]> git.saurik.com Git - redis.git/blob - src/redis.c
A filed called slave_read_only added in INFO output.
[redis.git] / src / redis.c
1 /*
2 * Copyright (c) 2009-2010, Salvatore Sanfilippo <antirez at gmail dot com>
3 * All rights reserved.
4 *
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions are met:
7 *
8 * * Redistributions of source code must retain the above copyright notice,
9 * this list of conditions and the following disclaimer.
10 * * Redistributions in binary form must reproduce the above copyright
11 * notice, this list of conditions and the following disclaimer in the
12 * documentation and/or other materials provided with the distribution.
13 * * Neither the name of Redis nor the names of its contributors may be used
14 * to endorse or promote products derived from this software without
15 * specific prior written permission.
16 *
17 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
18 * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20 * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
21 * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
22 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
23 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
24 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
25 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
26 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
27 * POSSIBILITY OF SUCH DAMAGE.
28 */
29
30 #include "redis.h"
31 #include "slowlog.h"
32 #include "bio.h"
33
34 #include <time.h>
35 #include <signal.h>
36 #include <sys/wait.h>
37 #include <errno.h>
38 #include <assert.h>
39 #include <ctype.h>
40 #include <stdarg.h>
41 #include <arpa/inet.h>
42 #include <sys/stat.h>
43 #include <fcntl.h>
44 #include <sys/time.h>
45 #include <sys/resource.h>
46 #include <sys/uio.h>
47 #include <limits.h>
48 #include <float.h>
49 #include <math.h>
50 #include <sys/resource.h>
51 #include <sys/utsname.h>
52
53 /* Our shared "common" objects */
54
55 struct sharedObjectsStruct shared;
56
57 /* Global vars that are actually used as constants. The following double
58 * values are used for double on-disk serialization, and are initialized
59 * at runtime to avoid strange compiler optimizations. */
60
61 double R_Zero, R_PosInf, R_NegInf, R_Nan;
62
63 /*================================= Globals ================================= */
64
65 /* Global vars */
66 struct redisServer server; /* server global state */
67 struct redisCommand *commandTable;
68
69 /* Our command table.
70 *
71 * Every entry is composed of the following fields:
72 *
73 * name: a string representing the command name.
74 * function: pointer to the C function implementing the command.
75 * arity: number of arguments, it is possible to use -N to say >= N
76 * sflags: command flags as string. See below for a table of flags.
77 * flags: flags as bitmask. Computed by Redis using the 'sflags' field.
78 * get_keys_proc: an optional function to get key arguments from a command.
79 * This is only used when the following three fields are not
80 * enough to specify what arguments are keys.
81 * first_key_index: first argument that is a key
82 * last_key_index: last argument that is a key
83 * key_step: step to get all the keys from first to last argument. For instance
84 * in MSET the step is two since arguments are key,val,key,val,...
85 * microseconds: microseconds of total execution time for this command.
86 * calls: total number of calls of this command.
87 *
88 * The flags, microseconds and calls fields are computed by Redis and should
89 * always be set to zero.
90 *
91 * Command flags are expressed using strings where every character represents
92 * a flag. Later the populateCommandTable() function will take care of
93 * populating the real 'flags' field using this characters.
94 *
95 * This is the meaning of the flags:
96 *
97 * w: write command (may modify the key space).
98 * r: read command (will never modify the key space).
99 * m: may increase memory usage once called. Don't allow if out of memory.
100 * a: admin command, like SAVE or SHUTDOWN.
101 * p: Pub/Sub related command.
102 * f: force replication of this command, regarless of server.dirty.
103 * s: command not allowed in scripts.
104 * R: random command. Command is not deterministic, that is, the same command
105 * with the same arguments, with the same key space, may have different
106 * results. For instance SPOP and RANDOMKEY are two random commands.
107 * S: Sort command output array if called from script, so that the output
108 * is deterministic.
109 * l: Allow command while loading the database.
110 * t: Allow command while a slave has stale data but is not allowed to
111 * server this data. Normally no command is accepted in this condition
112 * but just a few.
113 * M: Do not automatically propagate the command on MONITOR.
114 */
115 struct redisCommand redisCommandTable[] = {
116 {"get",getCommand,2,"r",0,NULL,1,1,1,0,0},
117 {"set",setCommand,3,"wm",0,noPreloadGetKeys,1,1,1,0,0},
118 {"setnx",setnxCommand,3,"wm",0,noPreloadGetKeys,1,1,1,0,0},
119 {"setex",setexCommand,4,"wm",0,noPreloadGetKeys,1,1,1,0,0},
120 {"psetex",psetexCommand,4,"wm",0,noPreloadGetKeys,1,1,1,0,0},
121 {"append",appendCommand,3,"wm",0,NULL,1,1,1,0,0},
122 {"strlen",strlenCommand,2,"r",0,NULL,1,1,1,0,0},
123 {"del",delCommand,-2,"w",0,noPreloadGetKeys,1,-1,1,0,0},
124 {"exists",existsCommand,2,"r",0,NULL,1,1,1,0,0},
125 {"setbit",setbitCommand,4,"wm",0,NULL,1,1,1,0,0},
126 {"getbit",getbitCommand,3,"r",0,NULL,1,1,1,0,0},
127 {"setrange",setrangeCommand,4,"wm",0,NULL,1,1,1,0,0},
128 {"getrange",getrangeCommand,4,"r",0,NULL,1,1,1,0,0},
129 {"substr",getrangeCommand,4,"r",0,NULL,1,1,1,0,0},
130 {"incr",incrCommand,2,"wm",0,NULL,1,1,1,0,0},
131 {"decr",decrCommand,2,"wm",0,NULL,1,1,1,0,0},
132 {"mget",mgetCommand,-2,"r",0,NULL,1,-1,1,0,0},
133 {"rpush",rpushCommand,-3,"wm",0,NULL,1,1,1,0,0},
134 {"lpush",lpushCommand,-3,"wm",0,NULL,1,1,1,0,0},
135 {"rpushx",rpushxCommand,3,"wm",0,NULL,1,1,1,0,0},
136 {"lpushx",lpushxCommand,3,"wm",0,NULL,1,1,1,0,0},
137 {"linsert",linsertCommand,5,"wm",0,NULL,1,1,1,0,0},
138 {"rpop",rpopCommand,2,"w",0,NULL,1,1,1,0,0},
139 {"lpop",lpopCommand,2,"w",0,NULL,1,1,1,0,0},
140 {"brpop",brpopCommand,-3,"ws",0,NULL,1,1,1,0,0},
141 {"brpoplpush",brpoplpushCommand,4,"wms",0,NULL,1,2,1,0,0},
142 {"blpop",blpopCommand,-3,"ws",0,NULL,1,-2,1,0,0},
143 {"llen",llenCommand,2,"r",0,NULL,1,1,1,0,0},
144 {"lindex",lindexCommand,3,"r",0,NULL,1,1,1,0,0},
145 {"lset",lsetCommand,4,"wm",0,NULL,1,1,1,0,0},
146 {"lrange",lrangeCommand,4,"r",0,NULL,1,1,1,0,0},
147 {"ltrim",ltrimCommand,4,"w",0,NULL,1,1,1,0,0},
148 {"lrem",lremCommand,4,"w",0,NULL,1,1,1,0,0},
149 {"rpoplpush",rpoplpushCommand,3,"wm",0,NULL,1,2,1,0,0},
150 {"sadd",saddCommand,-3,"wm",0,NULL,1,1,1,0,0},
151 {"srem",sremCommand,-3,"w",0,NULL,1,1,1,0,0},
152 {"smove",smoveCommand,4,"w",0,NULL,1,2,1,0,0},
153 {"sismember",sismemberCommand,3,"r",0,NULL,1,1,1,0,0},
154 {"scard",scardCommand,2,"r",0,NULL,1,1,1,0,0},
155 {"spop",spopCommand,2,"wRs",0,NULL,1,1,1,0,0},
156 {"srandmember",srandmemberCommand,-2,"rR",0,NULL,1,1,1,0,0},
157 {"sinter",sinterCommand,-2,"rS",0,NULL,1,-1,1,0,0},
158 {"sinterstore",sinterstoreCommand,-3,"wm",0,NULL,1,-1,1,0,0},
159 {"sunion",sunionCommand,-2,"rS",0,NULL,1,-1,1,0,0},
160 {"sunionstore",sunionstoreCommand,-3,"wm",0,NULL,1,-1,1,0,0},
161 {"sdiff",sdiffCommand,-2,"rS",0,NULL,1,-1,1,0,0},
162 {"sdiffstore",sdiffstoreCommand,-3,"wm",0,NULL,1,-1,1,0,0},
163 {"smembers",sinterCommand,2,"rS",0,NULL,1,1,1,0,0},
164 {"zadd",zaddCommand,-4,"wm",0,NULL,1,1,1,0,0},
165 {"zincrby",zincrbyCommand,4,"wm",0,NULL,1,1,1,0,0},
166 {"zrem",zremCommand,-3,"w",0,NULL,1,1,1,0,0},
167 {"zremrangebyscore",zremrangebyscoreCommand,4,"w",0,NULL,1,1,1,0,0},
168 {"zremrangebyrank",zremrangebyrankCommand,4,"w",0,NULL,1,1,1,0,0},
169 {"zunionstore",zunionstoreCommand,-4,"wm",0,zunionInterGetKeys,0,0,0,0,0},
170 {"zinterstore",zinterstoreCommand,-4,"wm",0,zunionInterGetKeys,0,0,0,0,0},
171 {"zrange",zrangeCommand,-4,"r",0,NULL,1,1,1,0,0},
172 {"zrangebyscore",zrangebyscoreCommand,-4,"r",0,NULL,1,1,1,0,0},
173 {"zrevrangebyscore",zrevrangebyscoreCommand,-4,"r",0,NULL,1,1,1,0,0},
174 {"zcount",zcountCommand,4,"r",0,NULL,1,1,1,0,0},
175 {"zrevrange",zrevrangeCommand,-4,"r",0,NULL,1,1,1,0,0},
176 {"zcard",zcardCommand,2,"r",0,NULL,1,1,1,0,0},
177 {"zscore",zscoreCommand,3,"r",0,NULL,1,1,1,0,0},
178 {"zrank",zrankCommand,3,"r",0,NULL,1,1,1,0,0},
179 {"zrevrank",zrevrankCommand,3,"r",0,NULL,1,1,1,0,0},
180 {"hset",hsetCommand,4,"wm",0,NULL,1,1,1,0,0},
181 {"hsetnx",hsetnxCommand,4,"wm",0,NULL,1,1,1,0,0},
182 {"hget",hgetCommand,3,"r",0,NULL,1,1,1,0,0},
183 {"hmset",hmsetCommand,-4,"wm",0,NULL,1,1,1,0,0},
184 {"hmget",hmgetCommand,-3,"r",0,NULL,1,1,1,0,0},
185 {"hincrby",hincrbyCommand,4,"wm",0,NULL,1,1,1,0,0},
186 {"hincrbyfloat",hincrbyfloatCommand,4,"wm",0,NULL,1,1,1,0,0},
187 {"hdel",hdelCommand,-3,"w",0,NULL,1,1,1,0,0},
188 {"hlen",hlenCommand,2,"r",0,NULL,1,1,1,0,0},
189 {"hkeys",hkeysCommand,2,"rS",0,NULL,1,1,1,0,0},
190 {"hvals",hvalsCommand,2,"rS",0,NULL,1,1,1,0,0},
191 {"hgetall",hgetallCommand,2,"r",0,NULL,1,1,1,0,0},
192 {"hexists",hexistsCommand,3,"r",0,NULL,1,1,1,0,0},
193 {"incrby",incrbyCommand,3,"wm",0,NULL,1,1,1,0,0},
194 {"decrby",decrbyCommand,3,"wm",0,NULL,1,1,1,0,0},
195 {"incrbyfloat",incrbyfloatCommand,3,"wm",0,NULL,1,1,1,0,0},
196 {"getset",getsetCommand,3,"wm",0,NULL,1,1,1,0,0},
197 {"mset",msetCommand,-3,"wm",0,NULL,1,-1,2,0,0},
198 {"msetnx",msetnxCommand,-3,"wm",0,NULL,1,-1,2,0,0},
199 {"randomkey",randomkeyCommand,1,"rR",0,NULL,0,0,0,0,0},
200 {"select",selectCommand,2,"r",0,NULL,0,0,0,0,0},
201 {"move",moveCommand,3,"w",0,NULL,1,1,1,0,0},
202 {"rename",renameCommand,3,"w",0,renameGetKeys,1,2,1,0,0},
203 {"renamenx",renamenxCommand,3,"w",0,renameGetKeys,1,2,1,0,0},
204 {"expire",expireCommand,3,"w",0,NULL,1,1,1,0,0},
205 {"expireat",expireatCommand,3,"w",0,NULL,1,1,1,0,0},
206 {"pexpire",pexpireCommand,3,"w",0,NULL,1,1,1,0,0},
207 {"pexpireat",pexpireatCommand,3,"w",0,NULL,1,1,1,0,0},
208 {"keys",keysCommand,2,"rS",0,NULL,0,0,0,0,0},
209 {"dbsize",dbsizeCommand,1,"r",0,NULL,0,0,0,0,0},
210 {"auth",authCommand,2,"rs",0,NULL,0,0,0,0,0},
211 {"ping",pingCommand,1,"r",0,NULL,0,0,0,0,0},
212 {"echo",echoCommand,2,"r",0,NULL,0,0,0,0,0},
213 {"save",saveCommand,1,"ars",0,NULL,0,0,0,0,0},
214 {"bgsave",bgsaveCommand,1,"ar",0,NULL,0,0,0,0,0},
215 {"bgrewriteaof",bgrewriteaofCommand,1,"ar",0,NULL,0,0,0,0,0},
216 {"shutdown",shutdownCommand,-1,"ar",0,NULL,0,0,0,0,0},
217 {"lastsave",lastsaveCommand,1,"r",0,NULL,0,0,0,0,0},
218 {"type",typeCommand,2,"r",0,NULL,1,1,1,0,0},
219 {"multi",multiCommand,1,"rs",0,NULL,0,0,0,0,0},
220 {"exec",execCommand,1,"sM",0,NULL,0,0,0,0,0},
221 {"discard",discardCommand,1,"rs",0,NULL,0,0,0,0,0},
222 {"sync",syncCommand,1,"ars",0,NULL,0,0,0,0,0},
223 {"replconf",replconfCommand,-1,"ars",0,NULL,0,0,0,0,0},
224 {"flushdb",flushdbCommand,1,"w",0,NULL,0,0,0,0,0},
225 {"flushall",flushallCommand,1,"w",0,NULL,0,0,0,0,0},
226 {"sort",sortCommand,-2,"wm",0,NULL,1,1,1,0,0},
227 {"info",infoCommand,-1,"rlt",0,NULL,0,0,0,0,0},
228 {"monitor",monitorCommand,1,"ars",0,NULL,0,0,0,0,0},
229 {"ttl",ttlCommand,2,"r",0,NULL,1,1,1,0,0},
230 {"pttl",pttlCommand,2,"r",0,NULL,1,1,1,0,0},
231 {"persist",persistCommand,2,"w",0,NULL,1,1,1,0,0},
232 {"slaveof",slaveofCommand,3,"ast",0,NULL,0,0,0,0,0},
233 {"debug",debugCommand,-2,"as",0,NULL,0,0,0,0,0},
234 {"config",configCommand,-2,"ar",0,NULL,0,0,0,0,0},
235 {"subscribe",subscribeCommand,-2,"rpslt",0,NULL,0,0,0,0,0},
236 {"unsubscribe",unsubscribeCommand,-1,"rpslt",0,NULL,0,0,0,0,0},
237 {"psubscribe",psubscribeCommand,-2,"rpslt",0,NULL,0,0,0,0,0},
238 {"punsubscribe",punsubscribeCommand,-1,"rpslt",0,NULL,0,0,0,0,0},
239 {"publish",publishCommand,3,"pflt",0,NULL,0,0,0,0,0},
240 {"watch",watchCommand,-2,"rs",0,noPreloadGetKeys,1,-1,1,0,0},
241 {"unwatch",unwatchCommand,1,"rs",0,NULL,0,0,0,0,0},
242 {"restore",restoreCommand,4,"awm",0,NULL,1,1,1,0,0},
243 {"migrate",migrateCommand,6,"aw",0,NULL,0,0,0,0,0},
244 {"dump",dumpCommand,2,"ar",0,NULL,1,1,1,0,0},
245 {"object",objectCommand,-2,"r",0,NULL,2,2,2,0,0},
246 {"client",clientCommand,-2,"ar",0,NULL,0,0,0,0,0},
247 {"eval",evalCommand,-3,"s",0,zunionInterGetKeys,0,0,0,0,0},
248 {"evalsha",evalShaCommand,-3,"s",0,zunionInterGetKeys,0,0,0,0,0},
249 {"slowlog",slowlogCommand,-2,"r",0,NULL,0,0,0,0,0},
250 {"script",scriptCommand,-2,"ras",0,NULL,0,0,0,0,0},
251 {"time",timeCommand,1,"rR",0,NULL,0,0,0,0,0},
252 {"bitop",bitopCommand,-4,"wm",0,NULL,2,-1,1,0,0},
253 {"bitcount",bitcountCommand,-2,"r",0,NULL,1,1,1,0,0}
254 };
255
256 /*============================ Utility functions ============================ */
257
258 /* Low level logging. To use only for very big messages, otherwise
259 * redisLog() is to prefer. */
260 void redisLogRaw(int level, const char *msg) {
261 const int syslogLevelMap[] = { LOG_DEBUG, LOG_INFO, LOG_NOTICE, LOG_WARNING };
262 const char *c = ".-*#";
263 FILE *fp;
264 char buf[64];
265 int rawmode = (level & REDIS_LOG_RAW);
266
267 level &= 0xff; /* clear flags */
268 if (level < server.verbosity) return;
269
270 fp = (server.logfile == NULL) ? stdout : fopen(server.logfile,"a");
271 if (!fp) return;
272
273 if (rawmode) {
274 fprintf(fp,"%s",msg);
275 } else {
276 int off;
277 struct timeval tv;
278
279 gettimeofday(&tv,NULL);
280 off = strftime(buf,sizeof(buf),"%d %b %H:%M:%S.",localtime(&tv.tv_sec));
281 snprintf(buf+off,sizeof(buf)-off,"%03d",(int)tv.tv_usec/1000);
282 fprintf(fp,"[%d] %s %c %s\n",(int)getpid(),buf,c[level],msg);
283 }
284 fflush(fp);
285
286 if (server.logfile) fclose(fp);
287
288 if (server.syslog_enabled) syslog(syslogLevelMap[level], "%s", msg);
289 }
290
291 /* Like redisLogRaw() but with printf-alike support. This is the funciton that
292 * is used across the code. The raw version is only used in order to dump
293 * the INFO output on crash. */
294 void redisLog(int level, const char *fmt, ...) {
295 va_list ap;
296 char msg[REDIS_MAX_LOGMSG_LEN];
297
298 if ((level&0xff) < server.verbosity) return;
299
300 va_start(ap, fmt);
301 vsnprintf(msg, sizeof(msg), fmt, ap);
302 va_end(ap);
303
304 redisLogRaw(level,msg);
305 }
306
307 /* Log a fixed message without printf-alike capabilities, in a way that is
308 * safe to call from a signal handler.
309 *
310 * We actually use this only for signals that are not fatal from the point
311 * of view of Redis. Signals that are going to kill the server anyway and
312 * where we need printf-alike features are served by redisLog(). */
313 void redisLogFromHandler(int level, const char *msg) {
314 int fd;
315 char buf[64];
316
317 if ((level&0xff) < server.verbosity ||
318 (server.logfile == NULL && server.daemonize)) return;
319 fd = server.logfile ?
320 open(server.logfile, O_APPEND|O_CREAT|O_WRONLY, 0644) :
321 STDOUT_FILENO;
322 if (fd == -1) return;
323 ll2string(buf,sizeof(buf),getpid());
324 if (write(fd,"[",1) == -1) goto err;
325 if (write(fd,buf,strlen(buf)) == -1) goto err;
326 if (write(fd," | signal handler] (",20) == -1) goto err;
327 ll2string(buf,sizeof(buf),time(NULL));
328 if (write(fd,buf,strlen(buf)) == -1) goto err;
329 if (write(fd,") ",2) == -1) goto err;
330 if (write(fd,msg,strlen(msg)) == -1) goto err;
331 if (write(fd,"\n",1) == -1) goto err;
332 err:
333 if (server.logfile) close(fd);
334 }
335
336 /* Return the UNIX time in microseconds */
337 long long ustime(void) {
338 struct timeval tv;
339 long long ust;
340
341 gettimeofday(&tv, NULL);
342 ust = ((long long)tv.tv_sec)*1000000;
343 ust += tv.tv_usec;
344 return ust;
345 }
346
347 /* Return the UNIX time in milliseconds */
348 long long mstime(void) {
349 return ustime()/1000;
350 }
351
352 /* After an RDB dump or AOF rewrite we exit from children using _exit() instead of
353 * exit(), because the latter may interact with the same file objects used by
354 * the parent process. However if we are testing the coverage normal exit() is
355 * used in order to obtain the right coverage information. */
356 void exitFromChild(int retcode) {
357 #ifdef COVERAGE_TEST
358 exit(retcode);
359 #else
360 _exit(retcode);
361 #endif
362 }
363
364 /*====================== Hash table type implementation ==================== */
365
366 /* This is an hash table type that uses the SDS dynamic strings libary as
367 * keys and radis objects as values (objects can hold SDS strings,
368 * lists, sets). */
369
370 void dictVanillaFree(void *privdata, void *val)
371 {
372 DICT_NOTUSED(privdata);
373 zfree(val);
374 }
375
376 void dictListDestructor(void *privdata, void *val)
377 {
378 DICT_NOTUSED(privdata);
379 listRelease((list*)val);
380 }
381
382 int dictSdsKeyCompare(void *privdata, const void *key1,
383 const void *key2)
384 {
385 int l1,l2;
386 DICT_NOTUSED(privdata);
387
388 l1 = sdslen((sds)key1);
389 l2 = sdslen((sds)key2);
390 if (l1 != l2) return 0;
391 return memcmp(key1, key2, l1) == 0;
392 }
393
394 /* A case insensitive version used for the command lookup table. */
395 int dictSdsKeyCaseCompare(void *privdata, const void *key1,
396 const void *key2)
397 {
398 DICT_NOTUSED(privdata);
399
400 return strcasecmp(key1, key2) == 0;
401 }
402
403 void dictRedisObjectDestructor(void *privdata, void *val)
404 {
405 DICT_NOTUSED(privdata);
406
407 if (val == NULL) return; /* Values of swapped out keys as set to NULL */
408 decrRefCount(val);
409 }
410
411 void dictSdsDestructor(void *privdata, void *val)
412 {
413 DICT_NOTUSED(privdata);
414
415 sdsfree(val);
416 }
417
418 int dictObjKeyCompare(void *privdata, const void *key1,
419 const void *key2)
420 {
421 const robj *o1 = key1, *o2 = key2;
422 return dictSdsKeyCompare(privdata,o1->ptr,o2->ptr);
423 }
424
425 unsigned int dictObjHash(const void *key) {
426 const robj *o = key;
427 return dictGenHashFunction(o->ptr, sdslen((sds)o->ptr));
428 }
429
430 unsigned int dictSdsHash(const void *key) {
431 return dictGenHashFunction((unsigned char*)key, sdslen((char*)key));
432 }
433
434 unsigned int dictSdsCaseHash(const void *key) {
435 return dictGenCaseHashFunction((unsigned char*)key, sdslen((char*)key));
436 }
437
438 int dictEncObjKeyCompare(void *privdata, const void *key1,
439 const void *key2)
440 {
441 robj *o1 = (robj*) key1, *o2 = (robj*) key2;
442 int cmp;
443
444 if (o1->encoding == REDIS_ENCODING_INT &&
445 o2->encoding == REDIS_ENCODING_INT)
446 return o1->ptr == o2->ptr;
447
448 o1 = getDecodedObject(o1);
449 o2 = getDecodedObject(o2);
450 cmp = dictSdsKeyCompare(privdata,o1->ptr,o2->ptr);
451 decrRefCount(o1);
452 decrRefCount(o2);
453 return cmp;
454 }
455
456 unsigned int dictEncObjHash(const void *key) {
457 robj *o = (robj*) key;
458
459 if (o->encoding == REDIS_ENCODING_RAW) {
460 return dictGenHashFunction(o->ptr, sdslen((sds)o->ptr));
461 } else {
462 if (o->encoding == REDIS_ENCODING_INT) {
463 char buf[32];
464 int len;
465
466 len = ll2string(buf,32,(long)o->ptr);
467 return dictGenHashFunction((unsigned char*)buf, len);
468 } else {
469 unsigned int hash;
470
471 o = getDecodedObject(o);
472 hash = dictGenHashFunction(o->ptr, sdslen((sds)o->ptr));
473 decrRefCount(o);
474 return hash;
475 }
476 }
477 }
478
479 /* Sets type hash table */
480 dictType setDictType = {
481 dictEncObjHash, /* hash function */
482 NULL, /* key dup */
483 NULL, /* val dup */
484 dictEncObjKeyCompare, /* key compare */
485 dictRedisObjectDestructor, /* key destructor */
486 NULL /* val destructor */
487 };
488
489 /* Sorted sets hash (note: a skiplist is used in addition to the hash table) */
490 dictType zsetDictType = {
491 dictEncObjHash, /* hash function */
492 NULL, /* key dup */
493 NULL, /* val dup */
494 dictEncObjKeyCompare, /* key compare */
495 dictRedisObjectDestructor, /* key destructor */
496 NULL /* val destructor */
497 };
498
499 /* Db->dict, keys are sds strings, vals are Redis objects. */
500 dictType dbDictType = {
501 dictSdsHash, /* hash function */
502 NULL, /* key dup */
503 NULL, /* val dup */
504 dictSdsKeyCompare, /* key compare */
505 dictSdsDestructor, /* key destructor */
506 dictRedisObjectDestructor /* val destructor */
507 };
508
509 /* Db->expires */
510 dictType keyptrDictType = {
511 dictSdsHash, /* hash function */
512 NULL, /* key dup */
513 NULL, /* val dup */
514 dictSdsKeyCompare, /* key compare */
515 NULL, /* key destructor */
516 NULL /* val destructor */
517 };
518
519 /* Command table. sds string -> command struct pointer. */
520 dictType commandTableDictType = {
521 dictSdsCaseHash, /* hash function */
522 NULL, /* key dup */
523 NULL, /* val dup */
524 dictSdsKeyCaseCompare, /* key compare */
525 dictSdsDestructor, /* key destructor */
526 NULL /* val destructor */
527 };
528
529 /* Hash type hash table (note that small hashes are represented with zimpaps) */
530 dictType hashDictType = {
531 dictEncObjHash, /* hash function */
532 NULL, /* key dup */
533 NULL, /* val dup */
534 dictEncObjKeyCompare, /* key compare */
535 dictRedisObjectDestructor, /* key destructor */
536 dictRedisObjectDestructor /* val destructor */
537 };
538
539 /* Keylist hash table type has unencoded redis objects as keys and
540 * lists as values. It's used for blocking operations (BLPOP) and to
541 * map swapped keys to a list of clients waiting for this keys to be loaded. */
542 dictType keylistDictType = {
543 dictObjHash, /* hash function */
544 NULL, /* key dup */
545 NULL, /* val dup */
546 dictObjKeyCompare, /* key compare */
547 dictRedisObjectDestructor, /* key destructor */
548 dictListDestructor /* val destructor */
549 };
550
551 int htNeedsResize(dict *dict) {
552 long long size, used;
553
554 size = dictSlots(dict);
555 used = dictSize(dict);
556 return (size && used && size > DICT_HT_INITIAL_SIZE &&
557 (used*100/size < REDIS_HT_MINFILL));
558 }
559
560 /* If the percentage of used slots in the HT reaches REDIS_HT_MINFILL
561 * we resize the hash table to save memory */
562 void tryResizeHashTables(void) {
563 int j;
564
565 for (j = 0; j < server.dbnum; j++) {
566 if (htNeedsResize(server.db[j].dict))
567 dictResize(server.db[j].dict);
568 if (htNeedsResize(server.db[j].expires))
569 dictResize(server.db[j].expires);
570 }
571 }
572
573 /* Our hash table implementation performs rehashing incrementally while
574 * we write/read from the hash table. Still if the server is idle, the hash
575 * table will use two tables for a long time. So we try to use 1 millisecond
576 * of CPU time at every serverCron() loop in order to rehash some key. */
577 void incrementallyRehash(void) {
578 int j;
579
580 for (j = 0; j < server.dbnum; j++) {
581 /* Keys dictionary */
582 if (dictIsRehashing(server.db[j].dict)) {
583 dictRehashMilliseconds(server.db[j].dict,1);
584 break; /* already used our millisecond for this loop... */
585 }
586 /* Expires */
587 if (dictIsRehashing(server.db[j].expires)) {
588 dictRehashMilliseconds(server.db[j].expires,1);
589 break; /* already used our millisecond for this loop... */
590 }
591 }
592 }
593
594 /* This function is called once a background process of some kind terminates,
595 * as we want to avoid resizing the hash tables when there is a child in order
596 * to play well with copy-on-write (otherwise when a resize happens lots of
597 * memory pages are copied). The goal of this function is to update the ability
598 * for dict.c to resize the hash tables accordingly to the fact we have o not
599 * running childs. */
600 void updateDictResizePolicy(void) {
601 if (server.rdb_child_pid == -1 && server.aof_child_pid == -1)
602 dictEnableResize();
603 else
604 dictDisableResize();
605 }
606
607 /* ======================= Cron: called every 100 ms ======================== */
608
609 /* Try to expire a few timed out keys. The algorithm used is adaptive and
610 * will use few CPU cycles if there are few expiring keys, otherwise
611 * it will get more aggressive to avoid that too much memory is used by
612 * keys that can be removed from the keyspace. */
613 void activeExpireCycle(void) {
614 int j, iteration = 0;
615 long long start = ustime(), timelimit;
616
617 /* We can use at max REDIS_EXPIRELOOKUPS_TIME_PERC percentage of CPU time
618 * per iteration. Since this function gets called with a frequency of
619 * REDIS_HZ times per second, the following is the max amount of
620 * microseconds we can spend in this function. */
621 timelimit = 1000000*REDIS_EXPIRELOOKUPS_TIME_PERC/REDIS_HZ/100;
622 if (timelimit <= 0) timelimit = 1;
623
624 for (j = 0; j < server.dbnum; j++) {
625 int expired;
626 redisDb *db = server.db+j;
627
628 /* Continue to expire if at the end of the cycle more than 25%
629 * of the keys were expired. */
630 do {
631 unsigned long num = dictSize(db->expires);
632 unsigned long slots = dictSlots(db->expires);
633 long long now = mstime();
634
635 /* When there are less than 1% filled slots getting random
636 * keys is expensive, so stop here waiting for better times...
637 * The dictionary will be resized asap. */
638 if (num && slots > DICT_HT_INITIAL_SIZE &&
639 (num*100/slots < 1)) break;
640
641 /* The main collection cycle. Sample random keys among keys
642 * with an expire set, checking for expired ones. */
643 expired = 0;
644 if (num > REDIS_EXPIRELOOKUPS_PER_CRON)
645 num = REDIS_EXPIRELOOKUPS_PER_CRON;
646 while (num--) {
647 dictEntry *de;
648 long long t;
649
650 if ((de = dictGetRandomKey(db->expires)) == NULL) break;
651 t = dictGetSignedIntegerVal(de);
652 if (now > t) {
653 sds key = dictGetKey(de);
654 robj *keyobj = createStringObject(key,sdslen(key));
655
656 propagateExpire(db,keyobj);
657 dbDelete(db,keyobj);
658 decrRefCount(keyobj);
659 expired++;
660 server.stat_expiredkeys++;
661 }
662 }
663 /* We can't block forever here even if there are many keys to
664 * expire. So after a given amount of milliseconds return to the
665 * caller waiting for the other active expire cycle. */
666 iteration++;
667 if ((iteration & 0xf) == 0 && /* check once every 16 cycles. */
668 (ustime()-start) > timelimit) return;
669 } while (expired > REDIS_EXPIRELOOKUPS_PER_CRON/4);
670 }
671 }
672
673 void updateLRUClock(void) {
674 server.lruclock = (server.unixtime/REDIS_LRU_CLOCK_RESOLUTION) &
675 REDIS_LRU_CLOCK_MAX;
676 }
677
678
679 /* Add a sample to the operations per second array of samples. */
680 void trackOperationsPerSecond(void) {
681 long long t = mstime() - server.ops_sec_last_sample_time;
682 long long ops = server.stat_numcommands - server.ops_sec_last_sample_ops;
683 long long ops_sec;
684
685 ops_sec = t > 0 ? (ops*1000/t) : 0;
686
687 server.ops_sec_samples[server.ops_sec_idx] = ops_sec;
688 server.ops_sec_idx = (server.ops_sec_idx+1) % REDIS_OPS_SEC_SAMPLES;
689 server.ops_sec_last_sample_time = mstime();
690 server.ops_sec_last_sample_ops = server.stat_numcommands;
691 }
692
693 /* Return the mean of all the samples. */
694 long long getOperationsPerSecond(void) {
695 int j;
696 long long sum = 0;
697
698 for (j = 0; j < REDIS_OPS_SEC_SAMPLES; j++)
699 sum += server.ops_sec_samples[j];
700 return sum / REDIS_OPS_SEC_SAMPLES;
701 }
702
703 /* Check for timeouts. Returns non-zero if the client was terminated */
704 int clientsCronHandleTimeout(redisClient *c) {
705 time_t now = server.unixtime;
706
707 if (server.maxidletime &&
708 !(c->flags & REDIS_SLAVE) && /* no timeout for slaves */
709 !(c->flags & REDIS_MASTER) && /* no timeout for masters */
710 !(c->flags & REDIS_BLOCKED) && /* no timeout for BLPOP */
711 dictSize(c->pubsub_channels) == 0 && /* no timeout for pubsub */
712 listLength(c->pubsub_patterns) == 0 &&
713 (now - c->lastinteraction > server.maxidletime))
714 {
715 redisLog(REDIS_VERBOSE,"Closing idle client");
716 freeClient(c);
717 return 1;
718 } else if (c->flags & REDIS_BLOCKED) {
719 if (c->bpop.timeout != 0 && c->bpop.timeout < now) {
720 addReply(c,shared.nullmultibulk);
721 unblockClientWaitingData(c);
722 }
723 }
724 return 0;
725 }
726
727 /* The client query buffer is an sds.c string that can end with a lot of
728 * free space not used, this function reclaims space if needed.
729 *
730 * The funciton always returns 0 as it never terminates the client. */
731 int clientsCronResizeQueryBuffer(redisClient *c) {
732 size_t querybuf_size = sdsAllocSize(c->querybuf);
733 time_t idletime = server.unixtime - c->lastinteraction;
734
735 /* There are two conditions to resize the query buffer:
736 * 1) Query buffer is > BIG_ARG and too big for latest peak.
737 * 2) Client is inactive and the buffer is bigger than 1k. */
738 if (((querybuf_size > REDIS_MBULK_BIG_ARG) &&
739 (querybuf_size/(c->querybuf_peak+1)) > 2) ||
740 (querybuf_size > 1024 && idletime > 2))
741 {
742 /* Only resize the query buffer if it is actually wasting space. */
743 if (sdsavail(c->querybuf) > 1024) {
744 c->querybuf = sdsRemoveFreeSpace(c->querybuf);
745 }
746 }
747 /* Reset the peak again to capture the peak memory usage in the next
748 * cycle. */
749 c->querybuf_peak = 0;
750 return 0;
751 }
752
753 void clientsCron(void) {
754 /* Make sure to process at least 1/(REDIS_HZ*10) of clients per call.
755 * Since this function is called REDIS_HZ times per second we are sure that
756 * in the worst case we process all the clients in 10 seconds.
757 * In normal conditions (a reasonable number of clients) we process
758 * all the clients in a shorter time. */
759 int numclients = listLength(server.clients);
760 int iterations = numclients/(REDIS_HZ*10);
761
762 if (iterations < 50)
763 iterations = (numclients < 50) ? numclients : 50;
764 while(listLength(server.clients) && iterations--) {
765 redisClient *c;
766 listNode *head;
767
768 /* Rotate the list, take the current head, process.
769 * This way if the client must be removed from the list it's the
770 * first element and we don't incur into O(N) computation. */
771 listRotate(server.clients);
772 head = listFirst(server.clients);
773 c = listNodeValue(head);
774 /* The following functions do different service checks on the client.
775 * The protocol is that they return non-zero if the client was
776 * terminated. */
777 if (clientsCronHandleTimeout(c)) continue;
778 if (clientsCronResizeQueryBuffer(c)) continue;
779 }
780 }
781
782 /* This is our timer interrupt, called REDIS_HZ times per second.
783 * Here is where we do a number of things that need to be done asynchronously.
784 * For instance:
785 *
786 * - Active expired keys collection (it is also performed in a lazy way on
787 * lookup).
788 * - Software watchdong.
789 * - Update some statistic.
790 * - Incremental rehashing of the DBs hash tables.
791 * - Triggering BGSAVE / AOF rewrite, and handling of terminated children.
792 * - Clients timeout of differnet kinds.
793 * - Replication reconnection.
794 * - Many more...
795 *
796 * Everything directly called here will be called REDIS_HZ times per second,
797 * so in order to throttle execution of things we want to do less frequently
798 * a macro is used: run_with_period(milliseconds) { .... }
799 */
800
801 int serverCron(struct aeEventLoop *eventLoop, long long id, void *clientData) {
802 int j;
803 REDIS_NOTUSED(eventLoop);
804 REDIS_NOTUSED(id);
805 REDIS_NOTUSED(clientData);
806
807 /* Software watchdog: deliver the SIGALRM that will reach the signal
808 * handler if we don't return here fast enough. */
809 if (server.watchdog_period) watchdogScheduleSignal(server.watchdog_period);
810
811 /* We take a cached value of the unix time in the global state because
812 * with virtual memory and aging there is to store the current time
813 * in objects at every object access, and accuracy is not needed.
814 * To access a global var is faster than calling time(NULL) */
815 server.unixtime = time(NULL);
816
817 run_with_period(100) trackOperationsPerSecond();
818
819 /* We have just 22 bits per object for LRU information.
820 * So we use an (eventually wrapping) LRU clock with 10 seconds resolution.
821 * 2^22 bits with 10 seconds resoluton is more or less 1.5 years.
822 *
823 * Note that even if this will wrap after 1.5 years it's not a problem,
824 * everything will still work but just some object will appear younger
825 * to Redis. But for this to happen a given object should never be touched
826 * for 1.5 years.
827 *
828 * Note that you can change the resolution altering the
829 * REDIS_LRU_CLOCK_RESOLUTION define.
830 */
831 updateLRUClock();
832
833 /* Record the max memory used since the server was started. */
834 if (zmalloc_used_memory() > server.stat_peak_memory)
835 server.stat_peak_memory = zmalloc_used_memory();
836
837 /* We received a SIGTERM, shutting down here in a safe way, as it is
838 * not ok doing so inside the signal handler. */
839 if (server.shutdown_asap) {
840 if (prepareForShutdown(0) == REDIS_OK) exit(0);
841 redisLog(REDIS_WARNING,"SIGTERM received but errors trying to shut down the server, check the logs for more information");
842 }
843
844 /* Show some info about non-empty databases */
845 run_with_period(5000) {
846 for (j = 0; j < server.dbnum; j++) {
847 long long size, used, vkeys;
848
849 size = dictSlots(server.db[j].dict);
850 used = dictSize(server.db[j].dict);
851 vkeys = dictSize(server.db[j].expires);
852 if (used || vkeys) {
853 redisLog(REDIS_VERBOSE,"DB %d: %lld keys (%lld volatile) in %lld slots HT.",j,used,vkeys,size);
854 /* dictPrintStats(server.dict); */
855 }
856 }
857 }
858
859 /* We don't want to resize the hash tables while a bacground saving
860 * is in progress: the saving child is created using fork() that is
861 * implemented with a copy-on-write semantic in most modern systems, so
862 * if we resize the HT while there is the saving child at work actually
863 * a lot of memory movements in the parent will cause a lot of pages
864 * copied. */
865 if (server.rdb_child_pid == -1 && server.aof_child_pid == -1) {
866 tryResizeHashTables();
867 if (server.activerehashing) incrementallyRehash();
868 }
869
870 /* Show information about connected clients */
871 if (!server.sentinel_mode) {
872 run_with_period(5000) {
873 redisLog(REDIS_VERBOSE,
874 "%d clients connected (%d slaves), %zu bytes in use",
875 listLength(server.clients)-listLength(server.slaves),
876 listLength(server.slaves),
877 zmalloc_used_memory());
878 }
879 }
880
881 /* We need to do a few operations on clients asynchronously. */
882 clientsCron();
883
884 /* Start a scheduled AOF rewrite if this was requested by the user while
885 * a BGSAVE was in progress. */
886 if (server.rdb_child_pid == -1 && server.aof_child_pid == -1 &&
887 server.aof_rewrite_scheduled)
888 {
889 rewriteAppendOnlyFileBackground();
890 }
891
892 /* Check if a background saving or AOF rewrite in progress terminated. */
893 if (server.rdb_child_pid != -1 || server.aof_child_pid != -1) {
894 int statloc;
895 pid_t pid;
896
897 if ((pid = wait3(&statloc,WNOHANG,NULL)) != 0) {
898 int exitcode = WEXITSTATUS(statloc);
899 int bysignal = 0;
900
901 if (WIFSIGNALED(statloc)) bysignal = WTERMSIG(statloc);
902
903 if (pid == server.rdb_child_pid) {
904 backgroundSaveDoneHandler(exitcode,bysignal);
905 } else {
906 backgroundRewriteDoneHandler(exitcode,bysignal);
907 }
908 updateDictResizePolicy();
909 }
910 } else {
911 /* If there is not a background saving/rewrite in progress check if
912 * we have to save/rewrite now */
913 for (j = 0; j < server.saveparamslen; j++) {
914 struct saveparam *sp = server.saveparams+j;
915
916 if (server.dirty >= sp->changes &&
917 server.unixtime-server.lastsave > sp->seconds) {
918 redisLog(REDIS_NOTICE,"%d changes in %d seconds. Saving...",
919 sp->changes, sp->seconds);
920 rdbSaveBackground(server.rdb_filename);
921 break;
922 }
923 }
924
925 /* Trigger an AOF rewrite if needed */
926 if (server.rdb_child_pid == -1 &&
927 server.aof_child_pid == -1 &&
928 server.aof_rewrite_perc &&
929 server.aof_current_size > server.aof_rewrite_min_size)
930 {
931 long long base = server.aof_rewrite_base_size ?
932 server.aof_rewrite_base_size : 1;
933 long long growth = (server.aof_current_size*100/base) - 100;
934 if (growth >= server.aof_rewrite_perc) {
935 redisLog(REDIS_NOTICE,"Starting automatic rewriting of AOF on %lld%% growth",growth);
936 rewriteAppendOnlyFileBackground();
937 }
938 }
939 }
940
941
942 /* If we postponed an AOF buffer flush, let's try to do it every time the
943 * cron function is called. */
944 if (server.aof_flush_postponed_start) flushAppendOnlyFile(0);
945
946 /* Expire a few keys per cycle, only if this is a master.
947 * On slaves we wait for DEL operations synthesized by the master
948 * in order to guarantee a strict consistency. */
949 if (server.masterhost == NULL) activeExpireCycle();
950
951 /* Close clients that need to be closed asynchronous */
952 freeClientsInAsyncFreeQueue();
953
954 /* Replication cron function -- used to reconnect to master and
955 * to detect transfer failures. */
956 run_with_period(1000) replicationCron();
957
958 /* Run the sentinel timer if we are in sentinel mode. */
959 run_with_period(100) {
960 if (server.sentinel_mode) sentinelTimer();
961 }
962
963 server.cronloops++;
964 return 1000/REDIS_HZ;
965 }
966
967 /* This function gets called every time Redis is entering the
968 * main loop of the event driven library, that is, before to sleep
969 * for ready file descriptors. */
970 void beforeSleep(struct aeEventLoop *eventLoop) {
971 REDIS_NOTUSED(eventLoop);
972 listNode *ln;
973 redisClient *c;
974
975 /* Try to process pending commands for clients that were just unblocked. */
976 while (listLength(server.unblocked_clients)) {
977 ln = listFirst(server.unblocked_clients);
978 redisAssert(ln != NULL);
979 c = ln->value;
980 listDelNode(server.unblocked_clients,ln);
981 c->flags &= ~REDIS_UNBLOCKED;
982
983 /* Process remaining data in the input buffer. */
984 if (c->querybuf && sdslen(c->querybuf) > 0) {
985 server.current_client = c;
986 processInputBuffer(c);
987 server.current_client = NULL;
988 }
989 }
990
991 /* Write the AOF buffer on disk */
992 flushAppendOnlyFile(0);
993 }
994
995 /* =========================== Server initialization ======================== */
996
997 void createSharedObjects(void) {
998 int j;
999
1000 shared.crlf = createObject(REDIS_STRING,sdsnew("\r\n"));
1001 shared.ok = createObject(REDIS_STRING,sdsnew("+OK\r\n"));
1002 shared.err = createObject(REDIS_STRING,sdsnew("-ERR\r\n"));
1003 shared.emptybulk = createObject(REDIS_STRING,sdsnew("$0\r\n\r\n"));
1004 shared.czero = createObject(REDIS_STRING,sdsnew(":0\r\n"));
1005 shared.cone = createObject(REDIS_STRING,sdsnew(":1\r\n"));
1006 shared.cnegone = createObject(REDIS_STRING,sdsnew(":-1\r\n"));
1007 shared.nullbulk = createObject(REDIS_STRING,sdsnew("$-1\r\n"));
1008 shared.nullmultibulk = createObject(REDIS_STRING,sdsnew("*-1\r\n"));
1009 shared.emptymultibulk = createObject(REDIS_STRING,sdsnew("*0\r\n"));
1010 shared.pong = createObject(REDIS_STRING,sdsnew("+PONG\r\n"));
1011 shared.queued = createObject(REDIS_STRING,sdsnew("+QUEUED\r\n"));
1012 shared.wrongtypeerr = createObject(REDIS_STRING,sdsnew(
1013 "-ERR Operation against a key holding the wrong kind of value\r\n"));
1014 shared.nokeyerr = createObject(REDIS_STRING,sdsnew(
1015 "-ERR no such key\r\n"));
1016 shared.syntaxerr = createObject(REDIS_STRING,sdsnew(
1017 "-ERR syntax error\r\n"));
1018 shared.sameobjecterr = createObject(REDIS_STRING,sdsnew(
1019 "-ERR source and destination objects are the same\r\n"));
1020 shared.outofrangeerr = createObject(REDIS_STRING,sdsnew(
1021 "-ERR index out of range\r\n"));
1022 shared.noscripterr = createObject(REDIS_STRING,sdsnew(
1023 "-NOSCRIPT No matching script. Please use EVAL.\r\n"));
1024 shared.loadingerr = createObject(REDIS_STRING,sdsnew(
1025 "-LOADING Redis is loading the dataset in memory\r\n"));
1026 shared.slowscripterr = createObject(REDIS_STRING,sdsnew(
1027 "-BUSY Redis is busy running a script. You can only call SCRIPT KILL or SHUTDOWN NOSAVE.\r\n"));
1028 shared.masterdownerr = createObject(REDIS_STRING,sdsnew(
1029 "-MASTERDOWN Link with MASTER is down and slave-serve-stale-data is set to 'no'.\r\n"));
1030 shared.bgsaveerr = createObject(REDIS_STRING,sdsnew(
1031 "-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"));
1032 shared.roslaveerr = createObject(REDIS_STRING,sdsnew(
1033 "-READONLY You can't write against a read only slave.\r\n"));
1034 shared.oomerr = createObject(REDIS_STRING,sdsnew(
1035 "-OOM command not allowed when used memory > 'maxmemory'.\r\n"));
1036 shared.space = createObject(REDIS_STRING,sdsnew(" "));
1037 shared.colon = createObject(REDIS_STRING,sdsnew(":"));
1038 shared.plus = createObject(REDIS_STRING,sdsnew("+"));
1039
1040 for (j = 0; j < REDIS_SHARED_SELECT_CMDS; j++) {
1041 shared.select[j] = createObject(REDIS_STRING,
1042 sdscatprintf(sdsempty(),"select %d\r\n", j));
1043 }
1044 shared.messagebulk = createStringObject("$7\r\nmessage\r\n",13);
1045 shared.pmessagebulk = createStringObject("$8\r\npmessage\r\n",14);
1046 shared.subscribebulk = createStringObject("$9\r\nsubscribe\r\n",15);
1047 shared.unsubscribebulk = createStringObject("$11\r\nunsubscribe\r\n",18);
1048 shared.psubscribebulk = createStringObject("$10\r\npsubscribe\r\n",17);
1049 shared.punsubscribebulk = createStringObject("$12\r\npunsubscribe\r\n",19);
1050 shared.del = createStringObject("DEL",3);
1051 shared.rpop = createStringObject("RPOP",4);
1052 shared.lpop = createStringObject("LPOP",4);
1053 shared.lpush = createStringObject("LPUSH",5);
1054 for (j = 0; j < REDIS_SHARED_INTEGERS; j++) {
1055 shared.integers[j] = createObject(REDIS_STRING,(void*)(long)j);
1056 shared.integers[j]->encoding = REDIS_ENCODING_INT;
1057 }
1058 for (j = 0; j < REDIS_SHARED_BULKHDR_LEN; j++) {
1059 shared.mbulkhdr[j] = createObject(REDIS_STRING,
1060 sdscatprintf(sdsempty(),"*%d\r\n",j));
1061 shared.bulkhdr[j] = createObject(REDIS_STRING,
1062 sdscatprintf(sdsempty(),"$%d\r\n",j));
1063 }
1064 }
1065
1066 void initServerConfig() {
1067 getRandomHexChars(server.runid,REDIS_RUN_ID_SIZE);
1068 server.runid[REDIS_RUN_ID_SIZE] = '\0';
1069 server.arch_bits = (sizeof(long) == 8) ? 64 : 32;
1070 server.port = REDIS_SERVERPORT;
1071 server.bindaddr = NULL;
1072 server.unixsocket = NULL;
1073 server.unixsocketperm = 0;
1074 server.ipfd = -1;
1075 server.sofd = -1;
1076 server.dbnum = REDIS_DEFAULT_DBNUM;
1077 server.verbosity = REDIS_NOTICE;
1078 server.maxidletime = REDIS_MAXIDLETIME;
1079 server.client_max_querybuf_len = REDIS_MAX_QUERYBUF_LEN;
1080 server.saveparams = NULL;
1081 server.loading = 0;
1082 server.logfile = NULL; /* NULL = log on standard output */
1083 server.syslog_enabled = 0;
1084 server.syslog_ident = zstrdup("redis");
1085 server.syslog_facility = LOG_LOCAL0;
1086 server.daemonize = 0;
1087 server.aof_state = REDIS_AOF_OFF;
1088 server.aof_fsync = AOF_FSYNC_EVERYSEC;
1089 server.aof_no_fsync_on_rewrite = 0;
1090 server.aof_rewrite_perc = REDIS_AOF_REWRITE_PERC;
1091 server.aof_rewrite_min_size = REDIS_AOF_REWRITE_MIN_SIZE;
1092 server.aof_rewrite_base_size = 0;
1093 server.aof_rewrite_scheduled = 0;
1094 server.aof_last_fsync = time(NULL);
1095 server.aof_rewrite_time_last = -1;
1096 server.aof_rewrite_time_start = -1;
1097 server.aof_lastbgrewrite_status = REDIS_OK;
1098 server.aof_delayed_fsync = 0;
1099 server.aof_fd = -1;
1100 server.aof_selected_db = -1; /* Make sure the first time will not match */
1101 server.aof_flush_postponed_start = 0;
1102 server.pidfile = zstrdup("/var/run/redis.pid");
1103 server.rdb_filename = zstrdup("dump.rdb");
1104 server.aof_filename = zstrdup("appendonly.aof");
1105 server.requirepass = NULL;
1106 server.rdb_compression = 1;
1107 server.rdb_checksum = 1;
1108 server.activerehashing = 1;
1109 server.maxclients = REDIS_MAX_CLIENTS;
1110 server.bpop_blocked_clients = 0;
1111 server.maxmemory = 0;
1112 server.maxmemory_policy = REDIS_MAXMEMORY_VOLATILE_LRU;
1113 server.maxmemory_samples = 3;
1114 server.hash_max_ziplist_entries = REDIS_HASH_MAX_ZIPLIST_ENTRIES;
1115 server.hash_max_ziplist_value = REDIS_HASH_MAX_ZIPLIST_VALUE;
1116 server.list_max_ziplist_entries = REDIS_LIST_MAX_ZIPLIST_ENTRIES;
1117 server.list_max_ziplist_value = REDIS_LIST_MAX_ZIPLIST_VALUE;
1118 server.set_max_intset_entries = REDIS_SET_MAX_INTSET_ENTRIES;
1119 server.zset_max_ziplist_entries = REDIS_ZSET_MAX_ZIPLIST_ENTRIES;
1120 server.zset_max_ziplist_value = REDIS_ZSET_MAX_ZIPLIST_VALUE;
1121 server.shutdown_asap = 0;
1122 server.repl_ping_slave_period = REDIS_REPL_PING_SLAVE_PERIOD;
1123 server.repl_timeout = REDIS_REPL_TIMEOUT;
1124 server.lua_caller = NULL;
1125 server.lua_time_limit = REDIS_LUA_TIME_LIMIT;
1126 server.lua_client = NULL;
1127 server.lua_timedout = 0;
1128
1129 updateLRUClock();
1130 resetServerSaveParams();
1131
1132 appendServerSaveParams(60*60,1); /* save after 1 hour and 1 change */
1133 appendServerSaveParams(300,100); /* save after 5 minutes and 100 changes */
1134 appendServerSaveParams(60,10000); /* save after 1 minute and 10000 changes */
1135 /* Replication related */
1136 server.masterauth = NULL;
1137 server.masterhost = NULL;
1138 server.masterport = 6379;
1139 server.master = NULL;
1140 server.repl_state = REDIS_REPL_NONE;
1141 server.repl_syncio_timeout = REDIS_REPL_SYNCIO_TIMEOUT;
1142 server.repl_serve_stale_data = 1;
1143 server.repl_slave_ro = 1;
1144 server.repl_down_since = time(NULL);
1145 server.slave_priority = REDIS_DEFAULT_SLAVE_PRIORITY;
1146
1147 /* Client output buffer limits */
1148 server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_NORMAL].hard_limit_bytes = 0;
1149 server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_NORMAL].soft_limit_bytes = 0;
1150 server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_NORMAL].soft_limit_seconds = 0;
1151 server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_SLAVE].hard_limit_bytes = 1024*1024*256;
1152 server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_SLAVE].soft_limit_bytes = 1024*1024*64;
1153 server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_SLAVE].soft_limit_seconds = 60;
1154 server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_PUBSUB].hard_limit_bytes = 1024*1024*32;
1155 server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_PUBSUB].soft_limit_bytes = 1024*1024*8;
1156 server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_PUBSUB].soft_limit_seconds = 60;
1157
1158 /* Double constants initialization */
1159 R_Zero = 0.0;
1160 R_PosInf = 1.0/R_Zero;
1161 R_NegInf = -1.0/R_Zero;
1162 R_Nan = R_Zero/R_Zero;
1163
1164 /* Command table -- we intiialize it here as it is part of the
1165 * initial configuration, since command names may be changed via
1166 * redis.conf using the rename-command directive. */
1167 server.commands = dictCreate(&commandTableDictType,NULL);
1168 populateCommandTable();
1169 server.delCommand = lookupCommandByCString("del");
1170 server.multiCommand = lookupCommandByCString("multi");
1171 server.lpushCommand = lookupCommandByCString("lpush");
1172 server.lpopCommand = lookupCommandByCString("lpop");
1173 server.rpopCommand = lookupCommandByCString("rpop");
1174
1175 /* Slow log */
1176 server.slowlog_log_slower_than = REDIS_SLOWLOG_LOG_SLOWER_THAN;
1177 server.slowlog_max_len = REDIS_SLOWLOG_MAX_LEN;
1178
1179 /* Debugging */
1180 server.assert_failed = "<no assertion failed>";
1181 server.assert_file = "<no file>";
1182 server.assert_line = 0;
1183 server.bug_report_start = 0;
1184 server.watchdog_period = 0;
1185 }
1186
1187 /* This function will try to raise the max number of open files accordingly to
1188 * the configured max number of clients. It will also account for 32 additional
1189 * file descriptors as we need a few more for persistence, listening
1190 * sockets, log files and so forth.
1191 *
1192 * If it will not be possible to set the limit accordingly to the configured
1193 * max number of clients, the function will do the reverse setting
1194 * server.maxclients to the value that we can actually handle. */
1195 void adjustOpenFilesLimit(void) {
1196 rlim_t maxfiles = server.maxclients+32;
1197 struct rlimit limit;
1198
1199 if (getrlimit(RLIMIT_NOFILE,&limit) == -1) {
1200 redisLog(REDIS_WARNING,"Unable to obtain the current NOFILE limit (%s), assuming 1024 and setting the max clients configuration accordingly.",
1201 strerror(errno));
1202 server.maxclients = 1024-32;
1203 } else {
1204 rlim_t oldlimit = limit.rlim_cur;
1205
1206 /* Set the max number of files if the current limit is not enough
1207 * for our needs. */
1208 if (oldlimit < maxfiles) {
1209 rlim_t f;
1210
1211 f = maxfiles;
1212 while(f > oldlimit) {
1213 limit.rlim_cur = f;
1214 limit.rlim_max = f;
1215 if (setrlimit(RLIMIT_NOFILE,&limit) != -1) break;
1216 f -= 128;
1217 }
1218 if (f < oldlimit) f = oldlimit;
1219 if (f != maxfiles) {
1220 server.maxclients = f-32;
1221 redisLog(REDIS_WARNING,"Unable to set the max number of files limit to %d (%s), setting the max clients configuration to %d.",
1222 (int) maxfiles, strerror(errno), (int) server.maxclients);
1223 } else {
1224 redisLog(REDIS_NOTICE,"Max number of open files set to %d",
1225 (int) maxfiles);
1226 }
1227 }
1228 }
1229 }
1230
1231 void initServer() {
1232 int j;
1233
1234 signal(SIGHUP, SIG_IGN);
1235 signal(SIGPIPE, SIG_IGN);
1236 setupSignalHandlers();
1237
1238 if (server.syslog_enabled) {
1239 openlog(server.syslog_ident, LOG_PID | LOG_NDELAY | LOG_NOWAIT,
1240 server.syslog_facility);
1241 }
1242
1243 server.current_client = NULL;
1244 server.clients = listCreate();
1245 server.clients_to_close = listCreate();
1246 server.slaves = listCreate();
1247 server.monitors = listCreate();
1248 server.unblocked_clients = listCreate();
1249 server.ready_keys = listCreate();
1250
1251 createSharedObjects();
1252 adjustOpenFilesLimit();
1253 server.el = aeCreateEventLoop(server.maxclients+1024);
1254 server.db = zmalloc(sizeof(redisDb)*server.dbnum);
1255
1256 if (server.port != 0) {
1257 server.ipfd = anetTcpServer(server.neterr,server.port,server.bindaddr);
1258 if (server.ipfd == ANET_ERR) {
1259 redisLog(REDIS_WARNING, "Opening port %d: %s",
1260 server.port, server.neterr);
1261 exit(1);
1262 }
1263 }
1264 if (server.unixsocket != NULL) {
1265 unlink(server.unixsocket); /* don't care if this fails */
1266 server.sofd = anetUnixServer(server.neterr,server.unixsocket,server.unixsocketperm);
1267 if (server.sofd == ANET_ERR) {
1268 redisLog(REDIS_WARNING, "Opening socket: %s", server.neterr);
1269 exit(1);
1270 }
1271 }
1272 if (server.ipfd < 0 && server.sofd < 0) {
1273 redisLog(REDIS_WARNING, "Configured to not listen anywhere, exiting.");
1274 exit(1);
1275 }
1276 for (j = 0; j < server.dbnum; j++) {
1277 server.db[j].dict = dictCreate(&dbDictType,NULL);
1278 server.db[j].expires = dictCreate(&keyptrDictType,NULL);
1279 server.db[j].blocking_keys = dictCreate(&keylistDictType,NULL);
1280 server.db[j].ready_keys = dictCreate(&setDictType,NULL);
1281 server.db[j].watched_keys = dictCreate(&keylistDictType,NULL);
1282 server.db[j].id = j;
1283 }
1284 server.pubsub_channels = dictCreate(&keylistDictType,NULL);
1285 server.pubsub_patterns = listCreate();
1286 listSetFreeMethod(server.pubsub_patterns,freePubsubPattern);
1287 listSetMatchMethod(server.pubsub_patterns,listMatchPubsubPattern);
1288 server.cronloops = 0;
1289 server.rdb_child_pid = -1;
1290 server.aof_child_pid = -1;
1291 aofRewriteBufferReset();
1292 server.aof_buf = sdsempty();
1293 server.lastsave = time(NULL);
1294 server.rdb_save_time_last = -1;
1295 server.rdb_save_time_start = -1;
1296 server.dirty = 0;
1297 server.stat_numcommands = 0;
1298 server.stat_numconnections = 0;
1299 server.stat_expiredkeys = 0;
1300 server.stat_evictedkeys = 0;
1301 server.stat_starttime = time(NULL);
1302 server.stat_keyspace_misses = 0;
1303 server.stat_keyspace_hits = 0;
1304 server.stat_peak_memory = 0;
1305 server.stat_fork_time = 0;
1306 server.stat_rejected_conn = 0;
1307 memset(server.ops_sec_samples,0,sizeof(server.ops_sec_samples));
1308 server.ops_sec_idx = 0;
1309 server.ops_sec_last_sample_time = mstime();
1310 server.ops_sec_last_sample_ops = 0;
1311 server.unixtime = time(NULL);
1312 server.lastbgsave_status = REDIS_OK;
1313 server.stop_writes_on_bgsave_err = 1;
1314 aeCreateTimeEvent(server.el, 1, serverCron, NULL, NULL);
1315 if (server.ipfd > 0 && aeCreateFileEvent(server.el,server.ipfd,AE_READABLE,
1316 acceptTcpHandler,NULL) == AE_ERR) redisPanic("Unrecoverable error creating server.ipfd file event.");
1317 if (server.sofd > 0 && aeCreateFileEvent(server.el,server.sofd,AE_READABLE,
1318 acceptUnixHandler,NULL) == AE_ERR) redisPanic("Unrecoverable error creating server.sofd file event.");
1319
1320 if (server.aof_state == REDIS_AOF_ON) {
1321 server.aof_fd = open(server.aof_filename,
1322 O_WRONLY|O_APPEND|O_CREAT,0644);
1323 if (server.aof_fd == -1) {
1324 redisLog(REDIS_WARNING, "Can't open the append-only file: %s",
1325 strerror(errno));
1326 exit(1);
1327 }
1328 }
1329
1330 /* 32 bit instances are limited to 4GB of address space, so if there is
1331 * no explicit limit in the user provided configuration we set a limit
1332 * at 3 GB using maxmemory with 'noeviction' policy'. This avoids
1333 * useless crashes of the Redis instance for out of memory. */
1334 if (server.arch_bits == 32 && server.maxmemory == 0) {
1335 redisLog(REDIS_WARNING,"Warning: 32 bit instance detected but no memory limit set. Setting 3 GB maxmemory limit with 'noeviction' policy now.");
1336 server.maxmemory = 3072LL*(1024*1024); /* 3 GB */
1337 server.maxmemory_policy = REDIS_MAXMEMORY_NO_EVICTION;
1338 }
1339
1340 scriptingInit();
1341 slowlogInit();
1342 bioInit();
1343 }
1344
1345 /* Populates the Redis Command Table starting from the hard coded list
1346 * we have on top of redis.c file. */
1347 void populateCommandTable(void) {
1348 int j;
1349 int numcommands = sizeof(redisCommandTable)/sizeof(struct redisCommand);
1350
1351 for (j = 0; j < numcommands; j++) {
1352 struct redisCommand *c = redisCommandTable+j;
1353 char *f = c->sflags;
1354 int retval;
1355
1356 while(*f != '\0') {
1357 switch(*f) {
1358 case 'w': c->flags |= REDIS_CMD_WRITE; break;
1359 case 'r': c->flags |= REDIS_CMD_READONLY; break;
1360 case 'm': c->flags |= REDIS_CMD_DENYOOM; break;
1361 case 'a': c->flags |= REDIS_CMD_ADMIN; break;
1362 case 'p': c->flags |= REDIS_CMD_PUBSUB; break;
1363 case 'f': c->flags |= REDIS_CMD_FORCE_REPLICATION; break;
1364 case 's': c->flags |= REDIS_CMD_NOSCRIPT; break;
1365 case 'R': c->flags |= REDIS_CMD_RANDOM; break;
1366 case 'S': c->flags |= REDIS_CMD_SORT_FOR_SCRIPT; break;
1367 case 'l': c->flags |= REDIS_CMD_LOADING; break;
1368 case 't': c->flags |= REDIS_CMD_STALE; break;
1369 case 'M': c->flags |= REDIS_CMD_SKIP_MONITOR; break;
1370 default: redisPanic("Unsupported command flag"); break;
1371 }
1372 f++;
1373 }
1374
1375 retval = dictAdd(server.commands, sdsnew(c->name), c);
1376 assert(retval == DICT_OK);
1377 }
1378 }
1379
1380 void resetCommandTableStats(void) {
1381 int numcommands = sizeof(redisCommandTable)/sizeof(struct redisCommand);
1382 int j;
1383
1384 for (j = 0; j < numcommands; j++) {
1385 struct redisCommand *c = redisCommandTable+j;
1386
1387 c->microseconds = 0;
1388 c->calls = 0;
1389 }
1390 }
1391
1392 /* ========================== Redis OP Array API ============================ */
1393
1394 void redisOpArrayInit(redisOpArray *oa) {
1395 oa->ops = NULL;
1396 oa->numops = 0;
1397 }
1398
1399 int redisOpArrayAppend(redisOpArray *oa, struct redisCommand *cmd, int dbid,
1400 robj **argv, int argc, int target)
1401 {
1402 redisOp *op;
1403
1404 oa->ops = zrealloc(oa->ops,sizeof(redisOp)*(oa->numops+1));
1405 op = oa->ops+oa->numops;
1406 op->cmd = cmd;
1407 op->dbid = dbid;
1408 op->argv = argv;
1409 op->argc = argc;
1410 op->target = target;
1411 oa->numops++;
1412 return oa->numops;
1413 }
1414
1415 void redisOpArrayFree(redisOpArray *oa) {
1416 while(oa->numops) {
1417 int j;
1418 redisOp *op;
1419
1420 oa->numops--;
1421 op = oa->ops+oa->numops;
1422 for (j = 0; j < op->argc; j++)
1423 decrRefCount(op->argv[j]);
1424 zfree(op->argv);
1425 }
1426 zfree(oa->ops);
1427 }
1428
1429 /* ====================== Commands lookup and execution ===================== */
1430
1431 struct redisCommand *lookupCommand(sds name) {
1432 return dictFetchValue(server.commands, name);
1433 }
1434
1435 struct redisCommand *lookupCommandByCString(char *s) {
1436 struct redisCommand *cmd;
1437 sds name = sdsnew(s);
1438
1439 cmd = dictFetchValue(server.commands, name);
1440 sdsfree(name);
1441 return cmd;
1442 }
1443
1444 /* Propagate the specified command (in the context of the specified database id)
1445 * to AOF and Slaves.
1446 *
1447 * flags are an xor between:
1448 * + REDIS_PROPAGATE_NONE (no propagation of command at all)
1449 * + REDIS_PROPAGATE_AOF (propagate into the AOF file if is enabled)
1450 * + REDIS_PROPAGATE_REPL (propagate into the replication link)
1451 */
1452 void propagate(struct redisCommand *cmd, int dbid, robj **argv, int argc,
1453 int flags)
1454 {
1455 if (server.aof_state != REDIS_AOF_OFF && flags & REDIS_PROPAGATE_AOF)
1456 feedAppendOnlyFile(cmd,dbid,argv,argc);
1457 if (flags & REDIS_PROPAGATE_REPL && listLength(server.slaves))
1458 replicationFeedSlaves(server.slaves,dbid,argv,argc);
1459 }
1460
1461 /* Used inside commands to schedule the propagation of additional commands
1462 * after the current command is propagated to AOF / Replication. */
1463 void alsoPropagate(struct redisCommand *cmd, int dbid, robj **argv, int argc,
1464 int target)
1465 {
1466 redisOpArrayAppend(&server.also_propagate,cmd,dbid,argv,argc,target);
1467 }
1468
1469 /* Call() is the core of Redis execution of a command */
1470 void call(redisClient *c, int flags) {
1471 long long dirty, start = ustime(), duration;
1472
1473 /* Sent the command to clients in MONITOR mode, only if the commands are
1474 * not geneated from reading an AOF. */
1475 if (listLength(server.monitors) &&
1476 !server.loading &&
1477 !(c->cmd->flags & REDIS_CMD_SKIP_MONITOR))
1478 {
1479 replicationFeedMonitors(c,server.monitors,c->db->id,c->argv,c->argc);
1480 }
1481
1482 /* Call the command. */
1483 redisOpArrayInit(&server.also_propagate);
1484 dirty = server.dirty;
1485 c->cmd->proc(c);
1486 dirty = server.dirty-dirty;
1487 duration = ustime()-start;
1488
1489 /* When EVAL is called loading the AOF we don't want commands called
1490 * from Lua to go into the slowlog or to populate statistics. */
1491 if (server.loading && c->flags & REDIS_LUA_CLIENT)
1492 flags &= ~(REDIS_CALL_SLOWLOG | REDIS_CALL_STATS);
1493
1494 /* Log the command into the Slow log if needed, and populate the
1495 * per-command statistics that we show in INFO commandstats. */
1496 if (flags & REDIS_CALL_SLOWLOG)
1497 slowlogPushEntryIfNeeded(c->argv,c->argc,duration);
1498 if (flags & REDIS_CALL_STATS) {
1499 c->cmd->microseconds += duration;
1500 c->cmd->calls++;
1501 }
1502
1503 /* Propagate the command into the AOF and replication link */
1504 if (flags & REDIS_CALL_PROPAGATE) {
1505 int flags = REDIS_PROPAGATE_NONE;
1506
1507 if (c->cmd->flags & REDIS_CMD_FORCE_REPLICATION)
1508 flags |= REDIS_PROPAGATE_REPL;
1509 if (dirty)
1510 flags |= (REDIS_PROPAGATE_REPL | REDIS_PROPAGATE_AOF);
1511 if (flags != REDIS_PROPAGATE_NONE)
1512 propagate(c->cmd,c->db->id,c->argv,c->argc,flags);
1513 }
1514 /* Commands such as LPUSH or BRPOPLPUSH may propagate an additional
1515 * PUSH command. */
1516 if (server.also_propagate.numops) {
1517 int j;
1518 redisOp *rop;
1519
1520 for (j = 0; j < server.also_propagate.numops; j++) {
1521 rop = &server.also_propagate.ops[j];
1522 propagate(rop->cmd, rop->dbid, rop->argv, rop->argc, rop->target);
1523 }
1524 redisOpArrayFree(&server.also_propagate);
1525 }
1526 server.stat_numcommands++;
1527 }
1528
1529 /* If this function gets called we already read a whole
1530 * command, argments are in the client argv/argc fields.
1531 * processCommand() execute the command or prepare the
1532 * server for a bulk read from the client.
1533 *
1534 * If 1 is returned the client is still alive and valid and
1535 * and other operations can be performed by the caller. Otherwise
1536 * if 0 is returned the client was destroied (i.e. after QUIT). */
1537 int processCommand(redisClient *c) {
1538 /* The QUIT command is handled separately. Normal command procs will
1539 * go through checking for replication and QUIT will cause trouble
1540 * when FORCE_REPLICATION is enabled and would be implemented in
1541 * a regular command proc. */
1542 if (!strcasecmp(c->argv[0]->ptr,"quit")) {
1543 addReply(c,shared.ok);
1544 c->flags |= REDIS_CLOSE_AFTER_REPLY;
1545 return REDIS_ERR;
1546 }
1547
1548 /* Now lookup the command and check ASAP about trivial error conditions
1549 * such as wrong arity, bad command name and so forth. */
1550 c->cmd = c->lastcmd = lookupCommand(c->argv[0]->ptr);
1551 if (!c->cmd) {
1552 addReplyErrorFormat(c,"unknown command '%s'",
1553 (char*)c->argv[0]->ptr);
1554 return REDIS_OK;
1555 } else if ((c->cmd->arity > 0 && c->cmd->arity != c->argc) ||
1556 (c->argc < -c->cmd->arity)) {
1557 addReplyErrorFormat(c,"wrong number of arguments for '%s' command",
1558 c->cmd->name);
1559 return REDIS_OK;
1560 }
1561
1562 /* Check if the user is authenticated */
1563 if (server.requirepass && !c->authenticated && c->cmd->proc != authCommand)
1564 {
1565 addReplyError(c,"operation not permitted");
1566 return REDIS_OK;
1567 }
1568
1569 /* Handle the maxmemory directive.
1570 *
1571 * First we try to free some memory if possible (if there are volatile
1572 * keys in the dataset). If there are not the only thing we can do
1573 * is returning an error. */
1574 if (server.maxmemory) {
1575 int retval = freeMemoryIfNeeded();
1576 if ((c->cmd->flags & REDIS_CMD_DENYOOM) && retval == REDIS_ERR) {
1577 addReply(c, shared.oomerr);
1578 return REDIS_OK;
1579 }
1580 }
1581
1582 /* Don't accept write commands if there are problems persisting on disk. */
1583 if (server.stop_writes_on_bgsave_err &&
1584 server.saveparamslen > 0
1585 && server.lastbgsave_status == REDIS_ERR &&
1586 c->cmd->flags & REDIS_CMD_WRITE)
1587 {
1588 addReply(c, shared.bgsaveerr);
1589 return REDIS_OK;
1590 }
1591
1592 /* Don't accept write commands if this is a read only slave. But
1593 * accept write commands if this is our master. */
1594 if (server.masterhost && server.repl_slave_ro &&
1595 !(c->flags & REDIS_MASTER) &&
1596 c->cmd->flags & REDIS_CMD_WRITE)
1597 {
1598 addReply(c, shared.roslaveerr);
1599 return REDIS_OK;
1600 }
1601
1602 /* Only allow SUBSCRIBE and UNSUBSCRIBE in the context of Pub/Sub */
1603 if ((dictSize(c->pubsub_channels) > 0 || listLength(c->pubsub_patterns) > 0)
1604 &&
1605 c->cmd->proc != subscribeCommand &&
1606 c->cmd->proc != unsubscribeCommand &&
1607 c->cmd->proc != psubscribeCommand &&
1608 c->cmd->proc != punsubscribeCommand) {
1609 addReplyError(c,"only (P)SUBSCRIBE / (P)UNSUBSCRIBE / QUIT allowed in this context");
1610 return REDIS_OK;
1611 }
1612
1613 /* Only allow INFO and SLAVEOF when slave-serve-stale-data is no and
1614 * we are a slave with a broken link with master. */
1615 if (server.masterhost && server.repl_state != REDIS_REPL_CONNECTED &&
1616 server.repl_serve_stale_data == 0 &&
1617 !(c->cmd->flags & REDIS_CMD_STALE))
1618 {
1619 addReply(c, shared.masterdownerr);
1620 return REDIS_OK;
1621 }
1622
1623 /* Loading DB? Return an error if the command has not the
1624 * REDIS_CMD_LOADING flag. */
1625 if (server.loading && !(c->cmd->flags & REDIS_CMD_LOADING)) {
1626 addReply(c, shared.loadingerr);
1627 return REDIS_OK;
1628 }
1629
1630 /* Lua script too slow? Only allow commands with REDIS_CMD_STALE flag. */
1631 if (server.lua_timedout &&
1632 c->cmd->proc != authCommand &&
1633 !(c->cmd->proc == shutdownCommand &&
1634 c->argc == 2 &&
1635 tolower(((char*)c->argv[1]->ptr)[0]) == 'n') &&
1636 !(c->cmd->proc == scriptCommand &&
1637 c->argc == 2 &&
1638 tolower(((char*)c->argv[1]->ptr)[0]) == 'k'))
1639 {
1640 addReply(c, shared.slowscripterr);
1641 return REDIS_OK;
1642 }
1643
1644 /* Exec the command */
1645 if (c->flags & REDIS_MULTI &&
1646 c->cmd->proc != execCommand && c->cmd->proc != discardCommand &&
1647 c->cmd->proc != multiCommand && c->cmd->proc != watchCommand)
1648 {
1649 queueMultiCommand(c);
1650 addReply(c,shared.queued);
1651 } else {
1652 call(c,REDIS_CALL_FULL);
1653 if (listLength(server.ready_keys))
1654 handleClientsBlockedOnLists();
1655 }
1656 return REDIS_OK;
1657 }
1658
1659 /*================================== Shutdown =============================== */
1660
1661 int prepareForShutdown(int flags) {
1662 int save = flags & REDIS_SHUTDOWN_SAVE;
1663 int nosave = flags & REDIS_SHUTDOWN_NOSAVE;
1664
1665 redisLog(REDIS_WARNING,"User requested shutdown...");
1666 /* Kill the saving child if there is a background saving in progress.
1667 We want to avoid race conditions, for instance our saving child may
1668 overwrite the synchronous saving did by SHUTDOWN. */
1669 if (server.rdb_child_pid != -1) {
1670 redisLog(REDIS_WARNING,"There is a child saving an .rdb. Killing it!");
1671 kill(server.rdb_child_pid,SIGKILL);
1672 rdbRemoveTempFile(server.rdb_child_pid);
1673 }
1674 if (server.aof_state != REDIS_AOF_OFF) {
1675 /* Kill the AOF saving child as the AOF we already have may be longer
1676 * but contains the full dataset anyway. */
1677 if (server.aof_child_pid != -1) {
1678 redisLog(REDIS_WARNING,
1679 "There is a child rewriting the AOF. Killing it!");
1680 kill(server.aof_child_pid,SIGKILL);
1681 }
1682 /* Append only file: fsync() the AOF and exit */
1683 redisLog(REDIS_NOTICE,"Calling fsync() on the AOF file.");
1684 aof_fsync(server.aof_fd);
1685 }
1686 if ((server.saveparamslen > 0 && !nosave) || save) {
1687 redisLog(REDIS_NOTICE,"Saving the final RDB snapshot before exiting.");
1688 /* Snapshotting. Perform a SYNC SAVE and exit */
1689 if (rdbSave(server.rdb_filename) != REDIS_OK) {
1690 /* Ooops.. error saving! The best we can do is to continue
1691 * operating. Note that if there was a background saving process,
1692 * in the next cron() Redis will be notified that the background
1693 * saving aborted, handling special stuff like slaves pending for
1694 * synchronization... */
1695 redisLog(REDIS_WARNING,"Error trying to save the DB, can't exit.");
1696 return REDIS_ERR;
1697 }
1698 }
1699 if (server.daemonize) {
1700 redisLog(REDIS_NOTICE,"Removing the pid file.");
1701 unlink(server.pidfile);
1702 }
1703 /* Close the listening sockets. Apparently this allows faster restarts. */
1704 if (server.ipfd != -1) close(server.ipfd);
1705 if (server.sofd != -1) close(server.sofd);
1706 if (server.unixsocket) {
1707 redisLog(REDIS_NOTICE,"Removing the unix socket file.");
1708 unlink(server.unixsocket); /* don't care if this fails */
1709 }
1710
1711 redisLog(REDIS_WARNING,"Redis is now ready to exit, bye bye...");
1712 return REDIS_OK;
1713 }
1714
1715 /*================================== Commands =============================== */
1716
1717 /* Return zero if strings are the same, non-zero if they are not.
1718 * The comparison is performed in a way that prevents an attacker to obtain
1719 * information about the nature of the strings just monitoring the execution
1720 * time of the function.
1721 *
1722 * Note that limiting the comparison length to strings up to 512 bytes we
1723 * can avoid leaking any information about the password length and any
1724 * possible branch misprediction related leak.
1725 */
1726 int time_independent_strcmp(char *a, char *b) {
1727 char bufa[REDIS_AUTHPASS_MAX_LEN], bufb[REDIS_AUTHPASS_MAX_LEN];
1728 /* The above two strlen perform len(a) + len(b) operations where either
1729 * a or b are fixed (our password) length, and the difference is only
1730 * relative to the length of the user provided string, so no information
1731 * leak is possible in the following two lines of code. */
1732 int alen = strlen(a);
1733 int blen = strlen(b);
1734 int j;
1735 int diff = 0;
1736
1737 /* We can't compare strings longer than our static buffers.
1738 * Note that this will never pass the first test in practical circumstances
1739 * so there is no info leak. */
1740 if (alen > sizeof(bufa) || blen > sizeof(bufb)) return 1;
1741
1742 memset(bufa,0,sizeof(bufa)); /* Constant time. */
1743 memset(bufb,0,sizeof(bufb)); /* Constant time. */
1744 /* Again the time of the following two copies is proportional to
1745 * len(a) + len(b) so no info is leaked. */
1746 memcpy(bufa,a,alen);
1747 memcpy(bufb,b,blen);
1748
1749 /* Always compare all the chars in the two buffers without
1750 * conditional expressions. */
1751 for (j = 0; j < sizeof(bufa); j++) {
1752 diff |= (bufa[j] ^ bufb[j]);
1753 }
1754 /* Length must be equal as well. */
1755 diff |= alen ^ blen;
1756 return diff; /* If zero strings are the same. */
1757 }
1758
1759 void authCommand(redisClient *c) {
1760 if (!server.requirepass) {
1761 addReplyError(c,"Client sent AUTH, but no password is set");
1762 } else if (!time_independent_strcmp(c->argv[1]->ptr, server.requirepass)) {
1763 c->authenticated = 1;
1764 addReply(c,shared.ok);
1765 } else {
1766 c->authenticated = 0;
1767 addReplyError(c,"invalid password");
1768 }
1769 }
1770
1771 void pingCommand(redisClient *c) {
1772 addReply(c,shared.pong);
1773 }
1774
1775 void echoCommand(redisClient *c) {
1776 addReplyBulk(c,c->argv[1]);
1777 }
1778
1779 void timeCommand(redisClient *c) {
1780 struct timeval tv;
1781
1782 /* gettimeofday() can only fail if &tv is a bad addresss so we
1783 * don't check for errors. */
1784 gettimeofday(&tv,NULL);
1785 addReplyMultiBulkLen(c,2);
1786 addReplyBulkLongLong(c,tv.tv_sec);
1787 addReplyBulkLongLong(c,tv.tv_usec);
1788 }
1789
1790 /* Convert an amount of bytes into a human readable string in the form
1791 * of 100B, 2G, 100M, 4K, and so forth. */
1792 void bytesToHuman(char *s, unsigned long long n) {
1793 double d;
1794
1795 if (n < 1024) {
1796 /* Bytes */
1797 sprintf(s,"%lluB",n);
1798 return;
1799 } else if (n < (1024*1024)) {
1800 d = (double)n/(1024);
1801 sprintf(s,"%.2fK",d);
1802 } else if (n < (1024LL*1024*1024)) {
1803 d = (double)n/(1024*1024);
1804 sprintf(s,"%.2fM",d);
1805 } else if (n < (1024LL*1024*1024*1024)) {
1806 d = (double)n/(1024LL*1024*1024);
1807 sprintf(s,"%.2fG",d);
1808 }
1809 }
1810
1811 /* Create the string returned by the INFO command. This is decoupled
1812 * by the INFO command itself as we need to report the same information
1813 * on memory corruption problems. */
1814 sds genRedisInfoString(char *section) {
1815 sds info = sdsempty();
1816 time_t uptime = server.unixtime-server.stat_starttime;
1817 int j, numcommands;
1818 struct rusage self_ru, c_ru;
1819 unsigned long lol, bib;
1820 int allsections = 0, defsections = 0;
1821 int sections = 0;
1822
1823 if (section) {
1824 allsections = strcasecmp(section,"all") == 0;
1825 defsections = strcasecmp(section,"default") == 0;
1826 }
1827
1828 getrusage(RUSAGE_SELF, &self_ru);
1829 getrusage(RUSAGE_CHILDREN, &c_ru);
1830 getClientsMaxBuffers(&lol,&bib);
1831
1832 /* Server */
1833 if (allsections || defsections || !strcasecmp(section,"server")) {
1834 struct utsname name;
1835 char *mode;
1836
1837 if (server.sentinel_mode) mode = "sentinel";
1838 else mode = "standalone";
1839
1840 if (sections++) info = sdscat(info,"\r\n");
1841 uname(&name);
1842 info = sdscatprintf(info,
1843 "# Server\r\n"
1844 "redis_version:%s\r\n"
1845 "redis_git_sha1:%s\r\n"
1846 "redis_git_dirty:%d\r\n"
1847 "redis_mode:%s\r\n"
1848 "os:%s %s %s\r\n"
1849 "arch_bits:%d\r\n"
1850 "multiplexing_api:%s\r\n"
1851 "gcc_version:%d.%d.%d\r\n"
1852 "process_id:%ld\r\n"
1853 "run_id:%s\r\n"
1854 "tcp_port:%d\r\n"
1855 "uptime_in_seconds:%ld\r\n"
1856 "uptime_in_days:%ld\r\n"
1857 "lru_clock:%ld\r\n",
1858 REDIS_VERSION,
1859 redisGitSHA1(),
1860 strtol(redisGitDirty(),NULL,10) > 0,
1861 mode,
1862 name.sysname, name.release, name.machine,
1863 server.arch_bits,
1864 aeGetApiName(),
1865 #ifdef __GNUC__
1866 __GNUC__,__GNUC_MINOR__,__GNUC_PATCHLEVEL__,
1867 #else
1868 0,0,0,
1869 #endif
1870 (long) getpid(),
1871 server.runid,
1872 server.port,
1873 uptime,
1874 uptime/(3600*24),
1875 (unsigned long) server.lruclock);
1876 }
1877
1878 /* Clients */
1879 if (allsections || defsections || !strcasecmp(section,"clients")) {
1880 if (sections++) info = sdscat(info,"\r\n");
1881 info = sdscatprintf(info,
1882 "# Clients\r\n"
1883 "connected_clients:%lu\r\n"
1884 "client_longest_output_list:%lu\r\n"
1885 "client_biggest_input_buf:%lu\r\n"
1886 "blocked_clients:%d\r\n",
1887 listLength(server.clients)-listLength(server.slaves),
1888 lol, bib,
1889 server.bpop_blocked_clients);
1890 }
1891
1892 /* Memory */
1893 if (allsections || defsections || !strcasecmp(section,"memory")) {
1894 char hmem[64];
1895 char peak_hmem[64];
1896
1897 bytesToHuman(hmem,zmalloc_used_memory());
1898 bytesToHuman(peak_hmem,server.stat_peak_memory);
1899 if (sections++) info = sdscat(info,"\r\n");
1900 info = sdscatprintf(info,
1901 "# Memory\r\n"
1902 "used_memory:%zu\r\n"
1903 "used_memory_human:%s\r\n"
1904 "used_memory_rss:%zu\r\n"
1905 "used_memory_peak:%zu\r\n"
1906 "used_memory_peak_human:%s\r\n"
1907 "used_memory_lua:%lld\r\n"
1908 "mem_fragmentation_ratio:%.2f\r\n"
1909 "mem_allocator:%s\r\n",
1910 zmalloc_used_memory(),
1911 hmem,
1912 zmalloc_get_rss(),
1913 server.stat_peak_memory,
1914 peak_hmem,
1915 ((long long)lua_gc(server.lua,LUA_GCCOUNT,0))*1024LL,
1916 zmalloc_get_fragmentation_ratio(),
1917 ZMALLOC_LIB
1918 );
1919 }
1920
1921 /* Persistence */
1922 if (allsections || defsections || !strcasecmp(section,"persistence")) {
1923 if (sections++) info = sdscat(info,"\r\n");
1924 info = sdscatprintf(info,
1925 "# Persistence\r\n"
1926 "loading:%d\r\n"
1927 "rdb_changes_since_last_save:%lld\r\n"
1928 "rdb_bgsave_in_progress:%d\r\n"
1929 "rdb_last_save_time:%ld\r\n"
1930 "rdb_last_bgsave_status:%s\r\n"
1931 "rdb_last_bgsave_time_sec:%ld\r\n"
1932 "rdb_current_bgsave_time_sec:%ld\r\n"
1933 "aof_enabled:%d\r\n"
1934 "aof_rewrite_in_progress:%d\r\n"
1935 "aof_rewrite_scheduled:%d\r\n"
1936 "aof_last_rewrite_time_sec:%ld\r\n"
1937 "aof_current_rewrite_time_sec:%ld\r\n"
1938 "aof_last_bgrewrite_status:%s\r\n",
1939 server.loading,
1940 server.dirty,
1941 server.rdb_child_pid != -1,
1942 server.lastsave,
1943 (server.lastbgsave_status == REDIS_OK) ? "ok" : "err",
1944 server.rdb_save_time_last,
1945 (server.rdb_child_pid == -1) ?
1946 -1 : time(NULL)-server.rdb_save_time_start,
1947 server.aof_state != REDIS_AOF_OFF,
1948 server.aof_child_pid != -1,
1949 server.aof_rewrite_scheduled,
1950 server.aof_rewrite_time_last,
1951 (server.aof_child_pid == -1) ?
1952 -1 : time(NULL)-server.aof_rewrite_time_start,
1953 (server.aof_lastbgrewrite_status == REDIS_OK) ? "ok" : "err");
1954
1955 if (server.aof_state != REDIS_AOF_OFF) {
1956 info = sdscatprintf(info,
1957 "aof_current_size:%lld\r\n"
1958 "aof_base_size:%lld\r\n"
1959 "aof_pending_rewrite:%d\r\n"
1960 "aof_buffer_length:%zu\r\n"
1961 "aof_rewrite_buffer_length:%lu\r\n"
1962 "aof_pending_bio_fsync:%llu\r\n"
1963 "aof_delayed_fsync:%lu\r\n",
1964 (long long) server.aof_current_size,
1965 (long long) server.aof_rewrite_base_size,
1966 server.aof_rewrite_scheduled,
1967 sdslen(server.aof_buf),
1968 aofRewriteBufferSize(),
1969 bioPendingJobsOfType(REDIS_BIO_AOF_FSYNC),
1970 server.aof_delayed_fsync);
1971 }
1972
1973 if (server.loading) {
1974 double perc;
1975 time_t eta, elapsed;
1976 off_t remaining_bytes = server.loading_total_bytes-
1977 server.loading_loaded_bytes;
1978
1979 perc = ((double)server.loading_loaded_bytes /
1980 server.loading_total_bytes) * 100;
1981
1982 elapsed = server.unixtime-server.loading_start_time;
1983 if (elapsed == 0) {
1984 eta = 1; /* A fake 1 second figure if we don't have
1985 enough info */
1986 } else {
1987 eta = (elapsed*remaining_bytes)/server.loading_loaded_bytes;
1988 }
1989
1990 info = sdscatprintf(info,
1991 "loading_start_time:%ld\r\n"
1992 "loading_total_bytes:%llu\r\n"
1993 "loading_loaded_bytes:%llu\r\n"
1994 "loading_loaded_perc:%.2f\r\n"
1995 "loading_eta_seconds:%ld\r\n"
1996 ,(unsigned long) server.loading_start_time,
1997 (unsigned long long) server.loading_total_bytes,
1998 (unsigned long long) server.loading_loaded_bytes,
1999 perc,
2000 eta
2001 );
2002 }
2003 }
2004
2005 /* Stats */
2006 if (allsections || defsections || !strcasecmp(section,"stats")) {
2007 if (sections++) info = sdscat(info,"\r\n");
2008 info = sdscatprintf(info,
2009 "# Stats\r\n"
2010 "total_connections_received:%lld\r\n"
2011 "total_commands_processed:%lld\r\n"
2012 "instantaneous_ops_per_sec:%lld\r\n"
2013 "rejected_connections:%lld\r\n"
2014 "expired_keys:%lld\r\n"
2015 "evicted_keys:%lld\r\n"
2016 "keyspace_hits:%lld\r\n"
2017 "keyspace_misses:%lld\r\n"
2018 "pubsub_channels:%ld\r\n"
2019 "pubsub_patterns:%lu\r\n"
2020 "latest_fork_usec:%lld\r\n",
2021 server.stat_numconnections,
2022 server.stat_numcommands,
2023 getOperationsPerSecond(),
2024 server.stat_rejected_conn,
2025 server.stat_expiredkeys,
2026 server.stat_evictedkeys,
2027 server.stat_keyspace_hits,
2028 server.stat_keyspace_misses,
2029 dictSize(server.pubsub_channels),
2030 listLength(server.pubsub_patterns),
2031 server.stat_fork_time);
2032 }
2033
2034 /* Replication */
2035 if (allsections || defsections || !strcasecmp(section,"replication")) {
2036 if (sections++) info = sdscat(info,"\r\n");
2037 info = sdscatprintf(info,
2038 "# Replication\r\n"
2039 "role:%s\r\n",
2040 server.masterhost == NULL ? "master" : "slave");
2041 if (server.masterhost) {
2042 info = sdscatprintf(info,
2043 "master_host:%s\r\n"
2044 "master_port:%d\r\n"
2045 "master_link_status:%s\r\n"
2046 "master_last_io_seconds_ago:%d\r\n"
2047 "master_sync_in_progress:%d\r\n"
2048 ,server.masterhost,
2049 server.masterport,
2050 (server.repl_state == REDIS_REPL_CONNECTED) ?
2051 "up" : "down",
2052 server.master ?
2053 ((int)(server.unixtime-server.master->lastinteraction)) : -1,
2054 server.repl_state == REDIS_REPL_TRANSFER
2055 );
2056
2057 if (server.repl_state == REDIS_REPL_TRANSFER) {
2058 info = sdscatprintf(info,
2059 "master_sync_left_bytes:%lld\r\n"
2060 "master_sync_last_io_seconds_ago:%d\r\n"
2061 , (long long)
2062 (server.repl_transfer_size - server.repl_transfer_read),
2063 (int)(server.unixtime-server.repl_transfer_lastio)
2064 );
2065 }
2066
2067 if (server.repl_state != REDIS_REPL_CONNECTED) {
2068 info = sdscatprintf(info,
2069 "master_link_down_since_seconds:%ld\r\n",
2070 (long)server.unixtime-server.repl_down_since);
2071 }
2072 info = sdscatprintf(info,
2073 "slave_priority:%d\r\n"
2074 "slave_read_only:%d\r\n",
2075 server.slave_priority,
2076 server.repl_slave_ro);
2077 }
2078 info = sdscatprintf(info,
2079 "connected_slaves:%lu\r\n",
2080 listLength(server.slaves));
2081 if (listLength(server.slaves)) {
2082 int slaveid = 0;
2083 listNode *ln;
2084 listIter li;
2085
2086 listRewind(server.slaves,&li);
2087 while((ln = listNext(&li))) {
2088 redisClient *slave = listNodeValue(ln);
2089 char *state = NULL;
2090 char ip[32];
2091 int port;
2092
2093 if (anetPeerToString(slave->fd,ip,&port) == -1) continue;
2094 switch(slave->replstate) {
2095 case REDIS_REPL_WAIT_BGSAVE_START:
2096 case REDIS_REPL_WAIT_BGSAVE_END:
2097 state = "wait_bgsave";
2098 break;
2099 case REDIS_REPL_SEND_BULK:
2100 state = "send_bulk";
2101 break;
2102 case REDIS_REPL_ONLINE:
2103 state = "online";
2104 break;
2105 }
2106 if (state == NULL) continue;
2107 info = sdscatprintf(info,"slave%d:%s,%d,%s\r\n",
2108 slaveid,ip,slave->slave_listening_port,state);
2109 slaveid++;
2110 }
2111 }
2112 }
2113
2114 /* CPU */
2115 if (allsections || defsections || !strcasecmp(section,"cpu")) {
2116 if (sections++) info = sdscat(info,"\r\n");
2117 info = sdscatprintf(info,
2118 "# CPU\r\n"
2119 "used_cpu_sys:%.2f\r\n"
2120 "used_cpu_user:%.2f\r\n"
2121 "used_cpu_sys_children:%.2f\r\n"
2122 "used_cpu_user_children:%.2f\r\n",
2123 (float)self_ru.ru_stime.tv_sec+(float)self_ru.ru_stime.tv_usec/1000000,
2124 (float)self_ru.ru_utime.tv_sec+(float)self_ru.ru_utime.tv_usec/1000000,
2125 (float)c_ru.ru_stime.tv_sec+(float)c_ru.ru_stime.tv_usec/1000000,
2126 (float)c_ru.ru_utime.tv_sec+(float)c_ru.ru_utime.tv_usec/1000000);
2127 }
2128
2129 /* cmdtime */
2130 if (allsections || !strcasecmp(section,"commandstats")) {
2131 if (sections++) info = sdscat(info,"\r\n");
2132 info = sdscatprintf(info, "# Commandstats\r\n");
2133 numcommands = sizeof(redisCommandTable)/sizeof(struct redisCommand);
2134 for (j = 0; j < numcommands; j++) {
2135 struct redisCommand *c = redisCommandTable+j;
2136
2137 if (!c->calls) continue;
2138 info = sdscatprintf(info,
2139 "cmdstat_%s:calls=%lld,usec=%lld,usec_per_call=%.2f\r\n",
2140 c->name, c->calls, c->microseconds,
2141 (c->calls == 0) ? 0 : ((float)c->microseconds/c->calls));
2142 }
2143 }
2144
2145 /* Key space */
2146 if (allsections || defsections || !strcasecmp(section,"keyspace")) {
2147 if (sections++) info = sdscat(info,"\r\n");
2148 info = sdscatprintf(info, "# Keyspace\r\n");
2149 for (j = 0; j < server.dbnum; j++) {
2150 long long keys, vkeys;
2151
2152 keys = dictSize(server.db[j].dict);
2153 vkeys = dictSize(server.db[j].expires);
2154 if (keys || vkeys) {
2155 info = sdscatprintf(info, "db%d:keys=%lld,expires=%lld\r\n",
2156 j, keys, vkeys);
2157 }
2158 }
2159 }
2160 return info;
2161 }
2162
2163 void infoCommand(redisClient *c) {
2164 char *section = c->argc == 2 ? c->argv[1]->ptr : "default";
2165
2166 if (c->argc > 2) {
2167 addReply(c,shared.syntaxerr);
2168 return;
2169 }
2170 sds info = genRedisInfoString(section);
2171 addReplySds(c,sdscatprintf(sdsempty(),"$%lu\r\n",
2172 (unsigned long)sdslen(info)));
2173 addReplySds(c,info);
2174 addReply(c,shared.crlf);
2175 }
2176
2177 void monitorCommand(redisClient *c) {
2178 /* ignore MONITOR if aleady slave or in monitor mode */
2179 if (c->flags & REDIS_SLAVE) return;
2180
2181 c->flags |= (REDIS_SLAVE|REDIS_MONITOR);
2182 c->slaveseldb = 0;
2183 listAddNodeTail(server.monitors,c);
2184 addReply(c,shared.ok);
2185 }
2186
2187 /* ============================ Maxmemory directive ======================== */
2188
2189 /* This function gets called when 'maxmemory' is set on the config file to limit
2190 * the max memory used by the server, before processing a command.
2191 *
2192 * The goal of the function is to free enough memory to keep Redis under the
2193 * configured memory limit.
2194 *
2195 * The function starts calculating how many bytes should be freed to keep
2196 * Redis under the limit, and enters a loop selecting the best keys to
2197 * evict accordingly to the configured policy.
2198 *
2199 * If all the bytes needed to return back under the limit were freed the
2200 * function returns REDIS_OK, otherwise REDIS_ERR is returned, and the caller
2201 * should block the execution of commands that will result in more memory
2202 * used by the server.
2203 */
2204 int freeMemoryIfNeeded(void) {
2205 size_t mem_used, mem_tofree, mem_freed;
2206 int slaves = listLength(server.slaves);
2207
2208 /* Remove the size of slaves output buffers and AOF buffer from the
2209 * count of used memory. */
2210 mem_used = zmalloc_used_memory();
2211 if (slaves) {
2212 listIter li;
2213 listNode *ln;
2214
2215 listRewind(server.slaves,&li);
2216 while((ln = listNext(&li))) {
2217 redisClient *slave = listNodeValue(ln);
2218 unsigned long obuf_bytes = getClientOutputBufferMemoryUsage(slave);
2219 if (obuf_bytes > mem_used)
2220 mem_used = 0;
2221 else
2222 mem_used -= obuf_bytes;
2223 }
2224 }
2225 if (server.aof_state != REDIS_AOF_OFF) {
2226 mem_used -= sdslen(server.aof_buf);
2227 mem_used -= aofRewriteBufferSize();
2228 }
2229
2230 /* Check if we are over the memory limit. */
2231 if (mem_used <= server.maxmemory) return REDIS_OK;
2232
2233 if (server.maxmemory_policy == REDIS_MAXMEMORY_NO_EVICTION)
2234 return REDIS_ERR; /* We need to free memory, but policy forbids. */
2235
2236 /* Compute how much memory we need to free. */
2237 mem_tofree = mem_used - server.maxmemory;
2238 mem_freed = 0;
2239 while (mem_freed < mem_tofree) {
2240 int j, k, keys_freed = 0;
2241
2242 for (j = 0; j < server.dbnum; j++) {
2243 long bestval = 0; /* just to prevent warning */
2244 sds bestkey = NULL;
2245 struct dictEntry *de;
2246 redisDb *db = server.db+j;
2247 dict *dict;
2248
2249 if (server.maxmemory_policy == REDIS_MAXMEMORY_ALLKEYS_LRU ||
2250 server.maxmemory_policy == REDIS_MAXMEMORY_ALLKEYS_RANDOM)
2251 {
2252 dict = server.db[j].dict;
2253 } else {
2254 dict = server.db[j].expires;
2255 }
2256 if (dictSize(dict) == 0) continue;
2257
2258 /* volatile-random and allkeys-random policy */
2259 if (server.maxmemory_policy == REDIS_MAXMEMORY_ALLKEYS_RANDOM ||
2260 server.maxmemory_policy == REDIS_MAXMEMORY_VOLATILE_RANDOM)
2261 {
2262 de = dictGetRandomKey(dict);
2263 bestkey = dictGetKey(de);
2264 }
2265
2266 /* volatile-lru and allkeys-lru policy */
2267 else if (server.maxmemory_policy == REDIS_MAXMEMORY_ALLKEYS_LRU ||
2268 server.maxmemory_policy == REDIS_MAXMEMORY_VOLATILE_LRU)
2269 {
2270 for (k = 0; k < server.maxmemory_samples; k++) {
2271 sds thiskey;
2272 long thisval;
2273 robj *o;
2274
2275 de = dictGetRandomKey(dict);
2276 thiskey = dictGetKey(de);
2277 /* When policy is volatile-lru we need an additonal lookup
2278 * to locate the real key, as dict is set to db->expires. */
2279 if (server.maxmemory_policy == REDIS_MAXMEMORY_VOLATILE_LRU)
2280 de = dictFind(db->dict, thiskey);
2281 o = dictGetVal(de);
2282 thisval = estimateObjectIdleTime(o);
2283
2284 /* Higher idle time is better candidate for deletion */
2285 if (bestkey == NULL || thisval > bestval) {
2286 bestkey = thiskey;
2287 bestval = thisval;
2288 }
2289 }
2290 }
2291
2292 /* volatile-ttl */
2293 else if (server.maxmemory_policy == REDIS_MAXMEMORY_VOLATILE_TTL) {
2294 for (k = 0; k < server.maxmemory_samples; k++) {
2295 sds thiskey;
2296 long thisval;
2297
2298 de = dictGetRandomKey(dict);
2299 thiskey = dictGetKey(de);
2300 thisval = (long) dictGetVal(de);
2301
2302 /* Expire sooner (minor expire unix timestamp) is better
2303 * candidate for deletion */
2304 if (bestkey == NULL || thisval < bestval) {
2305 bestkey = thiskey;
2306 bestval = thisval;
2307 }
2308 }
2309 }
2310
2311 /* Finally remove the selected key. */
2312 if (bestkey) {
2313 long long delta;
2314
2315 robj *keyobj = createStringObject(bestkey,sdslen(bestkey));
2316 propagateExpire(db,keyobj);
2317 /* We compute the amount of memory freed by dbDelete() alone.
2318 * It is possible that actually the memory needed to propagate
2319 * the DEL in AOF and replication link is greater than the one
2320 * we are freeing removing the key, but we can't account for
2321 * that otherwise we would never exit the loop.
2322 *
2323 * AOF and Output buffer memory will be freed eventually so
2324 * we only care about memory used by the key space. */
2325 delta = (long long) zmalloc_used_memory();
2326 dbDelete(db,keyobj);
2327 delta -= (long long) zmalloc_used_memory();
2328 mem_freed += delta;
2329 server.stat_evictedkeys++;
2330 decrRefCount(keyobj);
2331 keys_freed++;
2332
2333 /* When the memory to free starts to be big enough, we may
2334 * start spending so much time here that is impossible to
2335 * deliver data to the slaves fast enough, so we force the
2336 * transmission here inside the loop. */
2337 if (slaves) flushSlavesOutputBuffers();
2338 }
2339 }
2340 if (!keys_freed) return REDIS_ERR; /* nothing to free... */
2341 }
2342 return REDIS_OK;
2343 }
2344
2345 /* =================================== Main! ================================ */
2346
2347 #ifdef __linux__
2348 int linuxOvercommitMemoryValue(void) {
2349 FILE *fp = fopen("/proc/sys/vm/overcommit_memory","r");
2350 char buf[64];
2351
2352 if (!fp) return -1;
2353 if (fgets(buf,64,fp) == NULL) {
2354 fclose(fp);
2355 return -1;
2356 }
2357 fclose(fp);
2358
2359 return atoi(buf);
2360 }
2361
2362 void linuxOvercommitMemoryWarning(void) {
2363 if (linuxOvercommitMemoryValue() == 0) {
2364 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.");
2365 }
2366 }
2367 #endif /* __linux__ */
2368
2369 void createPidFile(void) {
2370 /* Try to write the pid file in a best-effort way. */
2371 FILE *fp = fopen(server.pidfile,"w");
2372 if (fp) {
2373 fprintf(fp,"%d\n",(int)getpid());
2374 fclose(fp);
2375 }
2376 }
2377
2378 void daemonize(void) {
2379 int fd;
2380
2381 if (fork() != 0) exit(0); /* parent exits */
2382 setsid(); /* create a new session */
2383
2384 /* Every output goes to /dev/null. If Redis is daemonized but
2385 * the 'logfile' is set to 'stdout' in the configuration file
2386 * it will not log at all. */
2387 if ((fd = open("/dev/null", O_RDWR, 0)) != -1) {
2388 dup2(fd, STDIN_FILENO);
2389 dup2(fd, STDOUT_FILENO);
2390 dup2(fd, STDERR_FILENO);
2391 if (fd > STDERR_FILENO) close(fd);
2392 }
2393 }
2394
2395 void version() {
2396 printf("Redis server v=%s sha=%s:%d malloc=%s bits=%d\n",
2397 REDIS_VERSION,
2398 redisGitSHA1(),
2399 atoi(redisGitDirty()) > 0,
2400 ZMALLOC_LIB,
2401 sizeof(long) == 4 ? 32 : 64);
2402 exit(0);
2403 }
2404
2405 void usage() {
2406 fprintf(stderr,"Usage: ./redis-server [/path/to/redis.conf] [options]\n");
2407 fprintf(stderr," ./redis-server - (read config from stdin)\n");
2408 fprintf(stderr," ./redis-server -v or --version\n");
2409 fprintf(stderr," ./redis-server -h or --help\n");
2410 fprintf(stderr," ./redis-server --test-memory <megabytes>\n\n");
2411 fprintf(stderr,"Examples:\n");
2412 fprintf(stderr," ./redis-server (run the server with default conf)\n");
2413 fprintf(stderr," ./redis-server /etc/redis/6379.conf\n");
2414 fprintf(stderr," ./redis-server --port 7777\n");
2415 fprintf(stderr," ./redis-server --port 7777 --slaveof 127.0.0.1 8888\n");
2416 fprintf(stderr," ./redis-server /etc/myredis.conf --loglevel verbose\n\n");
2417 fprintf(stderr,"Sentinel mode:\n");
2418 fprintf(stderr," ./redis-server /etc/sentinel.conf --sentinel\n");
2419 exit(1);
2420 }
2421
2422 void redisAsciiArt(void) {
2423 #include "asciilogo.h"
2424 char *buf = zmalloc(1024*16);
2425 char *mode = "stand alone";
2426
2427 if (server.sentinel_mode) mode = "sentinel";
2428
2429 snprintf(buf,1024*16,ascii_logo,
2430 REDIS_VERSION,
2431 redisGitSHA1(),
2432 strtol(redisGitDirty(),NULL,10) > 0,
2433 (sizeof(long) == 8) ? "64" : "32",
2434 mode, server.port,
2435 (long) getpid()
2436 );
2437 redisLogRaw(REDIS_NOTICE|REDIS_LOG_RAW,buf);
2438 zfree(buf);
2439 }
2440
2441 static void sigtermHandler(int sig) {
2442 REDIS_NOTUSED(sig);
2443
2444 redisLogFromHandler(REDIS_WARNING,"Received SIGTERM, scheduling shutdown...");
2445 server.shutdown_asap = 1;
2446 }
2447
2448 void setupSignalHandlers(void) {
2449 struct sigaction act;
2450
2451 /* When the SA_SIGINFO flag is set in sa_flags then sa_sigaction is used.
2452 * Otherwise, sa_handler is used. */
2453 sigemptyset(&act.sa_mask);
2454 act.sa_flags = 0;
2455 act.sa_handler = sigtermHandler;
2456 sigaction(SIGTERM, &act, NULL);
2457
2458 #ifdef HAVE_BACKTRACE
2459 sigemptyset(&act.sa_mask);
2460 act.sa_flags = SA_NODEFER | SA_RESETHAND | SA_SIGINFO;
2461 act.sa_sigaction = sigsegvHandler;
2462 sigaction(SIGSEGV, &act, NULL);
2463 sigaction(SIGBUS, &act, NULL);
2464 sigaction(SIGFPE, &act, NULL);
2465 sigaction(SIGILL, &act, NULL);
2466 #endif
2467 return;
2468 }
2469
2470 void memtest(size_t megabytes, int passes);
2471
2472 /* Returns 1 if there is --sentinel among the arguments or if
2473 * argv[0] is exactly "redis-sentinel". */
2474 int checkForSentinelMode(int argc, char **argv) {
2475 int j;
2476
2477 if (strstr(argv[0],"redis-sentinel") != NULL) return 1;
2478 for (j = 1; j < argc; j++)
2479 if (!strcmp(argv[j],"--sentinel")) return 1;
2480 return 0;
2481 }
2482
2483 /* Function called at startup to load RDB or AOF file in memory. */
2484 void loadDataFromDisk(void) {
2485 long long start = ustime();
2486 if (server.aof_state == REDIS_AOF_ON) {
2487 if (loadAppendOnlyFile(server.aof_filename) == REDIS_OK)
2488 redisLog(REDIS_NOTICE,"DB loaded from append only file: %.3f seconds",(float)(ustime()-start)/1000000);
2489 } else {
2490 if (rdbLoad(server.rdb_filename) == REDIS_OK) {
2491 redisLog(REDIS_NOTICE,"DB loaded from disk: %.3f seconds",
2492 (float)(ustime()-start)/1000000);
2493 } else if (errno != ENOENT) {
2494 redisLog(REDIS_WARNING,"Fatal error loading the DB. Exiting.");
2495 exit(1);
2496 }
2497 }
2498 }
2499
2500 void redisOutOfMemoryHandler(size_t allocation_size) {
2501 redisLog(REDIS_WARNING,"Out Of Memory allocating %zu bytes!",
2502 allocation_size);
2503 redisPanic("OOM");
2504 }
2505
2506 int main(int argc, char **argv) {
2507 struct timeval tv;
2508
2509 /* We need to initialize our libraries, and the server configuration. */
2510 zmalloc_enable_thread_safeness();
2511 zmalloc_set_oom_handler(redisOutOfMemoryHandler);
2512 srand(time(NULL)^getpid());
2513 gettimeofday(&tv,NULL);
2514 dictSetHashFunctionSeed(tv.tv_sec^tv.tv_usec^getpid());
2515 server.sentinel_mode = checkForSentinelMode(argc,argv);
2516 initServerConfig();
2517
2518 /* We need to init sentinel right now as parsing the configuration file
2519 * in sentinel mode will have the effect of populating the sentinel
2520 * data structures with master nodes to monitor. */
2521 if (server.sentinel_mode) {
2522 initSentinelConfig();
2523 initSentinel();
2524 }
2525
2526 if (argc >= 2) {
2527 int j = 1; /* First option to parse in argv[] */
2528 sds options = sdsempty();
2529 char *configfile = NULL;
2530
2531 /* Handle special options --help and --version */
2532 if (strcmp(argv[1], "-v") == 0 ||
2533 strcmp(argv[1], "--version") == 0) version();
2534 if (strcmp(argv[1], "--help") == 0 ||
2535 strcmp(argv[1], "-h") == 0) usage();
2536 if (strcmp(argv[1], "--test-memory") == 0) {
2537 if (argc == 3) {
2538 memtest(atoi(argv[2]),50);
2539 exit(0);
2540 } else {
2541 fprintf(stderr,"Please specify the amount of memory to test in megabytes.\n");
2542 fprintf(stderr,"Example: ./redis-server --test-memory 4096\n\n");
2543 exit(1);
2544 }
2545 }
2546
2547 /* First argument is the config file name? */
2548 if (argv[j][0] != '-' || argv[j][1] != '-')
2549 configfile = argv[j++];
2550 /* All the other options are parsed and conceptually appended to the
2551 * configuration file. For instance --port 6380 will generate the
2552 * string "port 6380\n" to be parsed after the actual file name
2553 * is parsed, if any. */
2554 while(j != argc) {
2555 if (argv[j][0] == '-' && argv[j][1] == '-') {
2556 /* Option name */
2557 if (sdslen(options)) options = sdscat(options,"\n");
2558 options = sdscat(options,argv[j]+2);
2559 options = sdscat(options," ");
2560 } else {
2561 /* Option argument */
2562 options = sdscatrepr(options,argv[j],strlen(argv[j]));
2563 options = sdscat(options," ");
2564 }
2565 j++;
2566 }
2567 resetServerSaveParams();
2568 loadServerConfig(configfile,options);
2569 sdsfree(options);
2570 } else {
2571 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");
2572 }
2573 if (server.daemonize) daemonize();
2574 initServer();
2575 if (server.daemonize) createPidFile();
2576 redisAsciiArt();
2577
2578 if (!server.sentinel_mode) {
2579 /* Things only needed when not runnign in Sentinel mode. */
2580 redisLog(REDIS_WARNING,"Server started, Redis version " REDIS_VERSION);
2581 #ifdef __linux__
2582 linuxOvercommitMemoryWarning();
2583 #endif
2584 loadDataFromDisk();
2585 if (server.ipfd > 0)
2586 redisLog(REDIS_NOTICE,"The server is now ready to accept connections on port %d", server.port);
2587 if (server.sofd > 0)
2588 redisLog(REDIS_NOTICE,"The server is now ready to accept connections at %s", server.unixsocket);
2589 }
2590
2591 /* Warning the user about suspicious maxmemory setting. */
2592 if (server.maxmemory > 0 && server.maxmemory < 1024*1024) {
2593 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);
2594 }
2595
2596 aeSetBeforeSleepProc(server.el,beforeSleep);
2597 aeMain(server.el);
2598 aeDeleteEventLoop(server.el);
2599 return 0;
2600 }
2601
2602 /* The End */