
    server.rdb_filename = zstrdup(REDIS_DEFAULT_RDB_FILENAME);
1426
    server.aof_filename = zstrdup(REDIS_DEFAULT_AOF_FILENAME);
1427
    server.requirepass = NULL;
1428 1429 1430 1431
    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
1432
    server.notify_keyspace_events = 0;
1433
    server.maxclients = REDIS_MAX_CLIENTS;
1434
    server.bpop_blocked_clients = 0;
1435 1436 1437
    server.maxmemory = REDIS_DEFAULT_MAXMEMORY;
    server.maxmemory_policy = REDIS_DEFAULT_MAXMEMORY_POLICY;
    server.maxmemory_samples = REDIS_DEFAULT_MAXMEMORY_SAMPLES;
1438 1439
    server.hash_max_ziplist_entries = REDIS_HASH_MAX_ZIPLIST_ENTRIES;
    server.hash_max_ziplist_value = REDIS_HASH_MAX_ZIPLIST_VALUE;
1440 1441
    server.list_max_ziplist_entries = REDIS_LIST_MAX_ZIPLIST_ENTRIES;
    server.list_max_ziplist_value = REDIS_LIST_MAX_ZIPLIST_VALUE;
1442
    server.set_max_intset_entries = REDIS_SET_MAX_INTSET_ENTRIES;
1443 1444
    server.zset_max_ziplist_entries = REDIS_ZSET_MAX_ZIPLIST_ENTRIES;
    server.zset_max_ziplist_value = REDIS_ZSET_MAX_ZIPLIST_VALUE;
1445
    server.hll_sparse_max_bytes = REDIS_DEFAULT_HLL_SPARSE_MAX_BYTES;
1446
    server.shutdown_asap = 0;
A
1fe4cd5  
antirez 已提交
1447 1448
    server.repl_ping_slave_period = REDIS_REPL_PING_SLAVE_PERIOD;
    server.repl_timeout = REDIS_REPL_TIMEOUT;
1449 1450
    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
1451
    server.cluster_enabled = 0;
1452
    server.cluster_node_timeout = REDIS_CLUSTER_DEFAULT_NODE_TIMEOUT;
1453
    server.cluster_migration_barrier = REDIS_CLUSTER_DEFAULT_MIGRATION_BARRIER;
1454
    server.cluster_slave_validity_factor = REDIS_CLUSTER_DEFAULT_SLAVE_VALIDITY;
1455
    server.cluster_require_full_coverage = REDIS_CLUSTER_DEFAULT_REQUIRE_FULL_COVERAGE;
1456
    server.cluster_configfile = zstrdup(REDIS_DEFAULT_CLUSTER_CONFIG_FILE);
1457
    server.lua_caller = NULL;
A
antirez 已提交
1458
    server.lua_time_limit = REDIS_LUA_TIME_LIMIT;
1459
    server.lua_client = NULL;
1460
    server.lua_timedout = 0;
A
antirez 已提交
1461
    server.migrate_cached_sockets = dictCreate(&migrateCacheDictType,NULL);
1462
    server.next_client_id = 1; /* Client IDs, start from 1 .*/
1463
    server.loading_process_events_interval_bytes = (1024*1024*2);
1464

1465
    server.lruclock = getLRUClock();
1466 1467 1468 1469 1470 1471 1472 1473 1474 1475
    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;
1476 1477
    server.cached_master = NULL;
    server.repl_master_initial_offset = -1;
A
antirez 已提交
1478
    server.repl_state = REDIS_REPL_NONE;
1479
    server.repl_syncio_timeout = REDIS_REPL_SYNCIO_TIMEOUT;
1480 1481
    server.repl_serve_stale_data = REDIS_DEFAULT_SLAVE_SERVE_STALE_DATA;
    server.repl_slave_ro = REDIS_DEFAULT_SLAVE_READ_ONLY;
1482
    server.repl_down_since = 0; /* Never connected, repl is down since EVER. */
1483
    server.repl_disable_tcp_nodelay = REDIS_DEFAULT_REPL_DISABLE_TCP_NODELAY;
1484
    server.repl_diskless_sync = REDIS_DEFAULT_REPL_DISKLESS_SYNC;
A
antirez 已提交
1485
    server.repl_diskless_sync_delay = REDIS_DEFAULT_REPL_DISKLESS_SYNC_DELAY;
1486
    server.slave_priority = REDIS_DEFAULT_SLAVE_PRIORITY;
1487 1488 1489 1490 1491 1492 1493 1494 1495 1496
    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);
1497

1498
    /* Client output buffer limits */
A
antirez 已提交
1499
    for (j = 0; j < REDIS_CLIENT_TYPE_COUNT; j++)
1500
        server.client_obuf_limits[j] = clientBufferLimitsDefaults[j];
1501

1502 1503 1504 1505 1506
    /* 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;
1507

G
guiquanz 已提交
1508
    /* Command table -- we initiialize it here as it is part of the
1509 1510 1511
     * initial configuration, since command names may be changed via
     * redis.conf using the rename-command directive. */
    server.commands = dictCreate(&commandTableDictType,NULL);
1512
    server.orig_commands = dictCreate(&commandTableDictType,NULL);
1513 1514 1515
    populateCommandTable();
    server.delCommand = lookupCommandByCString("del");
    server.multiCommand = lookupCommandByCString("multi");
1516
    server.lpushCommand = lookupCommandByCString("lpush");
1517 1518
    server.lpopCommand = lookupCommandByCString("lpop");
    server.rpopCommand = lookupCommandByCString("rpop");
1519

1520 1521 1522
    /* 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
1523

1524 1525 1526
    /* Latency monitor */
    server.latency_monitor_threshold = REDIS_DEFAULT_LATENCY_MONITOR_THRESHOLD;

A
antirez 已提交
1527
    /* Debugging */
A
antirez 已提交
1528 1529 1530 1531
    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
1532
    server.watchdog_period = 0;
1533 1534
}

1535
/* This function will try to raise the max number of open files accordingly to
1536
 * the configured max number of clients. It also reserves a number of file
1537
 * descriptors (REDIS_MIN_RESERVED_FDS) for extra operations of
1538
 * persistence, listening sockets, log files and so forth.
1539 1540 1541 1542 1543
 *
 * 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) {
1544
    rlim_t maxfiles = server.maxclients+REDIS_MIN_RESERVED_FDS;
1545 1546 1547 1548 1549
    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));
1550
        server.maxclients = 1024-REDIS_MIN_RESERVED_FDS;
1551 1552 1553 1554 1555 1556
    } 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) {
1557
            rlim_t f;
A
antirez 已提交
1558 1559 1560 1561
            int setrlimit_error = 0;

            /* Try to set the file limit to match 'maxfiles' or at least
             * to the higher value supported less than maxfiles. */
1562 1563
            f = maxfiles;
            while(f > oldlimit) {
1564
                rlim_t decr_step = 16;
A
antirez 已提交
1565

1566 1567 1568
                limit.rlim_cur = f;
                limit.rlim_max = f;
                if (setrlimit(RLIMIT_NOFILE,&limit) != -1) break;
A
antirez 已提交
1569 1570 1571 1572 1573 1574
                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;
1575
            }
A
antirez 已提交
1576 1577 1578

            /* Assume that the limit we get initially is still valid if
             * our last try was even lower. */
1579
            if (f < oldlimit) f = oldlimit;
A
antirez 已提交
1580

1581
            if (f != maxfiles) {
1582
                int old_maxclients = server.maxclients;
1583
                server.maxclients = f-REDIS_MIN_RESERVED_FDS;
1584 1585 1586 1587
                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 "
A
antirez 已提交
1588 1589 1590
                        "%llu. Exiting.",
                        (unsigned long long) oldlimit,
                        (unsigned long long) maxfiles);
1591 1592 1593 1594
                    exit(1);
                }
                redisLog(REDIS_WARNING,"You requested maxclients of %d "
                    "requiring at least %llu max file descriptors.",
A
antirez 已提交
1595 1596
                    old_maxclients,
                    (unsigned long long) maxfiles);
1597 1598
                redisLog(REDIS_WARNING,"Redis can't set maximum open files "
                    "to %llu because of OS error: %s.",
A
antirez 已提交
1599
                    (unsigned long long) maxfiles, strerror(setrlimit_error));
1600 1601 1602 1603
                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'.",
A
antirez 已提交
1604
                    (unsigned long long) oldlimit, server.maxclients);
1605
            } else {
1606 1607
                redisLog(REDIS_NOTICE,"Increased maximum number of open files "
                    "to %llu (it was originally set to %llu).",
A
antirez 已提交
1608 1609
                    (unsigned long long) maxfiles,
                    (unsigned long long) oldlimit);
1610 1611 1612 1613 1614
            }
        }
    }
}

1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642
/* 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. */
1643 1644
            fds[*count] = anetTcp6Server(server.neterr,port,NULL,
                server.tcp_backlog);
A
antirez 已提交
1645 1646 1647 1648
            if (fds[*count] != ANET_ERR) {
                anetNonBlock(NULL,fds[*count]);
                (*count)++;
            }
1649 1650
            fds[*count] = anetTcpServer(server.neterr,port,NULL,
                server.tcp_backlog);
A
antirez 已提交
1651 1652 1653 1654
            if (fds[*count] != ANET_ERR) {
                anetNonBlock(NULL,fds[*count]);
                (*count)++;
            }
1655 1656 1657 1658 1659 1660
            /* 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. */
1661 1662
            fds[*count] = anetTcp6Server(server.neterr,port,server.bindaddr[j],
                server.tcp_backlog);
1663 1664
        } else {
            /* Bind IPv4 address. */
1665 1666
            fds[*count] = anetTcpServer(server.neterr,port,server.bindaddr[j],
                server.tcp_backlog);
1667 1668 1669 1670 1671
        }
        if (fds[*count] == ANET_ERR) {
            redisLog(REDIS_WARNING,
                "Creating Server TCP listening socket %s:%d: %s",
                server.bindaddr[j] ? server.bindaddr[j] : "*",
1672
                port, server.neterr);
1673 1674
            return REDIS_ERR;
        }
A
antirez 已提交
1675
        anetNonBlock(NULL,fds[*count]);
1676 1677 1678 1679 1680
        (*count)++;
    }
    return REDIS_OK;
}

1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691
/* 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;
1692
    server.stat_fork_rate = 0;
1693 1694 1695 1696 1697 1698 1699 1700 1701 1702
    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;
}

1703
void initServer(void) {
1704 1705 1706 1707
    int j;

    signal(SIGHUP, SIG_IGN);
    signal(SIGPIPE, SIG_IGN);
1708
    setupSignalHandlers();
1709

J
Jonah H. Harris 已提交
1710 1711 1712 1713 1714
    if (server.syslog_enabled) {
        openlog(server.syslog_ident, LOG_PID | LOG_NDELAY | LOG_NOWAIT,
            server.syslog_facility);
    }

A
antirez 已提交
1715
    server.pid = getpid();
1716
    server.current_client = NULL;
1717
    server.clients = listCreate();
1718
    server.clients_to_close = listCreate();
1719 1720
    server.slaves = listCreate();
    server.monitors = listCreate();
1721
    server.slaveseldb = -1; /* Force to emit the first SELECT command. */
1722
    server.unblocked_clients = listCreate();
1723
    server.ready_keys = listCreate();
1724 1725
    server.clients_waiting_acks = listCreate();
    server.get_ack_from_slaves = 0;
1726
    server.clients_paused = 0;
1727

1728
    createSharedObjects();
1729
    adjustOpenFilesLimit();
A
antirez 已提交
1730
    server.el = aeCreateEventLoop(server.maxclients+REDIS_EVENTLOOP_FDSET_INCR);
1731
    server.db = zmalloc(sizeof(redisDb)*server.dbnum);
1732

1733
    /* Open the TCP listening socket for the user commands. */
1734 1735
    if (server.port != 0 &&
        listenToPort(server.port,server.ipfd,&server.ipfd_count) == REDIS_ERR)
1736
        exit(1);
1737 1738

    /* Open the listening Unix domain socket. */
1739 1740
    if (server.unixsocket != NULL) {
        unlink(server.unixsocket); /* don't care if this fails */
1741 1742
        server.sofd = anetUnixServer(server.neterr,server.unixsocket,
            server.unixsocketperm, server.tcp_backlog);
1743 1744 1745 1746
        if (server.sofd == ANET_ERR) {
            redisLog(REDIS_WARNING, "Opening socket: %s", server.neterr);
            exit(1);
        }
A
antirez 已提交
1747
        anetNonBlock(NULL,server.sofd);
1748
    }
1749 1750 1751

    /* Abort if there are no listening sockets at all. */
    if (server.ipfd_count == 0 && server.sofd < 0) {
1752
        redisLog(REDIS_WARNING, "Configured to not listen anywhere, exiting.");
1753 1754
        exit(1);
    }
1755 1756

    /* Create the Redis databases, and initialize other internal state. */
1757 1758 1759 1760
    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);
1761
        server.db[j].ready_keys = dictCreate(&setDictType,NULL);
1762
        server.db[j].watched_keys = dictCreate(&keylistDictType,NULL);
A
antirez 已提交
1763
        server.db[j].eviction_pool = evictionPoolAlloc();
1764
        server.db[j].id = j;
1765
        server.db[j].avg_ttl = 0;
1766 1767 1768 1769 1770 1771
    }
    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
1772
    server.rdb_child_pid = -1;
A
antirez 已提交
1773
    server.aof_child_pid = -1;
A
antirez 已提交
1774
    server.rdb_child_type = REDIS_RDB_CHILD_TYPE_NONE;
1775
    aofRewriteBufferReset();
A
antirez 已提交
1776
    server.aof_buf = sdsempty();
1777 1778
    server.lastsave = time(NULL); /* At startup we consider the DB saved. */
    server.lastbgsave_try = 0;    /* At startup we never tried to BGSAVE. */
1779 1780
    server.rdb_save_time_last = -1;
    server.rdb_save_time_start = -1;
1781
    server.dirty = 0;
1782 1783
    resetServerStats();
    /* A few stats we don't want to reset: server startup time, and peak mem. */
1784
    server.stat_starttime = time(NULL);
1785
    server.stat_peak_memory = 0;
A
antirez 已提交
1786
    server.resident_set_size = 0;
1787
    server.lastbgsave_status = REDIS_OK;
1788 1789
    server.aof_last_write_status = REDIS_OK;
    server.aof_last_write_errno = 0;
1790
    server.repl_good_slaves_count = 0;
1791
    updateCachedTime();
1792 1793 1794

    /* Create the serverCron() time event, that's our main way to process
     * background operations. */
stamhe's avatar
stamhe 已提交
1795
    if(aeCreateTimeEvent(server.el, 1, serverCron, NULL, NULL) == AE_ERR) {
1796
        redisPanic("Can't create the serverCron time event.");
stamhe's avatar
stamhe 已提交
1797 1798
        exit(1);
    }
1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809

    /* 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.");
            }
    }
1810
    if (server.sofd > 0 && aeCreateFileEvent(server.el,server.sofd,AE_READABLE,
A
antirez 已提交
1811
        acceptUnixHandler,NULL) == AE_ERR) redisPanic("Unrecoverable error creating server.sofd file event.");
1812

1813
    /* Open the AOF file if needed. */
1814
    if (server.aof_state == REDIS_AOF_ON) {
A
antirez 已提交
1815
        server.aof_fd = open(server.aof_filename,
1816
                               O_WRONLY|O_APPEND|O_CREAT,0644);
A
antirez 已提交
1817
        if (server.aof_fd == -1) {
1818 1819 1820 1821 1822 1823
            redisLog(REDIS_WARNING, "Can't open the append-only file: %s",
                strerror(errno));
            exit(1);
        }
    }

1824 1825
    /* 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
1826 1827
     * at 3 GB using maxmemory with 'noeviction' policy'. This avoids
     * useless crashes of the Redis instance for out of memory. */
1828
    if (server.arch_bits == 32 && server.maxmemory == 0) {
1829 1830
        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 */
1831 1832 1833
        server.maxmemory_policy = REDIS_MAXMEMORY_NO_EVICTION;
    }

A
antirez 已提交
1834
    if (server.cluster_enabled) clusterInit();
1835
    replicationScriptCacheInit();
1836
    scriptingInit();
1837
    slowlogInit();
1838
    latencyMonitorInit();
1839
    bioInit();
1840 1841
}

1842 1843 1844 1845
/* Populates the Redis Command Table starting from the hard coded list
 * we have on top of redis.c file. */
void populateCommandTable(void) {
    int j;
1846
    int numcommands = sizeof(redisCommandTable)/sizeof(struct redisCommand);
1847 1848

    for (j = 0; j < numcommands; j++) {
1849
        struct redisCommand *c = redisCommandTable+j;
1850
        char *f = c->sflags;
1851
        int retval1, retval2;
1852

1853 1854 1855 1856 1857 1858 1859
        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;
1860 1861
            case 's': c->flags |= REDIS_CMD_NOSCRIPT; break;
            case 'R': c->flags |= REDIS_CMD_RANDOM; break;
1862
            case 'S': c->flags |= REDIS_CMD_SORT_FOR_SCRIPT; break;
1863 1864
            case 'l': c->flags |= REDIS_CMD_LOADING; break;
            case 't': c->flags |= REDIS_CMD_STALE; break;
1865
            case 'M': c->flags |= REDIS_CMD_SKIP_MONITOR; break;
1866
            case 'k': c->flags |= REDIS_CMD_ASKING; break;
1867
            case 'F': c->flags |= REDIS_CMD_FAST; break;
1868 1869 1870 1871 1872
            default: redisPanic("Unsupported command flag"); break;
            }
            f++;
        }

1873 1874 1875 1876 1877
        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);
1878
    }
1879 1880
}

1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892
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;
    }
}

1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929
/* ========================== 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);
}

1930 1931
/* ====================== Commands lookup and execution ===================== */

1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942
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;
1943 1944
}

1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958
/* 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
1959
/* Propagate the specified command (in the context of the specified database id)
1960
 * to AOF and Slaves.
A
antirez 已提交
1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971
 *
 * 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);
1972
    if (flags & REDIS_PROPAGATE_REPL)
A
antirez 已提交
1973 1974 1975
        replicationFeedSlaves(server.slaves,dbid,argv,argc);
}

1976 1977
/* Used inside commands to schedule the propagation of additional commands
 * after the current command is propagated to AOF / Replication. */
1978 1979 1980
void alsoPropagate(struct redisCommand *cmd, int dbid, robj **argv, int argc,
                   int target)
{
1981
    redisOpArrayAppend(&server.also_propagate,cmd,dbid,argv,argc,target);
1982 1983
}

A
antirez 已提交
1984
/* It is possible to call the function forceCommandPropagation() inside a
J
Juarez Bochi 已提交
1985
 * Redis command implementation in order to to force the propagation of a
A
antirez 已提交
1986 1987 1988 1989 1990 1991
 * 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;
}

1992
/* Call() is the core of Redis execution of a command */
1993
void call(redisClient *c, int flags) {
1994
    long long dirty, start, duration;
A
antirez 已提交
1995
    int client_old_flags = c->flags;
1996

A
antirez 已提交
1997
    /* Sent the command to clients in MONITOR mode, only if the commands are
G
guiquanz 已提交
1998
     * not generated from reading an AOF. */
1999 2000 2001 2002
    if (listLength(server.monitors) &&
        !server.loading &&
        !(c->cmd->flags & REDIS_CMD_SKIP_MONITOR))
    {
2003
        replicationFeedMonitors(c,server.monitors,c->db->id,c->argv,c->argc);
2004
    }
A
antirez 已提交
2005 2006

    /* Call the command. */
A
antirez 已提交
2007
    c->flags &= ~(REDIS_FORCE_AOF|REDIS_FORCE_REPL);
2008
    redisOpArrayInit(&server.also_propagate);
2009
    dirty = server.dirty;
2010
    start = ustime();
2011
    c->cmd->proc(c);
2012
    duration = ustime()-start;
2013
    dirty = server.dirty-dirty;
A
antirez 已提交
2014
    if (dirty < 0) dirty = 0;
2015 2016 2017 2018 2019 2020

    /* 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
2021 2022 2023 2024 2025 2026 2027 2028 2029 2030
    /* 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
2031 2032
    /* Log the command into the Slow log if needed, and populate the
     * per-command statistics that we show in INFO commandstats. */
2033 2034 2035
    if (flags & REDIS_CALL_SLOWLOG && c->cmd->proc != execCommand) {
        char *latency_event = (c->cmd->flags & REDIS_CMD_FAST) ?
                              "fast-command" : "command";
2036
        latencyAddSampleIfNeeded(latency_event,duration/1000);
2037
        slowlogPushEntryIfNeeded(c->argv,c->argc,duration);
2038
    }
2039 2040 2041 2042
    if (flags & REDIS_CALL_STATS) {
        c->cmd->microseconds += duration;
        c->cmd->calls++;
    }
A
antirez 已提交
2043 2044

    /* Propagate the command into the AOF and replication link */
2045
    if (flags & REDIS_CALL_PROPAGATE) {
A
antirez 已提交
2046 2047
        int flags = REDIS_PROPAGATE_NONE;

A
antirez 已提交
2048 2049
        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
2050 2051 2052 2053
        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);
2054
    }
2055

A
antirez 已提交
2056 2057 2058 2059 2060
    /* 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);

2061 2062
    /* Handle the alsoPropagate() API to handle commands that want to propagate
     * multiple separated commands. */
2063
    if (server.also_propagate.numops) {
2064
        int j;
2065
        redisOp *rop;
2066

2067 2068 2069 2070 2071
        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);
2072
    }
2073 2074 2075 2076
    server.stat_numcommands++;
}

/* If this function gets called we already read a whole
2077
 * command, arguments are in the client argv/argc fields.
2078 2079 2080 2081
 * 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
2082 2083
 * other operations can be performed by the caller. Otherwise
 * if 0 is returned the client was destroyed (i.e. after QUIT). */
2084
int processCommand(redisClient *c) {
P
Pieter Noordhuis 已提交
2085 2086 2087 2088
    /* 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. */
2089
    if (!strcasecmp(c->argv[0]->ptr,"quit")) {
P
Pieter Noordhuis 已提交
2090
        addReply(c,shared.ok);
2091
        c->flags |= REDIS_CLOSE_AFTER_REPLY;
2092
        return REDIS_ERR;
2093 2094 2095
    }

    /* Now lookup the command and check ASAP about trivial error conditions
2096
     * such as wrong arity, bad command name and so forth. */
2097
    c->cmd = c->lastcmd = lookupCommand(c->argv[0]->ptr);
2098
    if (!c->cmd) {
A
antirez 已提交
2099
        flagTransaction(c);
2100 2101
        addReplyErrorFormat(c,"unknown command '%s'",
            (char*)c->argv[0]->ptr);
2102
        return REDIS_OK;
2103 2104
    } else if ((c->cmd->arity > 0 && c->cmd->arity != c->argc) ||
               (c->argc < -c->cmd->arity)) {
A
antirez 已提交
2105
        flagTransaction(c);
2106
        addReplyErrorFormat(c,"wrong number of arguments for '%s' command",
2107
            c->cmd->name);
2108
        return REDIS_OK;
2109 2110 2111
    }

    /* Check if the user is authenticated */
2112 2113
    if (server.requirepass && !c->authenticated && c->cmd->proc != authCommand)
    {
A
antirez 已提交
2114
        flagTransaction(c);
2115
        addReply(c,shared.noautherr);
2116
        return REDIS_OK;
2117 2118
    }

2119 2120 2121 2122
    /* 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
2123
    if (server.cluster_enabled &&
2124 2125 2126
        !(c->flags & REDIS_MASTER) &&
        !(c->cmd->getkeys_proc == NULL && c->cmd->firstkey == 0))
    {
A
antirez 已提交
2127 2128
        int hashslot;

2129
        if (server.cluster->state != REDIS_CLUSTER_OK) {
2130
            flagTransaction(c);
2131
            addReplySds(c,sdsnew("-CLUSTERDOWN The cluster is down. Use CLUSTER INFO for more information\r\n"));
A
antirez 已提交
2132 2133
            return REDIS_OK;
        } else {
2134 2135
            int error_code;
            clusterNode *n = getNodeByQuery(c,c->cmd,c->argv,c->argc,&hashslot,&error_code);
A
antirez 已提交
2136
            if (n == NULL) {
2137
                flagTransaction(c);
2138 2139 2140 2141 2142 2143 2144 2145 2146 2147
                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
2148
                return REDIS_OK;
2149
            } else if (n != server.cluster->myself) {
2150
                flagTransaction(c);
A
antirez 已提交
2151
                addReplySds(c,sdscatprintf(sdsempty(),
2152 2153
                    "-%s %d %s:%d\r\n",
                    (error_code == REDIS_CLUSTER_REDIR_ASK) ? "ASK" : "MOVED",
2154
                    hashslot,n->ip,n->port));
A
antirez 已提交
2155 2156 2157 2158 2159
                return REDIS_OK;
            }
        }
    }

2160 2161 2162 2163 2164
    /* 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. */
2165 2166 2167
    if (server.maxmemory) {
        int retval = freeMemoryIfNeeded();
        if ((c->cmd->flags & REDIS_CMD_DENYOOM) && retval == REDIS_ERR) {
A
antirez 已提交
2168
            flagTransaction(c);
2169
            addReply(c, shared.oomerr);
2170 2171
            return REDIS_OK;
        }
2172 2173
    }

2174 2175
    /* Don't accept write commands if there are problems persisting on disk
     * and if this is a master instance. */
2176 2177 2178 2179
    if (((server.stop_writes_on_bgsave_err &&
          server.saveparamslen > 0 &&
          server.lastbgsave_status == REDIS_ERR) ||
          server.aof_last_write_status == REDIS_ERR) &&
2180
        server.masterhost == NULL &&
2181 2182
        (c->cmd->flags & REDIS_CMD_WRITE ||
         c->cmd->proc == pingCommand))
2183
    {
A
antirez 已提交
2184
        flagTransaction(c);
2185 2186 2187 2188 2189 2190 2191
        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)));
2192 2193 2194
        return REDIS_OK;
    }

2195
    /* Don't accept write commands if there are not enough good slaves and
2196
     * user configured the min-slaves-to-write option. */
2197 2198
    if (server.masterhost == NULL &&
        server.repl_min_slaves_to_write &&
2199 2200
        server.repl_min_slaves_max_lag &&
        c->cmd->flags & REDIS_CMD_WRITE &&
2201 2202 2203 2204 2205 2206 2207
        server.repl_good_slaves_count < server.repl_min_slaves_to_write)
    {
        flagTransaction(c);
        addReply(c, shared.noreplicaserr);
        return REDIS_OK;
    }

2208
    /* Don't accept write commands if this is a read only slave. But
2209 2210 2211 2212 2213 2214 2215 2216 2217
     * 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;
    }

2218
    /* Only allow SUBSCRIBE and UNSUBSCRIBE in the context of Pub/Sub */
2219 2220
    if (c->flags & REDIS_PUBSUB &&
        c->cmd->proc != pingCommand &&
2221 2222 2223 2224
        c->cmd->proc != subscribeCommand &&
        c->cmd->proc != unsubscribeCommand &&
        c->cmd->proc != psubscribeCommand &&
        c->cmd->proc != punsubscribeCommand) {
2225
        addReplyError(c,"only (P)SUBSCRIBE / (P)UNSUBSCRIBE / QUIT allowed in this context");
2226
        return REDIS_OK;
2227 2228
    }

2229 2230
    /* 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
2231
    if (server.masterhost && server.repl_state != REDIS_REPL_CONNECTED &&
2232
        server.repl_serve_stale_data == 0 &&
2233
        !(c->cmd->flags & REDIS_CMD_STALE))
2234
    {
A
antirez 已提交
2235
        flagTransaction(c);
2236
        addReply(c, shared.masterdownerr);
2237 2238 2239
        return REDIS_OK;
    }

2240 2241 2242
    /* Loading DB? Return an error if the command has not the
     * REDIS_CMD_LOADING flag. */
    if (server.loading && !(c->cmd->flags & REDIS_CMD_LOADING)) {
2243 2244 2245 2246
        addReply(c, shared.loadingerr);
        return REDIS_OK;
    }

A
antirez 已提交
2247
    /* Lua script too slow? Only allow a limited number of commands. */
2248
    if (server.lua_timedout &&
2249
          c->cmd->proc != authCommand &&
A
antirez 已提交
2250
          c->cmd->proc != replconfCommand &&
2251
        !(c->cmd->proc == shutdownCommand &&
2252 2253 2254 2255 2256 2257
          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
2258
        flagTransaction(c);
2259 2260 2261 2262
        addReply(c, shared.slowscripterr);
        return REDIS_OK;
    }

2263 2264
    /* Exec the command */
    if (c->flags & REDIS_MULTI &&
2265 2266
        c->cmd->proc != execCommand && c->cmd->proc != discardCommand &&
        c->cmd->proc != multiCommand && c->cmd->proc != watchCommand)
2267
    {
2268
        queueMultiCommand(c);
2269 2270
        addReply(c,shared.queued);
    } else {
2271
        call(c,REDIS_CALL_FULL);
2272
        c->woff = server.master_repl_offset;
2273 2274
        if (listLength(server.ready_keys))
            handleClientsBlockedOnLists();
2275
    }
2276
    return REDIS_OK;
2277 2278 2279 2280
}

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

2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295
/* 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 */
    }
}

2296 2297 2298 2299
int prepareForShutdown(int flags) {
    int save = flags & REDIS_SHUTDOWN_SAVE;
    int nosave = flags & REDIS_SHUTDOWN_NOSAVE;

2300
    redisLog(REDIS_WARNING,"User requested shutdown...");
2301 2302 2303
    /* 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
2304
    if (server.rdb_child_pid != -1) {
2305
        redisLog(REDIS_WARNING,"There is a child saving an .rdb. Killing it!");
2306
        kill(server.rdb_child_pid,SIGUSR1);
A
antirez 已提交
2307
        rdbRemoveTempFile(server.rdb_child_pid);
2308
    }
2309
    if (server.aof_state != REDIS_AOF_OFF) {
2310 2311
        /* 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
2312
        if (server.aof_child_pid != -1) {
2313 2314 2315 2316 2317 2318
            /* If we have AOF enabled but haven't written the AOF yet, don't
             * shutdown or else the dataset will be lost. */
            if (server.aof_state == REDIS_AOF_WAIT_REWRITE) {
                redisLog(REDIS_WARNING, "Writing initial AOF, can't exit.");
                return REDIS_ERR;
            }
2319 2320
            redisLog(REDIS_WARNING,
                "There is a child rewriting the AOF. Killing it!");
2321
            kill(server.aof_child_pid,SIGUSR1);
2322
        }
2323
        /* Append only file: fsync() the AOF and exit */
2324
        redisLog(REDIS_NOTICE,"Calling fsync() on the AOF file.");
A
antirez 已提交
2325
        aof_fsync(server.aof_fd);
2326
    }
2327
    if ((server.saveparamslen > 0 && !nosave) || save) {
2328
        redisLog(REDIS_NOTICE,"Saving the final RDB snapshot before exiting.");
2329
        /* Snapshotting. Perform a SYNC SAVE and exit */
A
antirez 已提交
2330
        if (rdbSave(server.rdb_filename) != REDIS_OK) {
2331 2332 2333 2334 2335
            /* 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... */
2336
            redisLog(REDIS_WARNING,"Error trying to save the DB, can't exit.");
2337 2338 2339
            return REDIS_ERR;
        }
    }
2340 2341 2342 2343
    if (server.daemonize) {
        redisLog(REDIS_NOTICE,"Removing the pid file.");
        unlink(server.pidfile);
    }
2344
    /* Close the listening sockets. Apparently this allows faster restarts. */
2345
    closeListeningSockets(1);
2346 2347
    redisLog(REDIS_WARNING,"%s is now ready to exit, bye bye...",
        server.sentinel_mode ? "Sentinel" : "Redis");
2348 2349 2350 2351 2352
    return REDIS_OK;
}

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

2353
/* Return zero if strings are the same, non-zero if they are not.
2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367
 * 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. */
2368 2369 2370
    unsigned int alen = strlen(a);
    unsigned int blen = strlen(b);
    unsigned int j;
2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394
    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. */
}

2395
void authCommand(redisClient *c) {
2396 2397
    if (!server.requirepass) {
        addReplyError(c,"Client sent AUTH, but no password is set");
2398
    } else if (!time_independent_strcmp(c->argv[1]->ptr, server.requirepass)) {
2399 2400 2401 2402
      c->authenticated = 1;
      addReply(c,shared.ok);
    } else {
      c->authenticated = 0;
2403
      addReplyError(c,"invalid password");
2404 2405 2406
    }
}

2407 2408
/* The PING command. It works in a different way if the client is in
 * in Pub/Sub mode. */
2409
void pingCommand(redisClient *c) {
2410 2411
    /* The command takes zero or one arguments. */
    if (c->argc > 2) {
2412 2413
        addReplyErrorFormat(c,"wrong number of arguments for '%s' command",
            c->cmd->name);
2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429
        return;
    }

    if (c->flags & REDIS_PUBSUB) {
        addReply(c,shared.mbulkhdr[2]);
        addReplyBulkCBuffer(c,"pong",4);
        if (c->argc == 1)
            addReplyBulkCBuffer(c,"",0);
        else
            addReplyBulk(c,c->argv[1]);
    } else {
        if (c->argc == 1)
            addReply(c,shared.pong);
        else
            addReplyBulk(c,c->argv[1]);
    }
2430 2431 2432 2433 2434 2435
}

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

A
antirez 已提交
2436 2437 2438
void timeCommand(redisClient *c) {
    struct timeval tv;

G
guiquanz 已提交
2439
    /* gettimeofday() can only fail if &tv is a bad address so we
A
antirez 已提交
2440 2441 2442 2443 2444 2445 2446
     * don't check for errors. */
    gettimeofday(&tv,NULL);
    addReplyMultiBulkLen(c,2);
    addReplyBulkLongLong(c,tv.tv_sec);
    addReplyBulkLongLong(c,tv.tv_usec);
}

M
Matt Stancliff 已提交
2447

A
antirez 已提交
2448
/* Helper function for addReplyCommand() to output flags. */
2449
int addReplyCommandFlag(redisClient *c, struct redisCommand *cmd, int f, char *reply) {
M
Matt Stancliff 已提交
2450 2451 2452 2453 2454 2455
    if (cmd->flags & f) {
        addReplyStatus(c, reply);
        return 1;
    }
    return 0;
}
2456

A
antirez 已提交
2457
/* Output the representation of a Redis command. Used by the COMMAND command. */
2458
void addReplyCommand(redisClient *c, struct redisCommand *cmd) {
M
Matt Stancliff 已提交
2459 2460 2461 2462 2463 2464 2465 2466 2467 2468
    if (!cmd) {
        addReply(c, shared.nullbulk);
    } else {
        /* We are adding: command name, arg count, flags, first, last, offset */
        addReplyMultiBulkLen(c, 6);
        addReplyBulkCString(c, cmd->name);
        addReplyLongLong(c, cmd->arity);

        int flagcount = 0;
        void *flaglen = addDeferredMultiBulkLength(c);
2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480
        flagcount += addReplyCommandFlag(c,cmd,REDIS_CMD_WRITE, "write");
        flagcount += addReplyCommandFlag(c,cmd,REDIS_CMD_READONLY, "readonly");
        flagcount += addReplyCommandFlag(c,cmd,REDIS_CMD_DENYOOM, "denyoom");
        flagcount += addReplyCommandFlag(c,cmd,REDIS_CMD_ADMIN, "admin");
        flagcount += addReplyCommandFlag(c,cmd,REDIS_CMD_PUBSUB, "pubsub");
        flagcount += addReplyCommandFlag(c,cmd,REDIS_CMD_NOSCRIPT, "noscript");
        flagcount += addReplyCommandFlag(c,cmd,REDIS_CMD_RANDOM, "random");
        flagcount += addReplyCommandFlag(c,cmd,REDIS_CMD_SORT_FOR_SCRIPT,"sort_for_script");
        flagcount += addReplyCommandFlag(c,cmd,REDIS_CMD_LOADING, "loading");
        flagcount += addReplyCommandFlag(c,cmd,REDIS_CMD_STALE, "stale");
        flagcount += addReplyCommandFlag(c,cmd,REDIS_CMD_SKIP_MONITOR, "skip_monitor");
        flagcount += addReplyCommandFlag(c,cmd,REDIS_CMD_ASKING, "asking");
2481
        flagcount += addReplyCommandFlag(c,cmd,REDIS_CMD_FAST, "fast");
M
Matt Stancliff 已提交
2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492
        if (cmd->getkeys_proc) {
            addReplyStatus(c, "movablekeys");
            flagcount += 1;
        }
        setDeferredMultiBulkLength(c, flaglen, flagcount);

        addReplyLongLong(c, cmd->firstkey);
        addReplyLongLong(c, cmd->lastkey);
        addReplyLongLong(c, cmd->keystep);
    }
}
2493

A
antirez 已提交
2494
/* COMMAND <subcommand> <args> */
A
antirez 已提交
2495
void commandCommand(redisClient *c) {
M
Matt Stancliff 已提交
2496 2497 2498
    dictIterator *di;
    dictEntry *de;

A
antirez 已提交
2499 2500 2501 2502 2503 2504 2505 2506
    if (c->argc == 1) {
        addReplyMultiBulkLen(c, dictSize(server.commands));
        di = dictGetIterator(server.commands);
        while ((de = dictNext(di)) != NULL) {
            addReplyCommand(c, dictGetVal(de));
        }
        dictReleaseIterator(di);
    } else if (!strcasecmp(c->argv[1]->ptr, "info")) {
M
Matt Stancliff 已提交
2507 2508 2509
        int i;
        addReplyMultiBulkLen(c, c->argc-2);
        for (i = 2; i < c->argc; i++) {
2510
            addReplyCommand(c, dictFetchValue(server.commands, c->argv[i]->ptr));
M
Matt Stancliff 已提交
2511
        }
A
antirez 已提交
2512 2513
    } else if (!strcasecmp(c->argv[1]->ptr, "count") && c->argc == 2) {
        addReplyLongLong(c, dictSize(server.commands));
2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531
    } else if (!strcasecmp(c->argv[1]->ptr,"getkeys") && c->argc >= 3) {
        struct redisCommand *cmd = lookupCommand(c->argv[2]->ptr);
        int *keys, numkeys, j;

        if (!cmd) {
            addReplyErrorFormat(c,"Invalid command specified");
            return;
        } else if ((cmd->arity > 0 && cmd->arity != c->argc-2) ||
                   ((c->argc-2) < -cmd->arity))
        {
            addReplyError(c,"Invalid number of arguments specified for command");
            return;
        }

        keys = getKeysFromCommand(cmd,c->argv+2,c->argc-2,&numkeys);
        addReplyMultiBulkLen(c,numkeys);
        for (j = 0; j < numkeys; j++) addReplyBulk(c,c->argv[keys[j]+2]);
        getKeysFreeResult(keys);
A
antirez 已提交
2532
    } else {
A
antirez 已提交
2533
        addReplyError(c, "Unknown subcommand or wrong number of arguments.");
M
Matt Stancliff 已提交
2534 2535 2536 2537
        return;
    }
}

2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555
/* 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);
2556 2557 2558 2559 2560 2561 2562 2563 2564
    } else if (n < (1024LL*1024*1024*1024*1024)) {
        d = (double)n/(1024LL*1024*1024*1024);
        sprintf(s,"%.2fT",d);
    } else if (n < (1024LL*1024*1024*1024*1024*1024)) {
        d = (double)n/(1024LL*1024*1024*1024*1024);
        sprintf(s,"%.2fP",d);
    } else {
        /* Let's hope we never need this */
        sprintf(s,"%lluB",n);
2565 2566 2567 2568 2569 2570
    }
}

/* 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. */
2571 2572
sds genRedisInfoString(char *section) {
    sds info = sdsempty();
2573
    time_t uptime = server.unixtime-server.stat_starttime;
A
antirez 已提交
2574
    int j, numcommands;
2575
    struct rusage self_ru, c_ru;
2576
    unsigned long lol, bib;
2577 2578
    int allsections = 0, defsections = 0;
    int sections = 0;
2579

2580 2581 2582
    if (section == NULL) section = "default";
    allsections = strcasecmp(section,"all") == 0;
    defsections = strcasecmp(section,"default") == 0;
2583 2584 2585

    getrusage(RUSAGE_SELF, &self_ru);
    getrusage(RUSAGE_CHILDREN, &c_ru);
2586
    getClientsMaxBuffers(&lol,&bib);
2587 2588 2589

    /* Server */
    if (allsections || defsections || !strcasecmp(section,"server")) {
2590 2591
        static int call_uname = 1;
        static struct utsname name;
2592
        char *mode;
2593

2594 2595 2596
        if (server.cluster_enabled) mode = "cluster";
        else if (server.sentinel_mode) mode = "sentinel";
        else mode = "standalone";
2597

2598
        if (sections++) info = sdscat(info,"\r\n");
2599 2600 2601 2602 2603 2604 2605

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

2606
        info = sdscatprintf(info,
2607 2608 2609 2610
            "# Server\r\n"
            "redis_version:%s\r\n"
            "redis_git_sha1:%s\r\n"
            "redis_git_dirty:%d\r\n"
2611
            "redis_build_id:%llx\r\n"
2612
            "redis_mode:%s\r\n"
2613
            "os:%s %s %s\r\n"
2614
            "arch_bits:%d\r\n"
2615
            "multiplexing_api:%s\r\n"
A
antirez 已提交
2616
            "gcc_version:%d.%d.%d\r\n"
2617
            "process_id:%ld\r\n"
A
antirez 已提交
2618
            "run_id:%s\r\n"
2619
            "tcp_port:%d\r\n"
Y
YAMAMOTO Takashi 已提交
2620 2621
            "uptime_in_seconds:%jd\r\n"
            "uptime_in_days:%jd\r\n"
2622
            "hz:%d\r\n"
2623 2624
            "lru_clock:%ld\r\n"
            "config_file:%s\r\n",
2625 2626 2627
            REDIS_VERSION,
            redisGitSHA1(),
            strtol(redisGitDirty(),NULL,10) > 0,
2628
            (unsigned long long) redisBuildId(),
2629
            mode,
2630
            name.sysname, name.release, name.machine,
2631
            server.arch_bits,
2632
            aeGetApiName(),
A
antirez 已提交
2633 2634 2635 2636 2637
#ifdef __GNUC__
            __GNUC__,__GNUC_MINOR__,__GNUC_PATCHLEVEL__,
#else
            0,0,0,
#endif
2638
            (long) getpid(),
A
antirez 已提交
2639
            server.runid,
2640
            server.port,
Y
YAMAMOTO Takashi 已提交
2641 2642
            (intmax_t)uptime,
            (intmax_t)(uptime/(3600*24)),
2643
            server.hz,
2644 2645
            (unsigned long) server.lruclock,
            server.configfile ? server.configfile : "");
2646 2647 2648 2649 2650 2651 2652
    }

    /* Clients */
    if (allsections || defsections || !strcasecmp(section,"clients")) {
        if (sections++) info = sdscat(info,"\r\n");
        info = sdscatprintf(info,
            "# Clients\r\n"
2653
            "connected_clients:%lu\r\n"
2654 2655 2656 2657 2658 2659 2660 2661 2662 2663
            "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")) {
2664 2665
        char hmem[64];
        char peak_hmem[64];
2666
        size_t zmalloc_used = zmalloc_used_memory();
2667

2668 2669 2670 2671 2672
        /* 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)
2673 2674 2675
            server.stat_peak_memory = zmalloc_used;

        bytesToHuman(hmem,zmalloc_used);
2676
        bytesToHuman(peak_hmem,server.stat_peak_memory);
2677 2678 2679 2680 2681 2682
        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"
2683 2684
            "used_memory_peak:%zu\r\n"
            "used_memory_peak_human:%s\r\n"
A
antirez 已提交
2685
            "used_memory_lua:%lld\r\n"
2686
            "mem_fragmentation_ratio:%.2f\r\n"
2687
            "mem_allocator:%s\r\n",
2688
            zmalloc_used,
2689
            hmem,
A
antirez 已提交
2690
            server.resident_set_size,
2691 2692
            server.stat_peak_memory,
            peak_hmem,
A
antirez 已提交
2693
            ((long long)lua_gc(server.lua,LUA_GCCOUNT,0))*1024LL,
A
antirez 已提交
2694
            zmalloc_get_fragmentation_ratio(server.resident_set_size),
2695
            ZMALLOC_LIB
2696
            );
2697 2698
    }

2699 2700 2701
    /* Persistence */
    if (allsections || defsections || !strcasecmp(section,"persistence")) {
        if (sections++) info = sdscat(info,"\r\n");
2702
        info = sdscatprintf(info,
2703 2704
            "# Persistence\r\n"
            "loading:%d\r\n"
2705 2706
            "rdb_changes_since_last_save:%lld\r\n"
            "rdb_bgsave_in_progress:%d\r\n"
Y
YAMAMOTO Takashi 已提交
2707
            "rdb_last_save_time:%jd\r\n"
2708
            "rdb_last_bgsave_status:%s\r\n"
Y
YAMAMOTO Takashi 已提交
2709 2710
            "rdb_last_bgsave_time_sec:%jd\r\n"
            "rdb_current_bgsave_time_sec:%jd\r\n"
2711
            "aof_enabled:%d\r\n"
2712 2713
            "aof_rewrite_in_progress:%d\r\n"
            "aof_rewrite_scheduled:%d\r\n"
Y
YAMAMOTO Takashi 已提交
2714 2715
            "aof_last_rewrite_time_sec:%jd\r\n"
            "aof_current_rewrite_time_sec:%jd\r\n"
2716 2717
            "aof_last_bgrewrite_status:%s\r\n"
            "aof_last_write_status:%s\r\n",
2718 2719
            server.loading,
            server.dirty,
A
antirez 已提交
2720
            server.rdb_child_pid != -1,
Y
YAMAMOTO Takashi 已提交
2721
            (intmax_t)server.lastsave,
2722
            (server.lastbgsave_status == REDIS_OK) ? "ok" : "err",
Y
YAMAMOTO Takashi 已提交
2723 2724 2725
            (intmax_t)server.rdb_save_time_last,
            (intmax_t)((server.rdb_child_pid == -1) ?
                -1 : time(NULL)-server.rdb_save_time_start),
2726
            server.aof_state != REDIS_AOF_OFF,
2727
            server.aof_child_pid != -1,
2728
            server.aof_rewrite_scheduled,
Y
YAMAMOTO Takashi 已提交
2729 2730 2731
            (intmax_t)server.aof_rewrite_time_last,
            (intmax_t)((server.aof_child_pid == -1) ?
                -1 : time(NULL)-server.aof_rewrite_time_start),
2732 2733
            (server.aof_lastbgrewrite_status == REDIS_OK) ? "ok" : "err",
            (server.aof_last_write_status == REDIS_OK) ? "ok" : "err");
2734

2735
        if (server.aof_state != REDIS_AOF_OFF) {
2736 2737 2738
            info = sdscatprintf(info,
                "aof_current_size:%lld\r\n"
                "aof_base_size:%lld\r\n"
2739 2740
                "aof_pending_rewrite:%d\r\n"
                "aof_buffer_length:%zu\r\n"
2741
                "aof_rewrite_buffer_length:%lu\r\n"
2742 2743
                "aof_pending_bio_fsync:%llu\r\n"
                "aof_delayed_fsync:%lu\r\n",
2744 2745 2746
                (long long) server.aof_current_size,
                (long long) server.aof_rewrite_base_size,
                server.aof_rewrite_scheduled,
A
antirez 已提交
2747
                sdslen(server.aof_buf),
2748
                aofRewriteBufferSize(),
2749 2750
                bioPendingJobsOfType(REDIS_BIO_AOF_FSYNC),
                server.aof_delayed_fsync);
2751 2752
        }

2753 2754 2755 2756 2757 2758 2759 2760 2761
        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;

2762
            elapsed = server.unixtime-server.loading_start_time;
2763 2764 2765 2766 2767 2768 2769 2770
            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
2771
                "loading_start_time:%jd\r\n"
2772 2773 2774
                "loading_total_bytes:%llu\r\n"
                "loading_loaded_bytes:%llu\r\n"
                "loading_loaded_perc:%.2f\r\n"
Y
YAMAMOTO Takashi 已提交
2775 2776
                "loading_eta_seconds:%jd\r\n",
                (intmax_t) server.loading_start_time,
2777 2778 2779
                (unsigned long long) server.loading_total_bytes,
                (unsigned long long) server.loading_loaded_bytes,
                perc,
Y
YAMAMOTO Takashi 已提交
2780
                (intmax_t)eta
2781 2782
            );
        }
2783
    }
2784 2785 2786 2787

    /* Stats */
    if (allsections || defsections || !strcasecmp(section,"stats")) {
        if (sections++) info = sdscat(info,"\r\n");
2788
        info = sdscatprintf(info,
2789 2790 2791
            "# Stats\r\n"
            "total_connections_received:%lld\r\n"
            "total_commands_processed:%lld\r\n"
2792
            "instantaneous_ops_per_sec:%lld\r\n"
2793
            "rejected_connections:%lld\r\n"
2794 2795 2796
            "sync_full:%lld\r\n"
            "sync_partial_ok:%lld\r\n"
            "sync_partial_err:%lld\r\n"
2797 2798 2799 2800 2801
            "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"
2802
            "pubsub_patterns:%lu\r\n"
2803 2804
            "latest_fork_usec:%lld\r\n"
            "migrate_cached_sockets:%ld\r\n",
2805 2806
            server.stat_numconnections,
            server.stat_numcommands,
2807
            getOperationsPerSecond(),
2808
            server.stat_rejected_conn,
2809 2810 2811
            server.stat_sync_full,
            server.stat_sync_partial_ok,
            server.stat_sync_partial_err,
2812 2813 2814 2815 2816
            server.stat_expiredkeys,
            server.stat_evictedkeys,
            server.stat_keyspace_hits,
            server.stat_keyspace_misses,
            dictSize(server.pubsub_channels),
2817
            listLength(server.pubsub_patterns),
2818 2819
            server.stat_fork_time,
            dictSize(server.migrate_cached_sockets));
2820
    }
A
antirez 已提交
2821

2822 2823 2824 2825 2826 2827 2828 2829
    /* 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) {
2830 2831 2832 2833 2834 2835 2836
            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;

2837 2838 2839 2840 2841 2842
            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"
2843
                "slave_repl_offset:%lld\r\n"
2844 2845
                ,server.masterhost,
                server.masterport,
A
antirez 已提交
2846
                (server.repl_state == REDIS_REPL_CONNECTED) ?
2847 2848
                    "up" : "down",
                server.master ?
2849
                ((int)(server.unixtime-server.master->lastinteraction)) : -1,
2850
                server.repl_state == REDIS_REPL_TRANSFER,
2851
                slave_repl_offset
2852 2853
            );

A
antirez 已提交
2854
            if (server.repl_state == REDIS_REPL_TRANSFER) {
2855
                info = sdscatprintf(info,
2856
                    "master_sync_left_bytes:%lld\r\n"
2857
                    "master_sync_last_io_seconds_ago:%d\r\n"
2858 2859
                    , (long long)
                        (server.repl_transfer_size - server.repl_transfer_read),
2860
                    (int)(server.unixtime-server.repl_transfer_lastio)
2861 2862
                );
            }
2863

A
antirez 已提交
2864
            if (server.repl_state != REDIS_REPL_CONNECTED) {
2865
                info = sdscatprintf(info,
Y
YAMAMOTO Takashi 已提交
2866 2867
                    "master_link_down_since_seconds:%jd\r\n",
                    (intmax_t)server.unixtime-server.repl_down_since);
2868
            }
2869
            info = sdscatprintf(info,
2870 2871 2872 2873
                "slave_priority:%d\r\n"
                "slave_read_only:%d\r\n",
                server.slave_priority,
                server.repl_slave_ro);
A
antirez 已提交
2874
        }
2875

2876
        info = sdscatprintf(info,
2877
            "connected_slaves:%lu\r\n",
2878
            listLength(server.slaves));
2879 2880 2881 2882 2883 2884 2885 2886 2887 2888

        /* 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);
        }

2889 2890 2891 2892 2893 2894 2895 2896 2897
        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;
2898
                char ip[REDIS_IP_STR_LEN];
2899
                int port;
2900
                long lag = 0;
2901

2902
                if (anetPeerToString(slave->fd,ip,sizeof(ip),&port) == -1) continue;
2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915
                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;
2916 2917 2918 2919 2920
                if (slave->replstate == REDIS_REPL_ONLINE)
                    lag = time(NULL) - slave->repl_ack_time;

                info = sdscatprintf(info,
                    "slave%d:ip=%s,port=%d,state=%s,"
2921
                    "offset=%lld,lag=%ld\r\n",
2922
                    slaveid,ip,slave->slave_listening_port,state,
2923
                    slave->repl_ack_off, lag);
2924 2925 2926
                slaveid++;
            }
        }
2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937
        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
2938 2939
    }

2940 2941
    /* CPU */
    if (allsections || defsections || !strcasecmp(section,"cpu")) {
2942 2943
        if (sections++) info = sdscat(info,"\r\n");
        info = sdscatprintf(info,
2944
        "# CPU\r\n"
2945 2946
        "used_cpu_sys:%.2f\r\n"
        "used_cpu_user:%.2f\r\n"
2947 2948
        "used_cpu_sys_children:%.2f\r\n"
        "used_cpu_user_children:%.2f\r\n",
2949
        (float)self_ru.ru_stime.tv_sec+(float)self_ru.ru_stime.tv_usec/1000000,
2950 2951 2952
        (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);
2953
    }
2954

2955 2956 2957 2958
    /* cmdtime */
    if (allsections || !strcasecmp(section,"commandstats")) {
        if (sections++) info = sdscat(info,"\r\n");
        info = sdscatprintf(info, "# Commandstats\r\n");
2959
        numcommands = sizeof(redisCommandTable)/sizeof(struct redisCommand);
2960
        for (j = 0; j < numcommands; j++) {
2961
            struct redisCommand *c = redisCommandTable+j;
2962

2963 2964 2965 2966 2967
            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));
2968
        }
A
antirez 已提交
2969 2970
    }

2971
    /* Cluster */
2972 2973 2974 2975 2976 2977 2978 2979
    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);
    }

2980 2981 2982 2983 2984 2985
    /* 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;
2986

2987 2988 2989
            keys = dictSize(server.db[j].dict);
            vkeys = dictSize(server.db[j].expires);
            if (keys || vkeys) {
2990 2991 2992
                info = sdscatprintf(info,
                    "db%d:keys=%lld,expires=%lld,avg_ttl=%lld\r\n",
                    j, keys, vkeys, server.db[j].avg_ttl);
2993
            }
2994 2995 2996 2997 2998 2999
        }
    }
    return info;
}

void infoCommand(redisClient *c) {
3000 3001 3002 3003 3004 3005 3006
    char *section = c->argc == 2 ? c->argv[1]->ptr : "default";

    if (c->argc > 2) {
        addReply(c,shared.syntaxerr);
        return;
    }
    sds info = genRedisInfoString(section);
3007 3008 3009 3010 3011 3012 3013
    addReplySds(c,sdscatprintf(sdsempty(),"$%lu\r\n",
        (unsigned long)sdslen(info)));
    addReplySds(c,info);
    addReply(c,shared.crlf);
}

void monitorCommand(redisClient *c) {
3014
    /* ignore MONITOR if already slave or in monitor mode */
3015 3016 3017 3018 3019 3020 3021 3022 3023
    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
3024 3025 3026
/* freeMemoryIfNeeded() gets called when 'maxmemory' is set on the config
 * file to limit the max memory used by the server, before processing a
 * command.
3027
 *
3028 3029
 * The goal of the function is to free enough memory to keep Redis under the
 * configured memory limit.
3030
 *
3031 3032 3033 3034 3035 3036 3037 3038
 * 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
3039 3040 3041 3042 3043 3044
 *
 * ------------------------------------------------------------------------
 *
 * LRU approximation algorithm
 *
 * Redis uses an approximation of the LRU algorithm that runs in constant
J
Jan-Erik Rediger 已提交
3045
 * memory. Every time there is a key to expire, we sample N keys (with
A
antirez 已提交
3046 3047 3048 3049
 * 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
3050
 * with an old access time) if they are better than one of the current keys
A
antirez 已提交
3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082
 * 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. */
3083 3084

#define EVICTION_SAMPLES_ARRAY_SIZE 16
A
antirez 已提交
3085
void evictionPoolPopulate(dict *sampledict, dict *keydict, struct evictionPoolEntry *pool) {
3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103
    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
3104

3105
    for (j = 0; j < count; j++) {
A
antirez 已提交
3106 3107 3108
        unsigned long long idle;
        sds key;
        robj *o;
A
antirez 已提交
3109
        dictEntry *de;
A
antirez 已提交
3110

3111
        de = samples[j];
A
antirez 已提交
3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127
        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
3128
            /* Can't insert if the element is < the worst element we have
A
antirez 已提交
3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152
             * 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;
    }
3153
    if (samples != _samples) zfree(samples);
A
antirez 已提交
3154 3155
}

3156 3157 3158
int freeMemoryIfNeeded(void) {
    size_t mem_used, mem_tofree, mem_freed;
    int slaves = listLength(server.slaves);
3159
    mstime_t latency;
3160

3161 3162
    /* Remove the size of slaves output buffers and AOF buffer from the
     * count of used memory. */
3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177
    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;
        }
    }
3178 3179
    if (server.aof_state != REDIS_AOF_OFF) {
        mem_used -= sdslen(server.aof_buf);
3180
        mem_used -= aofRewriteBufferSize();
3181
    }
3182

3183 3184 3185 3186 3187 3188 3189 3190 3191
    /* 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;
3192
    latencyStartMonitor(latency);
3193 3194
    while (mem_freed < mem_tofree) {
        int j, k, keys_freed = 0;
3195

3196
        for (j = 0; j < server.dbnum; j++) {
3197
            long bestval = 0; /* just to prevent warning */
3198
            sds bestkey = NULL;
A
antirez 已提交
3199
            dictEntry *de;
3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212 3213 3214 3215 3216
            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);
3217
                bestkey = dictGetKey(de);
3218 3219 3220 3221 3222 3223
            }

            /* 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
3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 3234 3235 3236
                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,
3237
                            sizeof(pool[0])*(REDIS_EVICTION_POOL_SIZE-k-1));
A
antirez 已提交
3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251
                        /* 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;
                        }
3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262
                    }
                }
            }

            /* 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);
3263 3264
                    thiskey = dictGetKey(de);
                    thisval = (long) dictGetVal(de);
3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276

                    /* 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) {
3277 3278
                long long delta;

3279
                robj *keyobj = createStringObject(bestkey,sdslen(bestkey));
3280
                propagateExpire(db,keyobj);
3281 3282 3283 3284 3285 3286 3287 3288 3289
                /* 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();
3290
                dbDelete(db,keyobj);
3291 3292
                delta -= (long long) zmalloc_used_memory();
                mem_freed += delta;
3293
                server.stat_evictedkeys++;
3294 3295
                notifyKeyspaceEvent(REDIS_NOTIFY_EVICTED, "evicted",
                    keyobj, db->id);
3296
                decrRefCount(keyobj);
3297 3298 3299 3300 3301 3302
                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
3303
                if (slaves) flushSlavesOutputBuffers();
3304 3305
            }
        }
3306 3307 3308 3309 3310
        if (!keys_freed) {
            latencyEndMonitor(latency);
            latencyAddSampleIfNeeded("eviction-cycle",latency);
            return REDIS_ERR; /* nothing to free... */
        }
3311
    }
3312 3313
    latencyEndMonitor(latency);
    latencyAddSampleIfNeeded("eviction-cycle",latency);
3314
    return REDIS_OK;
3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329 3330 3331 3332 3333
}

/* =================================== 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);
}

3334
void linuxMemoryWarnings(void) {
3335 3336 3337
    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.");
    }
3338 3339 3340
    if (THPIsEnabled()) {
        redisLog(REDIS_WARNING,"WARNING you have Transparent Huge Pages (THP) support enabled in your kernel. This will create latency and memory usage issues with Redis. To fix this issue run the command 'echo never > /sys/kernel/mm/transparent_hugepage/enabled' as root, and add it to your /etc/rc.local in order to retain the setting after a reboot. Redis must be restarted after THP is disabled.");
    }
3341 3342 3343
}
#endif /* __linux__ */

3344 3345 3346 3347
void createPidFile(void) {
    /* Try to write the pid file in a best-effort way. */
    FILE *fp = fopen(server.pidfile,"w");
    if (fp) {
3348
        fprintf(fp,"%d\n",(int)getpid());
3349 3350 3351 3352
        fclose(fp);
    }
}

3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369
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);
    }
}

3370
void version(void) {
3371
    printf("Redis server v=%s sha=%s:%d malloc=%s bits=%d build=%llx\n",
A
antirez 已提交
3372 3373 3374 3375
        REDIS_VERSION,
        redisGitSHA1(),
        atoi(redisGitDirty()) > 0,
        ZMALLOC_LIB,
3376
        sizeof(long) == 4 ? 32 : 64,
3377
        (unsigned long long) redisBuildId());
3378 3379 3380
    exit(0);
}

3381
void usage(void) {
3382
    fprintf(stderr,"Usage: ./redis-server [/path/to/redis.conf] [options]\n");
3383
    fprintf(stderr,"       ./redis-server - (read config from stdin)\n");
3384
    fprintf(stderr,"       ./redis-server -v or --version\n");
3385 3386
    fprintf(stderr,"       ./redis-server -h or --help\n");
    fprintf(stderr,"       ./redis-server --test-memory <megabytes>\n\n");
3387 3388 3389 3390 3391
    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");
3392 3393 3394
    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");
3395 3396 3397
    exit(1);
}

A
antirez 已提交
3398 3399 3400
void redisAsciiArt(void) {
#include "asciilogo.h"
    char *buf = zmalloc(1024*16);
3401
    char *mode;
3402 3403 3404

    if (server.cluster_enabled) mode = "cluster";
    else if (server.sentinel_mode) mode = "sentinel";
3405
    else mode = "standalone";
A
antirez 已提交
3406

3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427
    if (server.syslog_enabled) {
        redisLog(REDIS_NOTICE,
            "Redis %s (%s/%d) %s bit, %s mode, port %d, pid %ld ready to start.",
            REDIS_VERSION,
            redisGitSHA1(),
            strtol(redisGitDirty(),NULL,10) > 0,
            (sizeof(long) == 8) ? "64" : "32",
            mode, server.port,
            (long) getpid()
        );
    } else {
        snprintf(buf,1024*16,ascii_logo,
            REDIS_VERSION,
            redisGitSHA1(),
            strtol(redisGitDirty(),NULL,10) > 0,
            (sizeof(long) == 8) ? "64" : "32",
            mode, server.port,
            (long) getpid()
        );
        redisLogRaw(REDIS_NOTICE|REDIS_LOG_RAW,buf);
    }
A
antirez 已提交
3428 3429 3430
    zfree(buf);
}

3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443 3444
static void sigShutdownHandler(int sig) {
    char *msg;

    switch (sig) {
    case SIGINT:
        msg = "Received SIGINT scheduling shutdown...";
        break;
    case SIGTERM:
        msg = "Received SIGTERM scheduling shutdown...";
        break;
    default:
        msg = "Received shutdown signal, scheduling shutdown...";
    };

3445 3446 3447 3448 3449 3450 3451 3452 3453 3454 3455 3456
    /* SIGINT is often delivered via Ctrl+C in an interactive session.
     * If we receive the signal the second time, we interpret this as
     * the user really wanting to quit ASAP without waiting to persist
     * on disk. */
    if (server.shutdown_asap && sig == SIGINT) {
        redisLogFromHandler(REDIS_WARNING, "You insist... exiting now.");
        rdbRemoveTempFile(getpid());
        exit(1); /* Exit with an error since this was not a clean shutdown. */
    } else if (server.loading) {
        exit(0);
    }

3457
    redisLogFromHandler(REDIS_WARNING, msg);
3458 3459 3460
    server.shutdown_asap = 1;
}

3461
void setupSignalHandlers(void) {
3462 3463
    struct sigaction act;

3464 3465 3466
    /* When the SA_SIGINFO flag is set in sa_flags then sa_sigaction is used.
     * Otherwise, sa_handler is used. */
    sigemptyset(&act.sa_mask);
3467
    act.sa_flags = 0;
3468
    act.sa_handler = sigShutdownHandler;
3469
    sigaction(SIGTERM, &act, NULL);
3470
    sigaction(SIGINT, &act, NULL);
3471

3472 3473
#ifdef HAVE_BACKTRACE
    sigemptyset(&act.sa_mask);
3474
    act.sa_flags = SA_NODEFER | SA_RESETHAND | SA_SIGINFO;
3475 3476 3477 3478 3479 3480 3481
    act.sa_sigaction = sigsegvHandler;
    sigaction(SIGSEGV, &act, NULL);
    sigaction(SIGBUS, &act, NULL);
    sigaction(SIGFPE, &act, NULL);
    sigaction(SIGILL, &act, NULL);
#endif
    return;
3482 3483
}

3484 3485
void memtest(size_t megabytes, int passes);

3486 3487 3488 3489 3490 3491 3492 3493 3494 3495 3496 3497 3498 3499 3500 3501 3502 3503 3504 3505 3506 3507
/* 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) {
3508
            redisLog(REDIS_WARNING,"Fatal error loading the DB: %s. Exiting.",strerror(errno));
3509 3510 3511 3512 3513
            exit(1);
        }
    }
}

A
antirez 已提交
3514 3515 3516
void redisOutOfMemoryHandler(size_t allocation_size) {
    redisLog(REDIS_WARNING,"Out Of Memory allocating %zu bytes!",
        allocation_size);
3517
    redisPanic("Redis aborting for OUT OF MEMORY");
A
antirez 已提交
3518 3519
}

3520
void redisSetProcTitle(char *title) {
3521
#ifdef USE_SETPROCTITLE
3522 3523 3524 3525 3526
    char *server_mode = "";
    if (server.cluster_enabled) server_mode = " [cluster]";
    else if (server.sentinel_mode) server_mode = " [sentinel]";

    setproctitle("%s %s:%d%s",
3527
        title,
A
antirez 已提交
3528
        server.bindaddr_count ? server.bindaddr[0] : "*",
3529 3530
        server.port,
        server_mode);
3531 3532 3533
#else
    REDIS_NOTUSED(title);
#endif
3534 3535
}

3536
int main(int argc, char **argv) {
3537
    struct timeval tv;
3538

A
antirez 已提交
3539
    /* We need to initialize our libraries, and the server configuration. */
3540 3541 3542
#ifdef INIT_SETPROCTITLE_REPLACEMENT
    spt_init(argc, argv);
#endif
3543
    setlocale(LC_COLLATE,"");
3544
    zmalloc_enable_thread_safeness();
A
antirez 已提交
3545
    zmalloc_set_oom_handler(redisOutOfMemoryHandler);
3546 3547 3548
    srand(time(NULL)^getpid());
    gettimeofday(&tv,NULL);
    dictSetHashFunctionSeed(tv.tv_sec^tv.tv_usec^getpid());
3549
    server.sentinel_mode = checkForSentinelMode(argc,argv);
3550
    initServerConfig();
3551

3552 3553 3554 3555 3556 3557 3558 3559
    /* 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();
    }

3560 3561 3562 3563 3564 3565
    if (argc >= 2) {
        int j = 1; /* First option to parse in argv[] */
        sds options = sdsempty();
        char *configfile = NULL;

        /* Handle special options --help and --version */
3566 3567
        if (strcmp(argv[1], "-v") == 0 ||
            strcmp(argv[1], "--version") == 0) version();
3568 3569
        if (strcmp(argv[1], "--help") == 0 ||
            strcmp(argv[1], "-h") == 0) usage();
3570 3571
        if (strcmp(argv[1], "--test-memory") == 0) {
            if (argc == 3) {
3572
                memtest(atoi(argv[2]),50);
3573 3574 3575 3576 3577 3578 3579 3580
                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);
            }
        }

3581 3582 3583 3584 3585 3586 3587 3588 3589 3590 3591 3592 3593 3594 3595 3596 3597 3598 3599 3600
        /* 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++;
        }
3601 3602 3603 3604 3605 3606 3607
        if (server.sentinel_mode && configfile && *configfile == '-') {
            redisLog(REDIS_WARNING,
                "Sentinel config from STDIN not allowed.");
            redisLog(REDIS_WARNING,
                "Sentinel needs config file on disk to save state.  Exiting...");
            exit(1);
        }
3608
        if (configfile) server.configfile = getAbsolutePath(configfile);
3609
        resetServerSaveParams();
3610 3611
        loadServerConfig(configfile,options);
        sdsfree(options);
3612
    } else {
3613
        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");
3614 3615 3616 3617
    }
    if (server.daemonize) daemonize();
    initServer();
    if (server.daemonize) createPidFile();
3618
    redisSetProcTitle(argv[0]);
3619
    redisAsciiArt();
3620 3621

    if (!server.sentinel_mode) {
A
antirez 已提交
3622
        /* Things not needed when running in Sentinel mode. */
3623 3624
        redisLog(REDIS_WARNING,"Server started, Redis version " REDIS_VERSION);
    #ifdef __linux__
3625
        linuxMemoryWarnings();
3626 3627
    #endif
        loadDataFromDisk();
3628 3629 3630 3631 3632 3633 3634 3635
        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);
            }
        }
3636
        if (server.ipfd_count > 0)
3637 3638 3639
            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);
3640
    } else {
A
antirez 已提交
3641
        sentinelIsRunning();
3642
    }
3643

3644 3645 3646 3647 3648
    /* 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);
    }

3649 3650 3651 3652 3653 3654
    aeSetBeforeSleepProc(server.el,beforeSleep);
    aeMain(server.el);
    aeDeleteEventLoop(server.el);
    return 0;
}

3655
/* The End */