]> git.saurik.com Git - redis.git/blob - src/redis.c
Merge pull request #209 from woowenjie/unstable
[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 #ifdef HAVE_BACKTRACE
35 #include <execinfo.h>
36 #include <ucontext.h>
37 #endif /* HAVE_BACKTRACE */
38
39 #include <time.h>
40 #include <signal.h>
41 #include <sys/wait.h>
42 #include <errno.h>
43 #include <assert.h>
44 #include <ctype.h>
45 #include <stdarg.h>
46 #include <arpa/inet.h>
47 #include <sys/stat.h>
48 #include <fcntl.h>
49 #include <sys/time.h>
50 #include <sys/resource.h>
51 #include <sys/uio.h>
52 #include <limits.h>
53 #include <float.h>
54 #include <math.h>
55 #include <sys/resource.h>
56
57 /* Our shared "common" objects */
58
59 struct sharedObjectsStruct shared;
60
61 /* Global vars that are actually used as constants. The following double
62 * values are used for double on-disk serialization, and are initialized
63 * at runtime to avoid strange compiler optimizations. */
64
65 double R_Zero, R_PosInf, R_NegInf, R_Nan;
66
67 /*================================= Globals ================================= */
68
69 /* Global vars */
70 struct redisServer server; /* server global state */
71 struct redisCommand *commandTable;
72
73 /* Our command table.
74 *
75 * Every entry is composed of the following fields:
76 *
77 * name: a string representing the command name.
78 * function: pointer to the C function implementing the command.
79 * arity: number of arguments, it is possible to use -N to say >= N
80 * sflags: command flags as string. See below for a table of flags.
81 * flags: flags as bitmask. Computed by Redis using the 'sflags' field.
82 * get_keys_proc: an optional function to get key arguments from a command.
83 * This is only used when the following three fields are not
84 * enough to specify what arguments are keys.
85 * first_key_index: first argument that is a key
86 * last_key_index: last argument that is a key
87 * key_step: step to get all the keys from first to last argument. For instance
88 * in MSET the step is two since arguments are key,val,key,val,...
89 * microseconds: microseconds of total execution time for this command.
90 * calls: total number of calls of this command.
91 *
92 * The flags, microseconds and calls fields are computed by Redis and should
93 * always be set to zero.
94 *
95 * Command flags are expressed using strings where every character represents
96 * a flag. Later the populateCommandTable() function will take care of
97 * populating the real 'flags' field using this characters.
98 *
99 * This is the meaning of the flags:
100 *
101 * w: write command (may modify the key space).
102 * r: read command (will never modify the key space).
103 * m: may increase memory usage once called. Don't allow if out of memory.
104 * a: admin command, like SAVE or SHUTDOWN.
105 * p: Pub/Sub related command.
106 * f: force replication of this command, regarless of server.dirty.
107 * s: command not allowed in scripts.
108 * R: random command. Command is not deterministic, that is, the same command
109 * with the same arguments, with the same key space, may have different
110 * results. For instance SPOP and RANDOMKEY are two random commands. */
111 struct redisCommand redisCommandTable[] = {
112 {"get",getCommand,2,"r",0,NULL,1,1,1,0,0},
113 {"set",setCommand,3,"wm",0,noPreloadGetKeys,1,1,1,0,0},
114 {"setnx",setnxCommand,3,"wm",0,noPreloadGetKeys,1,1,1,0,0},
115 {"setex",setexCommand,4,"wm",0,noPreloadGetKeys,1,1,1,0,0},
116 {"psetex",psetexCommand,4,"wm",0,noPreloadGetKeys,1,1,1,0,0},
117 {"append",appendCommand,3,"wm",0,NULL,1,1,1,0,0},
118 {"strlen",strlenCommand,2,"r",0,NULL,1,1,1,0,0},
119 {"del",delCommand,-2,"w",0,noPreloadGetKeys,1,-1,1,0,0},
120 {"exists",existsCommand,2,"r",0,NULL,1,1,1,0,0},
121 {"setbit",setbitCommand,4,"wm",0,NULL,1,1,1,0,0},
122 {"getbit",getbitCommand,3,"r",0,NULL,1,1,1,0,0},
123 {"setrange",setrangeCommand,4,"wm",0,NULL,1,1,1,0,0},
124 {"getrange",getrangeCommand,4,"r",0,NULL,1,1,1,0,0},
125 {"substr",getrangeCommand,4,"r",0,NULL,1,1,1,0,0},
126 {"incr",incrCommand,2,"wm",0,NULL,1,1,1,0,0},
127 {"decr",decrCommand,2,"wm",0,NULL,1,1,1,0,0},
128 {"mget",mgetCommand,-2,"r",0,NULL,1,-1,1,0,0},
129 {"rpush",rpushCommand,-3,"wm",0,NULL,1,1,1,0,0},
130 {"lpush",lpushCommand,-3,"wm",0,NULL,1,1,1,0,0},
131 {"rpushx",rpushxCommand,3,"wm",0,NULL,1,1,1,0,0},
132 {"lpushx",lpushxCommand,3,"wm",0,NULL,1,1,1,0,0},
133 {"linsert",linsertCommand,5,"wm",0,NULL,1,1,1,0,0},
134 {"rpop",rpopCommand,2,"w",0,NULL,1,1,1,0,0},
135 {"lpop",lpopCommand,2,"w",0,NULL,1,1,1,0,0},
136 {"brpop",brpopCommand,-3,"ws",0,NULL,1,1,1,0,0},
137 {"brpoplpush",brpoplpushCommand,4,"wms",0,NULL,1,2,1,0,0},
138 {"blpop",blpopCommand,-3,"ws",0,NULL,1,-2,1,0,0},
139 {"llen",llenCommand,2,"r",0,NULL,1,1,1,0,0},
140 {"lindex",lindexCommand,3,"r",0,NULL,1,1,1,0,0},
141 {"lset",lsetCommand,4,"wm",0,NULL,1,1,1,0,0},
142 {"lrange",lrangeCommand,4,"r",0,NULL,1,1,1,0,0},
143 {"ltrim",ltrimCommand,4,"w",0,NULL,1,1,1,0,0},
144 {"lrem",lremCommand,4,"w",0,NULL,1,1,1,0,0},
145 {"rpoplpush",rpoplpushCommand,3,"wm",0,NULL,1,2,1,0,0},
146 {"sadd",saddCommand,-3,"wm",0,NULL,1,1,1,0,0},
147 {"srem",sremCommand,-3,"w",0,NULL,1,1,1,0,0},
148 {"smove",smoveCommand,4,"w",0,NULL,1,2,1,0,0},
149 {"sismember",sismemberCommand,3,"r",0,NULL,1,1,1,0,0},
150 {"scard",scardCommand,2,"r",0,NULL,1,1,1,0,0},
151 {"spop",spopCommand,2,"wRs",0,NULL,1,1,1,0,0},
152 {"srandmember",srandmemberCommand,2,"rR",0,NULL,1,1,1,0,0},
153 {"sinter",sinterCommand,-2,"r",0,NULL,1,-1,1,0,0},
154 {"sinterstore",sinterstoreCommand,-3,"wm",0,NULL,1,-1,1,0,0},
155 {"sunion",sunionCommand,-2,"r",0,NULL,1,-1,1,0,0},
156 {"sunionstore",sunionstoreCommand,-3,"wm",0,NULL,1,-1,1,0,0},
157 {"sdiff",sdiffCommand,-2,"r",0,NULL,1,-1,1,0,0},
158 {"sdiffstore",sdiffstoreCommand,-3,"wm",0,NULL,1,-1,1,0,0},
159 {"smembers",sinterCommand,2,"r",0,NULL,1,1,1,0,0},
160 {"zadd",zaddCommand,-4,"wm",0,NULL,1,1,1,0,0},
161 {"zincrby",zincrbyCommand,4,"wm",0,NULL,1,1,1,0,0},
162 {"zrem",zremCommand,-3,"w",0,NULL,1,1,1,0,0},
163 {"zremrangebyscore",zremrangebyscoreCommand,4,"w",0,NULL,1,1,1,0,0},
164 {"zremrangebyrank",zremrangebyrankCommand,4,"w",0,NULL,1,1,1,0,0},
165 {"zunionstore",zunionstoreCommand,-4,"wm",0,zunionInterGetKeys,0,0,0,0,0},
166 {"zinterstore",zinterstoreCommand,-4,"wm",0,zunionInterGetKeys,0,0,0,0,0},
167 {"zrange",zrangeCommand,-4,"r",0,NULL,1,1,1,0,0},
168 {"zrangebyscore",zrangebyscoreCommand,-4,"r",0,NULL,1,1,1,0,0},
169 {"zrevrangebyscore",zrevrangebyscoreCommand,-4,"r",0,NULL,1,1,1,0,0},
170 {"zcount",zcountCommand,4,"r",0,NULL,1,1,1,0,0},
171 {"zrevrange",zrevrangeCommand,-4,"r",0,NULL,1,1,1,0,0},
172 {"zcard",zcardCommand,2,"r",0,NULL,1,1,1,0,0},
173 {"zscore",zscoreCommand,3,"r",0,NULL,1,1,1,0,0},
174 {"zrank",zrankCommand,3,"r",0,NULL,1,1,1,0,0},
175 {"zrevrank",zrevrankCommand,3,"r",0,NULL,1,1,1,0,0},
176 {"hset",hsetCommand,4,"wm",0,NULL,1,1,1,0,0},
177 {"hsetnx",hsetnxCommand,4,"wm",0,NULL,1,1,1,0,0},
178 {"hget",hgetCommand,3,"r",0,NULL,1,1,1,0,0},
179 {"hmset",hmsetCommand,-4,"wm",0,NULL,1,1,1,0,0},
180 {"hmget",hmgetCommand,-3,"r",0,NULL,1,1,1,0,0},
181 {"hincrby",hincrbyCommand,4,"wm",0,NULL,1,1,1,0,0},
182 {"hincrbyfloat",hincrbyfloatCommand,4,"wm",0,NULL,1,1,1,0,0},
183 {"hdel",hdelCommand,-3,"w",0,NULL,1,1,1,0,0},
184 {"hlen",hlenCommand,2,"r",0,NULL,1,1,1,0,0},
185 {"hkeys",hkeysCommand,2,"r",0,NULL,1,1,1,0,0},
186 {"hvals",hvalsCommand,2,"r",0,NULL,1,1,1,0,0},
187 {"hgetall",hgetallCommand,2,"r",0,NULL,1,1,1,0,0},
188 {"hexists",hexistsCommand,3,"r",0,NULL,1,1,1,0,0},
189 {"incrby",incrbyCommand,3,"wm",0,NULL,1,1,1,0,0},
190 {"decrby",decrbyCommand,3,"wm",0,NULL,1,1,1,0,0},
191 {"incrbyfloat",incrbyfloatCommand,3,"wm",0,NULL,1,1,1,0,0},
192 {"getset",getsetCommand,3,"wm",0,NULL,1,1,1,0,0},
193 {"mset",msetCommand,-3,"wm",0,NULL,1,-1,2,0,0},
194 {"msetnx",msetnxCommand,-3,"wm",0,NULL,1,-1,2,0,0},
195 {"randomkey",randomkeyCommand,1,"rR",0,NULL,0,0,0,0,0},
196 {"select",selectCommand,2,"r",0,NULL,0,0,0,0,0},
197 {"move",moveCommand,3,"w",0,NULL,1,1,1,0,0},
198 {"rename",renameCommand,3,"w",0,renameGetKeys,1,2,1,0,0},
199 {"renamenx",renamenxCommand,3,"w",0,renameGetKeys,1,2,1,0,0},
200 {"expire",expireCommand,3,"w",0,NULL,1,1,1,0,0},
201 {"expireat",expireatCommand,3,"w",0,NULL,1,1,1,0,0},
202 {"pexpire",pexpireCommand,3,"w",0,NULL,1,1,1,0,0},
203 {"pexpireat",pexpireatCommand,3,"w",0,NULL,1,1,1,0,0},
204 {"keys",keysCommand,2,"r",0,NULL,0,0,0,0,0},
205 {"dbsize",dbsizeCommand,1,"r",0,NULL,0,0,0,0,0},
206 {"auth",authCommand,2,"rs",0,NULL,0,0,0,0,0},
207 {"ping",pingCommand,1,"r",0,NULL,0,0,0,0,0},
208 {"echo",echoCommand,2,"r",0,NULL,0,0,0,0,0},
209 {"save",saveCommand,1,"ars",0,NULL,0,0,0,0,0},
210 {"bgsave",bgsaveCommand,1,"ar",0,NULL,0,0,0,0,0},
211 {"bgrewriteaof",bgrewriteaofCommand,1,"ar",0,NULL,0,0,0,0,0},
212 {"shutdown",shutdownCommand,-1,"ar",0,NULL,0,0,0,0,0},
213 {"lastsave",lastsaveCommand,1,"r",0,NULL,0,0,0,0,0},
214 {"type",typeCommand,2,"r",0,NULL,1,1,1,0,0},
215 {"multi",multiCommand,1,"rs",0,NULL,0,0,0,0,0},
216 {"exec",execCommand,1,"wms",0,NULL,0,0,0,0,0},
217 {"discard",discardCommand,1,"rs",0,NULL,0,0,0,0,0},
218 {"sync",syncCommand,1,"ars",0,NULL,0,0,0,0,0},
219 {"flushdb",flushdbCommand,1,"w",0,NULL,0,0,0,0,0},
220 {"flushall",flushallCommand,1,"w",0,NULL,0,0,0,0,0},
221 {"sort",sortCommand,-2,"wm",0,NULL,1,1,1,0,0},
222 {"info",infoCommand,-1,"r",0,NULL,0,0,0,0,0},
223 {"monitor",monitorCommand,1,"ars",0,NULL,0,0,0,0,0},
224 {"ttl",ttlCommand,2,"r",0,NULL,1,1,1,0,0},
225 {"pttl",pttlCommand,2,"r",0,NULL,1,1,1,0,0},
226 {"persist",persistCommand,2,"w",0,NULL,1,1,1,0,0},
227 {"slaveof",slaveofCommand,3,"aws",0,NULL,0,0,0,0,0},
228 {"debug",debugCommand,-2,"aws",0,NULL,0,0,0,0,0},
229 {"config",configCommand,-2,"ar",0,NULL,0,0,0,0,0},
230 {"subscribe",subscribeCommand,-2,"rps",0,NULL,0,0,0,0,0},
231 {"unsubscribe",unsubscribeCommand,-1,"rps",0,NULL,0,0,0,0,0},
232 {"psubscribe",psubscribeCommand,-2,"rps",0,NULL,0,0,0,0,0},
233 {"punsubscribe",punsubscribeCommand,-1,"rps",0,NULL,0,0,0,0,0},
234 {"publish",publishCommand,3,"rpf",0,NULL,0,0,0,0,0},
235 {"watch",watchCommand,-2,"rs",0,noPreloadGetKeys,1,-1,1,0,0},
236 {"unwatch",unwatchCommand,1,"rs",0,NULL,0,0,0,0,0},
237 {"cluster",clusterCommand,-2,"ar",0,NULL,0,0,0,0,0},
238 {"restore",restoreCommand,4,"awm",0,NULL,1,1,1,0,0},
239 {"migrate",migrateCommand,6,"aw",0,NULL,0,0,0,0,0},
240 {"asking",askingCommand,1,"r",0,NULL,0,0,0,0,0},
241 {"dump",dumpCommand,2,"ar",0,NULL,1,1,1,0,0},
242 {"object",objectCommand,-2,"r",0,NULL,2,2,2,0,0},
243 {"client",clientCommand,-2,"ar",0,NULL,0,0,0,0,0},
244 {"eval",evalCommand,-3,"wms",0,zunionInterGetKeys,0,0,0,0,0},
245 {"evalsha",evalShaCommand,-3,"wms",0,zunionInterGetKeys,0,0,0,0,0},
246 {"slowlog",slowlogCommand,-2,"r",0,NULL,0,0,0,0,0},
247 {"script",scriptCommand,-2,"ras",0,NULL,0,0,0,0,0}
248 };
249
250 /*============================ Utility functions ============================ */
251
252 /* Low level logging. To use only for very big messages, otherwise
253 * redisLog() is to prefer. */
254 void redisLogRaw(int level, const char *msg) {
255 const int syslogLevelMap[] = { LOG_DEBUG, LOG_INFO, LOG_NOTICE, LOG_WARNING };
256 const char *c = ".-*#";
257 time_t now = time(NULL);
258 FILE *fp;
259 char buf[64];
260 int rawmode = (level & REDIS_LOG_RAW);
261
262 level &= 0xff; /* clear flags */
263 if (level < server.verbosity) return;
264
265 fp = (server.logfile == NULL) ? stdout : fopen(server.logfile,"a");
266 if (!fp) return;
267
268 if (rawmode) {
269 fprintf(fp,"%s",msg);
270 } else {
271 strftime(buf,sizeof(buf),"%d %b %H:%M:%S",localtime(&now));
272 fprintf(fp,"[%d] %s %c %s\n",(int)getpid(),buf,c[level],msg);
273 }
274 fflush(fp);
275
276 if (server.logfile) fclose(fp);
277
278 if (server.syslog_enabled) syslog(syslogLevelMap[level], "%s", msg);
279 }
280
281 /* Like redisLogRaw() but with printf-alike support. This is the funciton that
282 * is used across the code. The raw version is only used in order to dump
283 * the INFO output on crash. */
284 void redisLog(int level, const char *fmt, ...) {
285 va_list ap;
286 char msg[REDIS_MAX_LOGMSG_LEN];
287
288 if ((level&0xff) < server.verbosity) return;
289
290 va_start(ap, fmt);
291 vsnprintf(msg, sizeof(msg), fmt, ap);
292 va_end(ap);
293
294 redisLogRaw(level,msg);
295 }
296
297 /* Redis generally does not try to recover from out of memory conditions
298 * when allocating objects or strings, it is not clear if it will be possible
299 * to report this condition to the client since the networking layer itself
300 * is based on heap allocation for send buffers, so we simply abort.
301 * At least the code will be simpler to read... */
302 void oom(const char *msg) {
303 redisLog(REDIS_WARNING, "%s: Out of memory\n",msg);
304 sleep(1);
305 abort();
306 }
307
308 /* Return the UNIX time in microseconds */
309 long long ustime(void) {
310 struct timeval tv;
311 long long ust;
312
313 gettimeofday(&tv, NULL);
314 ust = ((long long)tv.tv_sec)*1000000;
315 ust += tv.tv_usec;
316 return ust;
317 }
318
319 /* Return the UNIX time in milliseconds */
320 long long mstime(void) {
321 return ustime()/1000;
322 }
323
324 /*====================== Hash table type implementation ==================== */
325
326 /* This is an hash table type that uses the SDS dynamic strings libary as
327 * keys and radis objects as values (objects can hold SDS strings,
328 * lists, sets). */
329
330 void dictVanillaFree(void *privdata, void *val)
331 {
332 DICT_NOTUSED(privdata);
333 zfree(val);
334 }
335
336 void dictListDestructor(void *privdata, void *val)
337 {
338 DICT_NOTUSED(privdata);
339 listRelease((list*)val);
340 }
341
342 int dictSdsKeyCompare(void *privdata, const void *key1,
343 const void *key2)
344 {
345 int l1,l2;
346 DICT_NOTUSED(privdata);
347
348 l1 = sdslen((sds)key1);
349 l2 = sdslen((sds)key2);
350 if (l1 != l2) return 0;
351 return memcmp(key1, key2, l1) == 0;
352 }
353
354 /* A case insensitive version used for the command lookup table. */
355 int dictSdsKeyCaseCompare(void *privdata, const void *key1,
356 const void *key2)
357 {
358 DICT_NOTUSED(privdata);
359
360 return strcasecmp(key1, key2) == 0;
361 }
362
363 void dictRedisObjectDestructor(void *privdata, void *val)
364 {
365 DICT_NOTUSED(privdata);
366
367 if (val == NULL) return; /* Values of swapped out keys as set to NULL */
368 decrRefCount(val);
369 }
370
371 void dictSdsDestructor(void *privdata, void *val)
372 {
373 DICT_NOTUSED(privdata);
374
375 sdsfree(val);
376 }
377
378 int dictObjKeyCompare(void *privdata, const void *key1,
379 const void *key2)
380 {
381 const robj *o1 = key1, *o2 = key2;
382 return dictSdsKeyCompare(privdata,o1->ptr,o2->ptr);
383 }
384
385 unsigned int dictObjHash(const void *key) {
386 const robj *o = key;
387 return dictGenHashFunction(o->ptr, sdslen((sds)o->ptr));
388 }
389
390 unsigned int dictSdsHash(const void *key) {
391 return dictGenHashFunction((unsigned char*)key, sdslen((char*)key));
392 }
393
394 unsigned int dictSdsCaseHash(const void *key) {
395 return dictGenCaseHashFunction((unsigned char*)key, sdslen((char*)key));
396 }
397
398 int dictEncObjKeyCompare(void *privdata, const void *key1,
399 const void *key2)
400 {
401 robj *o1 = (robj*) key1, *o2 = (robj*) key2;
402 int cmp;
403
404 if (o1->encoding == REDIS_ENCODING_INT &&
405 o2->encoding == REDIS_ENCODING_INT)
406 return o1->ptr == o2->ptr;
407
408 o1 = getDecodedObject(o1);
409 o2 = getDecodedObject(o2);
410 cmp = dictSdsKeyCompare(privdata,o1->ptr,o2->ptr);
411 decrRefCount(o1);
412 decrRefCount(o2);
413 return cmp;
414 }
415
416 unsigned int dictEncObjHash(const void *key) {
417 robj *o = (robj*) key;
418
419 if (o->encoding == REDIS_ENCODING_RAW) {
420 return dictGenHashFunction(o->ptr, sdslen((sds)o->ptr));
421 } else {
422 if (o->encoding == REDIS_ENCODING_INT) {
423 char buf[32];
424 int len;
425
426 len = ll2string(buf,32,(long)o->ptr);
427 return dictGenHashFunction((unsigned char*)buf, len);
428 } else {
429 unsigned int hash;
430
431 o = getDecodedObject(o);
432 hash = dictGenHashFunction(o->ptr, sdslen((sds)o->ptr));
433 decrRefCount(o);
434 return hash;
435 }
436 }
437 }
438
439 /* Sets type hash table */
440 dictType setDictType = {
441 dictEncObjHash, /* hash function */
442 NULL, /* key dup */
443 NULL, /* val dup */
444 dictEncObjKeyCompare, /* key compare */
445 dictRedisObjectDestructor, /* key destructor */
446 NULL /* val destructor */
447 };
448
449 /* Sorted sets hash (note: a skiplist is used in addition to the hash table) */
450 dictType zsetDictType = {
451 dictEncObjHash, /* hash function */
452 NULL, /* key dup */
453 NULL, /* val dup */
454 dictEncObjKeyCompare, /* key compare */
455 dictRedisObjectDestructor, /* key destructor */
456 NULL /* val destructor */
457 };
458
459 /* Db->dict, keys are sds strings, vals are Redis objects. */
460 dictType dbDictType = {
461 dictSdsHash, /* hash function */
462 NULL, /* key dup */
463 NULL, /* val dup */
464 dictSdsKeyCompare, /* key compare */
465 dictSdsDestructor, /* key destructor */
466 dictRedisObjectDestructor /* val destructor */
467 };
468
469 /* Db->expires */
470 dictType keyptrDictType = {
471 dictSdsHash, /* hash function */
472 NULL, /* key dup */
473 NULL, /* val dup */
474 dictSdsKeyCompare, /* key compare */
475 NULL, /* key destructor */
476 NULL /* val destructor */
477 };
478
479 /* Command table. sds string -> command struct pointer. */
480 dictType commandTableDictType = {
481 dictSdsCaseHash, /* hash function */
482 NULL, /* key dup */
483 NULL, /* val dup */
484 dictSdsKeyCaseCompare, /* key compare */
485 dictSdsDestructor, /* key destructor */
486 NULL /* val destructor */
487 };
488
489 /* Hash type hash table (note that small hashes are represented with zimpaps) */
490 dictType hashDictType = {
491 dictEncObjHash, /* hash function */
492 NULL, /* key dup */
493 NULL, /* val dup */
494 dictEncObjKeyCompare, /* key compare */
495 dictRedisObjectDestructor, /* key destructor */
496 dictRedisObjectDestructor /* val destructor */
497 };
498
499 /* Keylist hash table type has unencoded redis objects as keys and
500 * lists as values. It's used for blocking operations (BLPOP) and to
501 * map swapped keys to a list of clients waiting for this keys to be loaded. */
502 dictType keylistDictType = {
503 dictObjHash, /* hash function */
504 NULL, /* key dup */
505 NULL, /* val dup */
506 dictObjKeyCompare, /* key compare */
507 dictRedisObjectDestructor, /* key destructor */
508 dictListDestructor /* val destructor */
509 };
510
511 /* Cluster nodes hash table, mapping nodes addresses 1.2.3.4:6379 to
512 * clusterNode structures. */
513 dictType clusterNodesDictType = {
514 dictSdsHash, /* hash function */
515 NULL, /* key dup */
516 NULL, /* val dup */
517 dictSdsKeyCompare, /* key compare */
518 dictSdsDestructor, /* key destructor */
519 NULL /* val destructor */
520 };
521
522 int htNeedsResize(dict *dict) {
523 long long size, used;
524
525 size = dictSlots(dict);
526 used = dictSize(dict);
527 return (size && used && size > DICT_HT_INITIAL_SIZE &&
528 (used*100/size < REDIS_HT_MINFILL));
529 }
530
531 /* If the percentage of used slots in the HT reaches REDIS_HT_MINFILL
532 * we resize the hash table to save memory */
533 void tryResizeHashTables(void) {
534 int j;
535
536 for (j = 0; j < server.dbnum; j++) {
537 if (htNeedsResize(server.db[j].dict))
538 dictResize(server.db[j].dict);
539 if (htNeedsResize(server.db[j].expires))
540 dictResize(server.db[j].expires);
541 }
542 }
543
544 /* Our hash table implementation performs rehashing incrementally while
545 * we write/read from the hash table. Still if the server is idle, the hash
546 * table will use two tables for a long time. So we try to use 1 millisecond
547 * of CPU time at every serverCron() loop in order to rehash some key. */
548 void incrementallyRehash(void) {
549 int j;
550
551 for (j = 0; j < server.dbnum; j++) {
552 if (dictIsRehashing(server.db[j].dict)) {
553 dictRehashMilliseconds(server.db[j].dict,1);
554 break; /* already used our millisecond for this loop... */
555 }
556 }
557 }
558
559 /* This function is called once a background process of some kind terminates,
560 * as we want to avoid resizing the hash tables when there is a child in order
561 * to play well with copy-on-write (otherwise when a resize happens lots of
562 * memory pages are copied). The goal of this function is to update the ability
563 * for dict.c to resize the hash tables accordingly to the fact we have o not
564 * running childs. */
565 void updateDictResizePolicy(void) {
566 if (server.bgsavechildpid == -1 && server.bgrewritechildpid == -1)
567 dictEnableResize();
568 else
569 dictDisableResize();
570 }
571
572 /* ======================= Cron: called every 100 ms ======================== */
573
574 /* Try to expire a few timed out keys. The algorithm used is adaptive and
575 * will use few CPU cycles if there are few expiring keys, otherwise
576 * it will get more aggressive to avoid that too much memory is used by
577 * keys that can be removed from the keyspace. */
578 void activeExpireCycle(void) {
579 int j;
580
581 for (j = 0; j < server.dbnum; j++) {
582 int expired;
583 redisDb *db = server.db+j;
584
585 /* Continue to expire if at the end of the cycle more than 25%
586 * of the keys were expired. */
587 do {
588 long num = dictSize(db->expires);
589 long long now = mstime();
590
591 expired = 0;
592 if (num > REDIS_EXPIRELOOKUPS_PER_CRON)
593 num = REDIS_EXPIRELOOKUPS_PER_CRON;
594 while (num--) {
595 dictEntry *de;
596 long long t;
597
598 if ((de = dictGetRandomKey(db->expires)) == NULL) break;
599 t = dictGetSignedIntegerVal(de);
600 if (now > t) {
601 sds key = dictGetKey(de);
602 robj *keyobj = createStringObject(key,sdslen(key));
603
604 propagateExpire(db,keyobj);
605 dbDelete(db,keyobj);
606 decrRefCount(keyobj);
607 expired++;
608 server.stat_expiredkeys++;
609 }
610 }
611 } while (expired > REDIS_EXPIRELOOKUPS_PER_CRON/4);
612 }
613 }
614
615 void updateLRUClock(void) {
616 server.lruclock = (time(NULL)/REDIS_LRU_CLOCK_RESOLUTION) &
617 REDIS_LRU_CLOCK_MAX;
618 }
619
620 int serverCron(struct aeEventLoop *eventLoop, long long id, void *clientData) {
621 int j, loops = server.cronloops;
622 REDIS_NOTUSED(eventLoop);
623 REDIS_NOTUSED(id);
624 REDIS_NOTUSED(clientData);
625
626 /* We take a cached value of the unix time in the global state because
627 * with virtual memory and aging there is to store the current time
628 * in objects at every object access, and accuracy is not needed.
629 * To access a global var is faster than calling time(NULL) */
630 server.unixtime = time(NULL);
631
632 /* We have just 22 bits per object for LRU information.
633 * So we use an (eventually wrapping) LRU clock with 10 seconds resolution.
634 * 2^22 bits with 10 seconds resoluton is more or less 1.5 years.
635 *
636 * Note that even if this will wrap after 1.5 years it's not a problem,
637 * everything will still work but just some object will appear younger
638 * to Redis. But for this to happen a given object should never be touched
639 * for 1.5 years.
640 *
641 * Note that you can change the resolution altering the
642 * REDIS_LRU_CLOCK_RESOLUTION define.
643 */
644 updateLRUClock();
645
646 /* Record the max memory used since the server was started. */
647 if (zmalloc_used_memory() > server.stat_peak_memory)
648 server.stat_peak_memory = zmalloc_used_memory();
649
650 /* We received a SIGTERM, shutting down here in a safe way, as it is
651 * not ok doing so inside the signal handler. */
652 if (server.shutdown_asap) {
653 if (prepareForShutdown(0) == REDIS_OK) exit(0);
654 redisLog(REDIS_WARNING,"SIGTERM received but errors trying to shut down the server, check the logs for more information");
655 }
656
657 /* Show some info about non-empty databases */
658 for (j = 0; j < server.dbnum; j++) {
659 long long size, used, vkeys;
660
661 size = dictSlots(server.db[j].dict);
662 used = dictSize(server.db[j].dict);
663 vkeys = dictSize(server.db[j].expires);
664 if (!(loops % 50) && (used || vkeys)) {
665 redisLog(REDIS_VERBOSE,"DB %d: %lld keys (%lld volatile) in %lld slots HT.",j,used,vkeys,size);
666 /* dictPrintStats(server.dict); */
667 }
668 }
669
670 /* We don't want to resize the hash tables while a bacground saving
671 * is in progress: the saving child is created using fork() that is
672 * implemented with a copy-on-write semantic in most modern systems, so
673 * if we resize the HT while there is the saving child at work actually
674 * a lot of memory movements in the parent will cause a lot of pages
675 * copied. */
676 if (server.bgsavechildpid == -1 && server.bgrewritechildpid == -1) {
677 if (!(loops % 10)) tryResizeHashTables();
678 if (server.activerehashing) incrementallyRehash();
679 }
680
681 /* Show information about connected clients */
682 if (!(loops % 50)) {
683 redisLog(REDIS_VERBOSE,"%d clients connected (%d slaves), %zu bytes in use",
684 listLength(server.clients)-listLength(server.slaves),
685 listLength(server.slaves),
686 zmalloc_used_memory());
687 }
688
689 /* Close connections of timedout clients */
690 if ((server.maxidletime && !(loops % 100)) || server.bpop_blocked_clients)
691 closeTimedoutClients();
692
693 /* Start a scheduled AOF rewrite if this was requested by the user while
694 * a BGSAVE was in progress. */
695 if (server.bgsavechildpid == -1 && server.bgrewritechildpid == -1 &&
696 server.aofrewrite_scheduled)
697 {
698 rewriteAppendOnlyFileBackground();
699 }
700
701 /* Check if a background saving or AOF rewrite in progress terminated. */
702 if (server.bgsavechildpid != -1 || server.bgrewritechildpid != -1) {
703 int statloc;
704 pid_t pid;
705
706 if ((pid = wait3(&statloc,WNOHANG,NULL)) != 0) {
707 int exitcode = WEXITSTATUS(statloc);
708 int bysignal = 0;
709
710 if (WIFSIGNALED(statloc)) bysignal = WTERMSIG(statloc);
711
712 if (pid == server.bgsavechildpid) {
713 backgroundSaveDoneHandler(exitcode,bysignal);
714 } else {
715 backgroundRewriteDoneHandler(exitcode,bysignal);
716 }
717 updateDictResizePolicy();
718 }
719 } else {
720 time_t now = time(NULL);
721
722 /* If there is not a background saving/rewrite in progress check if
723 * we have to save/rewrite now */
724 for (j = 0; j < server.saveparamslen; j++) {
725 struct saveparam *sp = server.saveparams+j;
726
727 if (server.dirty >= sp->changes &&
728 now-server.lastsave > sp->seconds) {
729 redisLog(REDIS_NOTICE,"%d changes in %d seconds. Saving...",
730 sp->changes, sp->seconds);
731 rdbSaveBackground(server.dbfilename);
732 break;
733 }
734 }
735
736 /* Trigger an AOF rewrite if needed */
737 if (server.bgsavechildpid == -1 &&
738 server.bgrewritechildpid == -1 &&
739 server.auto_aofrewrite_perc &&
740 server.appendonly_current_size > server.auto_aofrewrite_min_size)
741 {
742 long long base = server.auto_aofrewrite_base_size ?
743 server.auto_aofrewrite_base_size : 1;
744 long long growth = (server.appendonly_current_size*100/base) - 100;
745 if (growth >= server.auto_aofrewrite_perc) {
746 redisLog(REDIS_NOTICE,"Starting automatic rewriting of AOF on %lld%% growth",growth);
747 rewriteAppendOnlyFileBackground();
748 }
749 }
750 }
751
752
753 /* If we postponed an AOF buffer flush, let's try to do it every time the
754 * cron function is called. */
755 if (server.aof_flush_postponed_start) flushAppendOnlyFile(0);
756
757 /* Expire a few keys per cycle, only if this is a master.
758 * On slaves we wait for DEL operations synthesized by the master
759 * in order to guarantee a strict consistency. */
760 if (server.masterhost == NULL) activeExpireCycle();
761
762 /* Replication cron function -- used to reconnect to master and
763 * to detect transfer failures. */
764 if (!(loops % 10)) replicationCron();
765
766 /* Run other sub-systems specific cron jobs */
767 if (server.cluster_enabled && !(loops % 10)) clusterCron();
768
769 server.cronloops++;
770 return 100;
771 }
772
773 /* This function gets called every time Redis is entering the
774 * main loop of the event driven library, that is, before to sleep
775 * for ready file descriptors. */
776 void beforeSleep(struct aeEventLoop *eventLoop) {
777 REDIS_NOTUSED(eventLoop);
778 listNode *ln;
779 redisClient *c;
780
781 /* Try to process pending commands for clients that were just unblocked. */
782 while (listLength(server.unblocked_clients)) {
783 ln = listFirst(server.unblocked_clients);
784 redisAssert(ln != NULL);
785 c = ln->value;
786 listDelNode(server.unblocked_clients,ln);
787 c->flags &= ~REDIS_UNBLOCKED;
788
789 /* Process remaining data in the input buffer. */
790 if (c->querybuf && sdslen(c->querybuf) > 0)
791 processInputBuffer(c);
792 }
793
794 /* Write the AOF buffer on disk */
795 flushAppendOnlyFile(0);
796 }
797
798 /* =========================== Server initialization ======================== */
799
800 void createSharedObjects(void) {
801 int j;
802
803 shared.crlf = createObject(REDIS_STRING,sdsnew("\r\n"));
804 shared.ok = createObject(REDIS_STRING,sdsnew("+OK\r\n"));
805 shared.err = createObject(REDIS_STRING,sdsnew("-ERR\r\n"));
806 shared.emptybulk = createObject(REDIS_STRING,sdsnew("$0\r\n\r\n"));
807 shared.czero = createObject(REDIS_STRING,sdsnew(":0\r\n"));
808 shared.cone = createObject(REDIS_STRING,sdsnew(":1\r\n"));
809 shared.cnegone = createObject(REDIS_STRING,sdsnew(":-1\r\n"));
810 shared.nullbulk = createObject(REDIS_STRING,sdsnew("$-1\r\n"));
811 shared.nullmultibulk = createObject(REDIS_STRING,sdsnew("*-1\r\n"));
812 shared.emptymultibulk = createObject(REDIS_STRING,sdsnew("*0\r\n"));
813 shared.pong = createObject(REDIS_STRING,sdsnew("+PONG\r\n"));
814 shared.queued = createObject(REDIS_STRING,sdsnew("+QUEUED\r\n"));
815 shared.wrongtypeerr = createObject(REDIS_STRING,sdsnew(
816 "-ERR Operation against a key holding the wrong kind of value\r\n"));
817 shared.nokeyerr = createObject(REDIS_STRING,sdsnew(
818 "-ERR no such key\r\n"));
819 shared.syntaxerr = createObject(REDIS_STRING,sdsnew(
820 "-ERR syntax error\r\n"));
821 shared.sameobjecterr = createObject(REDIS_STRING,sdsnew(
822 "-ERR source and destination objects are the same\r\n"));
823 shared.outofrangeerr = createObject(REDIS_STRING,sdsnew(
824 "-ERR index out of range\r\n"));
825 shared.noscripterr = createObject(REDIS_STRING,sdsnew(
826 "-NOSCRIPT No matching script. Please use EVAL.\r\n"));
827 shared.loadingerr = createObject(REDIS_STRING,sdsnew(
828 "-LOADING Redis is loading the dataset in memory\r\n"));
829 shared.slowscripterr = createObject(REDIS_STRING,sdsnew(
830 "-BUSY Redis is busy running a script. You can only call SCRIPT KILL or SHUTDOWN NOSAVE.\r\n"));
831 shared.space = createObject(REDIS_STRING,sdsnew(" "));
832 shared.colon = createObject(REDIS_STRING,sdsnew(":"));
833 shared.plus = createObject(REDIS_STRING,sdsnew("+"));
834 shared.select0 = createStringObject("select 0\r\n",10);
835 shared.select1 = createStringObject("select 1\r\n",10);
836 shared.select2 = createStringObject("select 2\r\n",10);
837 shared.select3 = createStringObject("select 3\r\n",10);
838 shared.select4 = createStringObject("select 4\r\n",10);
839 shared.select5 = createStringObject("select 5\r\n",10);
840 shared.select6 = createStringObject("select 6\r\n",10);
841 shared.select7 = createStringObject("select 7\r\n",10);
842 shared.select8 = createStringObject("select 8\r\n",10);
843 shared.select9 = createStringObject("select 9\r\n",10);
844 shared.messagebulk = createStringObject("$7\r\nmessage\r\n",13);
845 shared.pmessagebulk = createStringObject("$8\r\npmessage\r\n",14);
846 shared.subscribebulk = createStringObject("$9\r\nsubscribe\r\n",15);
847 shared.unsubscribebulk = createStringObject("$11\r\nunsubscribe\r\n",18);
848 shared.psubscribebulk = createStringObject("$10\r\npsubscribe\r\n",17);
849 shared.punsubscribebulk = createStringObject("$12\r\npunsubscribe\r\n",19);
850 shared.mbulk3 = createStringObject("*3\r\n",4);
851 shared.mbulk4 = createStringObject("*4\r\n",4);
852 for (j = 0; j < REDIS_SHARED_INTEGERS; j++) {
853 shared.integers[j] = createObject(REDIS_STRING,(void*)(long)j);
854 shared.integers[j]->encoding = REDIS_ENCODING_INT;
855 }
856 }
857
858 void initServerConfig() {
859 server.port = REDIS_SERVERPORT;
860 server.bindaddr = NULL;
861 server.unixsocket = NULL;
862 server.unixsocketperm = 0;
863 server.ipfd = -1;
864 server.sofd = -1;
865 server.dbnum = REDIS_DEFAULT_DBNUM;
866 server.verbosity = REDIS_VERBOSE;
867 server.maxidletime = REDIS_MAXIDLETIME;
868 server.client_max_querybuf_len = REDIS_MAX_QUERYBUF_LEN;
869 server.saveparams = NULL;
870 server.loading = 0;
871 server.logfile = NULL; /* NULL = log on standard output */
872 server.syslog_enabled = 0;
873 server.syslog_ident = zstrdup("redis");
874 server.syslog_facility = LOG_LOCAL0;
875 server.daemonize = 0;
876 server.appendonly = 0;
877 server.appendfsync = APPENDFSYNC_EVERYSEC;
878 server.no_appendfsync_on_rewrite = 0;
879 server.auto_aofrewrite_perc = REDIS_AUTO_AOFREWRITE_PERC;
880 server.auto_aofrewrite_min_size = REDIS_AUTO_AOFREWRITE_MIN_SIZE;
881 server.auto_aofrewrite_base_size = 0;
882 server.aofrewrite_scheduled = 0;
883 server.lastfsync = time(NULL);
884 server.appendfd = -1;
885 server.appendseldb = -1; /* Make sure the first time will not match */
886 server.aof_flush_postponed_start = 0;
887 server.pidfile = zstrdup("/var/run/redis.pid");
888 server.dbfilename = zstrdup("dump.rdb");
889 server.appendfilename = zstrdup("appendonly.aof");
890 server.requirepass = NULL;
891 server.rdbcompression = 1;
892 server.activerehashing = 1;
893 server.maxclients = REDIS_MAX_CLIENTS;
894 server.bpop_blocked_clients = 0;
895 server.maxmemory = 0;
896 server.maxmemory_policy = REDIS_MAXMEMORY_VOLATILE_LRU;
897 server.maxmemory_samples = 3;
898 server.hash_max_zipmap_entries = REDIS_HASH_MAX_ZIPMAP_ENTRIES;
899 server.hash_max_zipmap_value = REDIS_HASH_MAX_ZIPMAP_VALUE;
900 server.list_max_ziplist_entries = REDIS_LIST_MAX_ZIPLIST_ENTRIES;
901 server.list_max_ziplist_value = REDIS_LIST_MAX_ZIPLIST_VALUE;
902 server.set_max_intset_entries = REDIS_SET_MAX_INTSET_ENTRIES;
903 server.zset_max_ziplist_entries = REDIS_ZSET_MAX_ZIPLIST_ENTRIES;
904 server.zset_max_ziplist_value = REDIS_ZSET_MAX_ZIPLIST_VALUE;
905 server.shutdown_asap = 0;
906 server.repl_ping_slave_period = REDIS_REPL_PING_SLAVE_PERIOD;
907 server.repl_timeout = REDIS_REPL_TIMEOUT;
908 server.cluster_enabled = 0;
909 server.cluster.configfile = zstrdup("nodes.conf");
910 server.lua_caller = NULL;
911 server.lua_time_limit = REDIS_LUA_TIME_LIMIT;
912 server.lua_client = NULL;
913 server.lua_timedout = 0;
914
915 updateLRUClock();
916 resetServerSaveParams();
917
918 appendServerSaveParams(60*60,1); /* save after 1 hour and 1 change */
919 appendServerSaveParams(300,100); /* save after 5 minutes and 100 changes */
920 appendServerSaveParams(60,10000); /* save after 1 minute and 10000 changes */
921 /* Replication related */
922 server.isslave = 0;
923 server.masterauth = NULL;
924 server.masterhost = NULL;
925 server.masterport = 6379;
926 server.master = NULL;
927 server.replstate = REDIS_REPL_NONE;
928 server.repl_syncio_timeout = REDIS_REPL_SYNCIO_TIMEOUT;
929 server.repl_serve_stale_data = 1;
930 server.repl_down_since = -1;
931
932 /* Double constants initialization */
933 R_Zero = 0.0;
934 R_PosInf = 1.0/R_Zero;
935 R_NegInf = -1.0/R_Zero;
936 R_Nan = R_Zero/R_Zero;
937
938 /* Command table -- we intiialize it here as it is part of the
939 * initial configuration, since command names may be changed via
940 * redis.conf using the rename-command directive. */
941 server.commands = dictCreate(&commandTableDictType,NULL);
942 populateCommandTable();
943 server.delCommand = lookupCommandByCString("del");
944 server.multiCommand = lookupCommandByCString("multi");
945
946 /* Slow log */
947 server.slowlog_log_slower_than = REDIS_SLOWLOG_LOG_SLOWER_THAN;
948 server.slowlog_max_len = REDIS_SLOWLOG_MAX_LEN;
949
950 /* Assert */
951 server.assert_failed = "<no assertion failed>";
952 server.assert_file = "<no file>";
953 server.assert_line = 0;
954 server.bug_report_start = 0;
955 }
956
957 void initServer() {
958 int j;
959
960 signal(SIGHUP, SIG_IGN);
961 signal(SIGPIPE, SIG_IGN);
962 setupSignalHandlers();
963
964 if (server.syslog_enabled) {
965 openlog(server.syslog_ident, LOG_PID | LOG_NDELAY | LOG_NOWAIT,
966 server.syslog_facility);
967 }
968
969 server.clients = listCreate();
970 server.slaves = listCreate();
971 server.monitors = listCreate();
972 server.unblocked_clients = listCreate();
973
974 createSharedObjects();
975 server.el = aeCreateEventLoop();
976 server.db = zmalloc(sizeof(redisDb)*server.dbnum);
977
978 if (server.port != 0) {
979 server.ipfd = anetTcpServer(server.neterr,server.port,server.bindaddr);
980 if (server.ipfd == ANET_ERR) {
981 redisLog(REDIS_WARNING, "Opening port %d: %s",
982 server.port, server.neterr);
983 exit(1);
984 }
985 }
986 if (server.unixsocket != NULL) {
987 unlink(server.unixsocket); /* don't care if this fails */
988 server.sofd = anetUnixServer(server.neterr,server.unixsocket,server.unixsocketperm);
989 if (server.sofd == ANET_ERR) {
990 redisLog(REDIS_WARNING, "Opening socket: %s", server.neterr);
991 exit(1);
992 }
993 }
994 if (server.ipfd < 0 && server.sofd < 0) {
995 redisLog(REDIS_WARNING, "Configured to not listen anywhere, exiting.");
996 exit(1);
997 }
998 for (j = 0; j < server.dbnum; j++) {
999 server.db[j].dict = dictCreate(&dbDictType,NULL);
1000 server.db[j].expires = dictCreate(&keyptrDictType,NULL);
1001 server.db[j].blocking_keys = dictCreate(&keylistDictType,NULL);
1002 server.db[j].watched_keys = dictCreate(&keylistDictType,NULL);
1003 server.db[j].id = j;
1004 }
1005 server.pubsub_channels = dictCreate(&keylistDictType,NULL);
1006 server.pubsub_patterns = listCreate();
1007 listSetFreeMethod(server.pubsub_patterns,freePubsubPattern);
1008 listSetMatchMethod(server.pubsub_patterns,listMatchPubsubPattern);
1009 server.cronloops = 0;
1010 server.bgsavechildpid = -1;
1011 server.bgrewritechildpid = -1;
1012 server.bgrewritebuf = sdsempty();
1013 server.aofbuf = sdsempty();
1014 server.lastsave = time(NULL);
1015 server.dirty = 0;
1016 server.stat_numcommands = 0;
1017 server.stat_numconnections = 0;
1018 server.stat_expiredkeys = 0;
1019 server.stat_evictedkeys = 0;
1020 server.stat_starttime = time(NULL);
1021 server.stat_keyspace_misses = 0;
1022 server.stat_keyspace_hits = 0;
1023 server.stat_peak_memory = 0;
1024 server.stat_fork_time = 0;
1025 server.stat_rejected_conn = 0;
1026 server.unixtime = time(NULL);
1027 aeCreateTimeEvent(server.el, 1, serverCron, NULL, NULL);
1028 if (server.ipfd > 0 && aeCreateFileEvent(server.el,server.ipfd,AE_READABLE,
1029 acceptTcpHandler,NULL) == AE_ERR) oom("creating file event");
1030 if (server.sofd > 0 && aeCreateFileEvent(server.el,server.sofd,AE_READABLE,
1031 acceptUnixHandler,NULL) == AE_ERR) oom("creating file event");
1032
1033 if (server.appendonly) {
1034 server.appendfd = open(server.appendfilename,O_WRONLY|O_APPEND|O_CREAT,0644);
1035 if (server.appendfd == -1) {
1036 redisLog(REDIS_WARNING, "Can't open the append-only file: %s",
1037 strerror(errno));
1038 exit(1);
1039 }
1040 }
1041
1042 if (server.cluster_enabled) clusterInit();
1043 scriptingInit();
1044 slowlogInit();
1045 bioInit();
1046 srand(time(NULL)^getpid());
1047
1048 /* Try to raise the max number of open files accordingly to the
1049 * configured max number of clients. Also account for 32 additional
1050 * file descriptors as we need a few more for persistence, listening
1051 * sockets, log files and so forth. */
1052 {
1053 rlim_t maxfiles = server.maxclients+32;
1054 struct rlimit limit;
1055
1056 if (maxfiles < 1024) maxfiles = 1024;
1057 if (getrlimit(RLIMIT_NOFILE,&limit) == -1) {
1058 redisLog(REDIS_WARNING,"Unable to obtain the current NOFILE limit (%s), assuming 1024 and setting the max clients configuration accordingly.",
1059 strerror(errno));
1060 server.maxclients = 1024-32;
1061 } else {
1062 rlim_t oldlimit = limit.rlim_cur;
1063
1064 /* Set the max number of files if the current limit is not enough
1065 * for our needs. */
1066 if (oldlimit < maxfiles) {
1067 limit.rlim_cur = maxfiles;
1068 limit.rlim_max = maxfiles;
1069 if (setrlimit(RLIMIT_NOFILE,&limit) == -1) {
1070 server.maxclients = oldlimit-32;
1071 redisLog(REDIS_WARNING,"Unable to set the max number of files limit to %d (%s), setting the max clients configuration to %d.",
1072 (int) maxfiles, strerror(errno), (int) server.maxclients);
1073 } else {
1074 redisLog(REDIS_NOTICE,"Max number of open files set to %d",
1075 (int) maxfiles);
1076 }
1077 }
1078 }
1079 }
1080 }
1081
1082 /* Populates the Redis Command Table starting from the hard coded list
1083 * we have on top of redis.c file. */
1084 void populateCommandTable(void) {
1085 int j;
1086 int numcommands = sizeof(redisCommandTable)/sizeof(struct redisCommand);
1087
1088 for (j = 0; j < numcommands; j++) {
1089 struct redisCommand *c = redisCommandTable+j;
1090 char *f = c->sflags;
1091 int retval;
1092
1093 while(*f != '\0') {
1094 switch(*f) {
1095 case 'w': c->flags |= REDIS_CMD_WRITE; break;
1096 case 'r': c->flags |= REDIS_CMD_READONLY; break;
1097 case 'm': c->flags |= REDIS_CMD_DENYOOM; break;
1098 case 'a': c->flags |= REDIS_CMD_ADMIN; break;
1099 case 'p': c->flags |= REDIS_CMD_PUBSUB; break;
1100 case 'f': c->flags |= REDIS_CMD_FORCE_REPLICATION; break;
1101 case 's': c->flags |= REDIS_CMD_NOSCRIPT; break;
1102 case 'R': c->flags |= REDIS_CMD_RANDOM; break;
1103 default: redisPanic("Unsupported command flag"); break;
1104 }
1105 f++;
1106 }
1107
1108 retval = dictAdd(server.commands, sdsnew(c->name), c);
1109 assert(retval == DICT_OK);
1110 }
1111 }
1112
1113 void resetCommandTableStats(void) {
1114 int numcommands = sizeof(redisCommandTable)/sizeof(struct redisCommand);
1115 int j;
1116
1117 for (j = 0; j < numcommands; j++) {
1118 struct redisCommand *c = redisCommandTable+j;
1119
1120 c->microseconds = 0;
1121 c->calls = 0;
1122 }
1123 }
1124
1125 /* ====================== Commands lookup and execution ===================== */
1126
1127 struct redisCommand *lookupCommand(sds name) {
1128 return dictFetchValue(server.commands, name);
1129 }
1130
1131 struct redisCommand *lookupCommandByCString(char *s) {
1132 struct redisCommand *cmd;
1133 sds name = sdsnew(s);
1134
1135 cmd = dictFetchValue(server.commands, name);
1136 sdsfree(name);
1137 return cmd;
1138 }
1139
1140 /* Call() is the core of Redis execution of a command */
1141 void call(redisClient *c) {
1142 long long dirty, start = ustime(), duration;
1143
1144 dirty = server.dirty;
1145 c->cmd->proc(c);
1146 dirty = server.dirty-dirty;
1147 duration = ustime()-start;
1148 c->cmd->microseconds += duration;
1149 slowlogPushEntryIfNeeded(c->argv,c->argc,duration);
1150 c->cmd->calls++;
1151
1152 if (server.appendonly && dirty > 0)
1153 feedAppendOnlyFile(c->cmd,c->db->id,c->argv,c->argc);
1154 if ((dirty > 0 || c->cmd->flags & REDIS_CMD_FORCE_REPLICATION) &&
1155 listLength(server.slaves))
1156 replicationFeedSlaves(server.slaves,c->db->id,c->argv,c->argc);
1157 if (listLength(server.monitors))
1158 replicationFeedMonitors(server.monitors,c->db->id,c->argv,c->argc);
1159 server.stat_numcommands++;
1160 }
1161
1162 /* If this function gets called we already read a whole
1163 * command, argments are in the client argv/argc fields.
1164 * processCommand() execute the command or prepare the
1165 * server for a bulk read from the client.
1166 *
1167 * If 1 is returned the client is still alive and valid and
1168 * and other operations can be performed by the caller. Otherwise
1169 * if 0 is returned the client was destroied (i.e. after QUIT). */
1170 int processCommand(redisClient *c) {
1171 /* The QUIT command is handled separately. Normal command procs will
1172 * go through checking for replication and QUIT will cause trouble
1173 * when FORCE_REPLICATION is enabled and would be implemented in
1174 * a regular command proc. */
1175 if (!strcasecmp(c->argv[0]->ptr,"quit")) {
1176 addReply(c,shared.ok);
1177 c->flags |= REDIS_CLOSE_AFTER_REPLY;
1178 return REDIS_ERR;
1179 }
1180
1181 /* Now lookup the command and check ASAP about trivial error conditions
1182 * such as wrong arity, bad command name and so forth. */
1183 c->cmd = c->lastcmd = lookupCommand(c->argv[0]->ptr);
1184 if (!c->cmd) {
1185 addReplyErrorFormat(c,"unknown command '%s'",
1186 (char*)c->argv[0]->ptr);
1187 return REDIS_OK;
1188 } else if ((c->cmd->arity > 0 && c->cmd->arity != c->argc) ||
1189 (c->argc < -c->cmd->arity)) {
1190 addReplyErrorFormat(c,"wrong number of arguments for '%s' command",
1191 c->cmd->name);
1192 return REDIS_OK;
1193 }
1194
1195 /* Check if the user is authenticated */
1196 if (server.requirepass && !c->authenticated && c->cmd->proc != authCommand)
1197 {
1198 addReplyError(c,"operation not permitted");
1199 return REDIS_OK;
1200 }
1201
1202 /* If cluster is enabled, redirect here */
1203 if (server.cluster_enabled &&
1204 !(c->cmd->getkeys_proc == NULL && c->cmd->firstkey == 0)) {
1205 int hashslot;
1206
1207 if (server.cluster.state != REDIS_CLUSTER_OK) {
1208 addReplyError(c,"The cluster is down. Check with CLUSTER INFO for more information");
1209 return REDIS_OK;
1210 } else {
1211 int ask;
1212 clusterNode *n = getNodeByQuery(c,c->cmd,c->argv,c->argc,&hashslot,&ask);
1213 if (n == NULL) {
1214 addReplyError(c,"Multi keys request invalid in cluster");
1215 return REDIS_OK;
1216 } else if (n != server.cluster.myself) {
1217 addReplySds(c,sdscatprintf(sdsempty(),
1218 "-%s %d %s:%d\r\n", ask ? "ASK" : "MOVED",
1219 hashslot,n->ip,n->port));
1220 return REDIS_OK;
1221 }
1222 }
1223 }
1224
1225 /* Handle the maxmemory directive.
1226 *
1227 * First we try to free some memory if possible (if there are volatile
1228 * keys in the dataset). If there are not the only thing we can do
1229 * is returning an error. */
1230 if (server.maxmemory) freeMemoryIfNeeded();
1231 if (server.maxmemory && (c->cmd->flags & REDIS_CMD_DENYOOM) &&
1232 zmalloc_used_memory() > server.maxmemory)
1233 {
1234 addReplyError(c,"command not allowed when used memory > 'maxmemory'");
1235 return REDIS_OK;
1236 }
1237
1238 /* Only allow SUBSCRIBE and UNSUBSCRIBE in the context of Pub/Sub */
1239 if ((dictSize(c->pubsub_channels) > 0 || listLength(c->pubsub_patterns) > 0)
1240 &&
1241 c->cmd->proc != subscribeCommand &&
1242 c->cmd->proc != unsubscribeCommand &&
1243 c->cmd->proc != psubscribeCommand &&
1244 c->cmd->proc != punsubscribeCommand) {
1245 addReplyError(c,"only (P)SUBSCRIBE / (P)UNSUBSCRIBE / QUIT allowed in this context");
1246 return REDIS_OK;
1247 }
1248
1249 /* Only allow INFO and SLAVEOF when slave-serve-stale-data is no and
1250 * we are a slave with a broken link with master. */
1251 if (server.masterhost && server.replstate != REDIS_REPL_CONNECTED &&
1252 server.repl_serve_stale_data == 0 &&
1253 c->cmd->proc != infoCommand && c->cmd->proc != slaveofCommand)
1254 {
1255 addReplyError(c,
1256 "link with MASTER is down and slave-serve-stale-data is set to no");
1257 return REDIS_OK;
1258 }
1259
1260 /* Loading DB? Return an error if the command is not INFO */
1261 if (server.loading && c->cmd->proc != infoCommand) {
1262 addReply(c, shared.loadingerr);
1263 return REDIS_OK;
1264 }
1265
1266 /* Lua script too slow? Only allow SHUTDOWN NOSAVE and SCRIPT KILL. */
1267 if (server.lua_timedout &&
1268 !(c->cmd->proc != shutdownCommand &&
1269 c->argc == 2 &&
1270 tolower(((char*)c->argv[1]->ptr)[0]) == 'n') &&
1271 !(c->cmd->proc == scriptCommand &&
1272 c->argc == 2 &&
1273 tolower(((char*)c->argv[1]->ptr)[0]) == 'k'))
1274 {
1275 addReply(c, shared.slowscripterr);
1276 return REDIS_OK;
1277 }
1278
1279 /* Exec the command */
1280 if (c->flags & REDIS_MULTI &&
1281 c->cmd->proc != execCommand && c->cmd->proc != discardCommand &&
1282 c->cmd->proc != multiCommand && c->cmd->proc != watchCommand)
1283 {
1284 queueMultiCommand(c);
1285 addReply(c,shared.queued);
1286 } else {
1287 call(c);
1288 }
1289 return REDIS_OK;
1290 }
1291
1292 /*================================== Shutdown =============================== */
1293
1294 int prepareForShutdown(int flags) {
1295 int save = flags & REDIS_SHUTDOWN_SAVE;
1296 int nosave = flags & REDIS_SHUTDOWN_NOSAVE;
1297
1298 redisLog(REDIS_WARNING,"User requested shutdown...");
1299 /* Kill the saving child if there is a background saving in progress.
1300 We want to avoid race conditions, for instance our saving child may
1301 overwrite the synchronous saving did by SHUTDOWN. */
1302 if (server.bgsavechildpid != -1) {
1303 redisLog(REDIS_WARNING,"There is a child saving an .rdb. Killing it!");
1304 kill(server.bgsavechildpid,SIGKILL);
1305 rdbRemoveTempFile(server.bgsavechildpid);
1306 }
1307 if (server.appendonly) {
1308 /* Kill the AOF saving child as the AOF we already have may be longer
1309 * but contains the full dataset anyway. */
1310 if (server.bgrewritechildpid != -1) {
1311 redisLog(REDIS_WARNING,
1312 "There is a child rewriting the AOF. Killing it!");
1313 kill(server.bgrewritechildpid,SIGKILL);
1314 }
1315 /* Append only file: fsync() the AOF and exit */
1316 redisLog(REDIS_NOTICE,"Calling fsync() on the AOF file.");
1317 aof_fsync(server.appendfd);
1318 }
1319 if ((server.saveparamslen > 0 && !nosave) || save) {
1320 redisLog(REDIS_NOTICE,"Saving the final RDB snapshot before exiting.");
1321 /* Snapshotting. Perform a SYNC SAVE and exit */
1322 if (rdbSave(server.dbfilename) != REDIS_OK) {
1323 /* Ooops.. error saving! The best we can do is to continue
1324 * operating. Note that if there was a background saving process,
1325 * in the next cron() Redis will be notified that the background
1326 * saving aborted, handling special stuff like slaves pending for
1327 * synchronization... */
1328 redisLog(REDIS_WARNING,"Error trying to save the DB, can't exit.");
1329 return REDIS_ERR;
1330 }
1331 }
1332 if (server.daemonize) {
1333 redisLog(REDIS_NOTICE,"Removing the pid file.");
1334 unlink(server.pidfile);
1335 }
1336 /* Close the listening sockets. Apparently this allows faster restarts. */
1337 if (server.ipfd != -1) close(server.ipfd);
1338 if (server.sofd != -1) close(server.sofd);
1339 if (server.unixsocket) {
1340 redisLog(REDIS_NOTICE,"Removing the unix socket file.");
1341 unlink(server.unixsocket); /* don't care if this fails */
1342 }
1343
1344 redisLog(REDIS_WARNING,"Redis is now ready to exit, bye bye...");
1345 return REDIS_OK;
1346 }
1347
1348 /*================================== Commands =============================== */
1349
1350 void authCommand(redisClient *c) {
1351 if (!server.requirepass) {
1352 addReplyError(c,"Client sent AUTH, but no password is set");
1353 } else if (!strcmp(c->argv[1]->ptr, server.requirepass)) {
1354 c->authenticated = 1;
1355 addReply(c,shared.ok);
1356 } else {
1357 c->authenticated = 0;
1358 addReplyError(c,"invalid password");
1359 }
1360 }
1361
1362 void pingCommand(redisClient *c) {
1363 addReply(c,shared.pong);
1364 }
1365
1366 void echoCommand(redisClient *c) {
1367 addReplyBulk(c,c->argv[1]);
1368 }
1369
1370 /* Convert an amount of bytes into a human readable string in the form
1371 * of 100B, 2G, 100M, 4K, and so forth. */
1372 void bytesToHuman(char *s, unsigned long long n) {
1373 double d;
1374
1375 if (n < 1024) {
1376 /* Bytes */
1377 sprintf(s,"%lluB",n);
1378 return;
1379 } else if (n < (1024*1024)) {
1380 d = (double)n/(1024);
1381 sprintf(s,"%.2fK",d);
1382 } else if (n < (1024LL*1024*1024)) {
1383 d = (double)n/(1024*1024);
1384 sprintf(s,"%.2fM",d);
1385 } else if (n < (1024LL*1024*1024*1024)) {
1386 d = (double)n/(1024LL*1024*1024);
1387 sprintf(s,"%.2fG",d);
1388 }
1389 }
1390
1391 /* Create the string returned by the INFO command. This is decoupled
1392 * by the INFO command itself as we need to report the same information
1393 * on memory corruption problems. */
1394 sds genRedisInfoString(char *section) {
1395 sds info = sdsempty();
1396 time_t uptime = time(NULL)-server.stat_starttime;
1397 int j, numcommands;
1398 struct rusage self_ru, c_ru;
1399 unsigned long lol, bib;
1400 int allsections = 0, defsections = 0;
1401 int sections = 0;
1402
1403 if (section) {
1404 allsections = strcasecmp(section,"all") == 0;
1405 defsections = strcasecmp(section,"default") == 0;
1406 }
1407
1408 getrusage(RUSAGE_SELF, &self_ru);
1409 getrusage(RUSAGE_CHILDREN, &c_ru);
1410 getClientsMaxBuffers(&lol,&bib);
1411
1412 /* Server */
1413 if (allsections || defsections || !strcasecmp(section,"server")) {
1414 if (sections++) info = sdscat(info,"\r\n");
1415 info = sdscatprintf(info,
1416 "# Server\r\n"
1417 "redis_version:%s\r\n"
1418 "redis_git_sha1:%s\r\n"
1419 "redis_git_dirty:%d\r\n"
1420 "arch_bits:%s\r\n"
1421 "multiplexing_api:%s\r\n"
1422 "process_id:%ld\r\n"
1423 "tcp_port:%d\r\n"
1424 "uptime_in_seconds:%ld\r\n"
1425 "uptime_in_days:%ld\r\n"
1426 "lru_clock:%ld\r\n",
1427 REDIS_VERSION,
1428 redisGitSHA1(),
1429 strtol(redisGitDirty(),NULL,10) > 0,
1430 (sizeof(long) == 8) ? "64" : "32",
1431 aeGetApiName(),
1432 (long) getpid(),
1433 server.port,
1434 uptime,
1435 uptime/(3600*24),
1436 (unsigned long) server.lruclock);
1437 }
1438
1439 /* Clients */
1440 if (allsections || defsections || !strcasecmp(section,"clients")) {
1441 if (sections++) info = sdscat(info,"\r\n");
1442 info = sdscatprintf(info,
1443 "# Clients\r\n"
1444 "connected_clients:%d\r\n"
1445 "client_longest_output_list:%lu\r\n"
1446 "client_biggest_input_buf:%lu\r\n"
1447 "blocked_clients:%d\r\n",
1448 listLength(server.clients)-listLength(server.slaves),
1449 lol, bib,
1450 server.bpop_blocked_clients);
1451 }
1452
1453 /* Memory */
1454 if (allsections || defsections || !strcasecmp(section,"memory")) {
1455 char hmem[64];
1456 char peak_hmem[64];
1457
1458 bytesToHuman(hmem,zmalloc_used_memory());
1459 bytesToHuman(peak_hmem,server.stat_peak_memory);
1460 if (sections++) info = sdscat(info,"\r\n");
1461 info = sdscatprintf(info,
1462 "# Memory\r\n"
1463 "used_memory:%zu\r\n"
1464 "used_memory_human:%s\r\n"
1465 "used_memory_rss:%zu\r\n"
1466 "used_memory_peak:%zu\r\n"
1467 "used_memory_peak_human:%s\r\n"
1468 "used_memory_lua:%lld\r\n"
1469 "mem_fragmentation_ratio:%.2f\r\n"
1470 "mem_allocator:%s\r\n",
1471 zmalloc_used_memory(),
1472 hmem,
1473 zmalloc_get_rss(),
1474 server.stat_peak_memory,
1475 peak_hmem,
1476 ((long long)lua_gc(server.lua,LUA_GCCOUNT,0))*1024LL,
1477 zmalloc_get_fragmentation_ratio(),
1478 ZMALLOC_LIB
1479 );
1480 }
1481
1482 /* Persistence */
1483 if (allsections || defsections || !strcasecmp(section,"persistence")) {
1484 if (sections++) info = sdscat(info,"\r\n");
1485 info = sdscatprintf(info,
1486 "# Persistence\r\n"
1487 "loading:%d\r\n"
1488 "aof_enabled:%d\r\n"
1489 "changes_since_last_save:%lld\r\n"
1490 "bgsave_in_progress:%d\r\n"
1491 "last_save_time:%ld\r\n"
1492 "bgrewriteaof_in_progress:%d\r\n",
1493 server.loading,
1494 server.appendonly,
1495 server.dirty,
1496 server.bgsavechildpid != -1,
1497 server.lastsave,
1498 server.bgrewritechildpid != -1);
1499
1500 if (server.appendonly) {
1501 info = sdscatprintf(info,
1502 "aof_current_size:%lld\r\n"
1503 "aof_base_size:%lld\r\n"
1504 "aof_pending_rewrite:%d\r\n",
1505 (long long) server.appendonly_current_size,
1506 (long long) server.auto_aofrewrite_base_size,
1507 server.aofrewrite_scheduled);
1508 }
1509
1510 if (server.loading) {
1511 double perc;
1512 time_t eta, elapsed;
1513 off_t remaining_bytes = server.loading_total_bytes-
1514 server.loading_loaded_bytes;
1515
1516 perc = ((double)server.loading_loaded_bytes /
1517 server.loading_total_bytes) * 100;
1518
1519 elapsed = time(NULL)-server.loading_start_time;
1520 if (elapsed == 0) {
1521 eta = 1; /* A fake 1 second figure if we don't have
1522 enough info */
1523 } else {
1524 eta = (elapsed*remaining_bytes)/server.loading_loaded_bytes;
1525 }
1526
1527 info = sdscatprintf(info,
1528 "loading_start_time:%ld\r\n"
1529 "loading_total_bytes:%llu\r\n"
1530 "loading_loaded_bytes:%llu\r\n"
1531 "loading_loaded_perc:%.2f\r\n"
1532 "loading_eta_seconds:%ld\r\n"
1533 ,(unsigned long) server.loading_start_time,
1534 (unsigned long long) server.loading_total_bytes,
1535 (unsigned long long) server.loading_loaded_bytes,
1536 perc,
1537 eta
1538 );
1539 }
1540 }
1541
1542 /* Stats */
1543 if (allsections || defsections || !strcasecmp(section,"stats")) {
1544 if (sections++) info = sdscat(info,"\r\n");
1545 info = sdscatprintf(info,
1546 "# Stats\r\n"
1547 "total_connections_received:%lld\r\n"
1548 "total_commands_processed:%lld\r\n"
1549 "rejected_connections:%lld\r\n"
1550 "expired_keys:%lld\r\n"
1551 "evicted_keys:%lld\r\n"
1552 "keyspace_hits:%lld\r\n"
1553 "keyspace_misses:%lld\r\n"
1554 "pubsub_channels:%ld\r\n"
1555 "pubsub_patterns:%u\r\n"
1556 "latest_fork_usec:%lld\r\n",
1557 server.stat_numconnections,
1558 server.stat_numcommands,
1559 server.stat_rejected_conn,
1560 server.stat_expiredkeys,
1561 server.stat_evictedkeys,
1562 server.stat_keyspace_hits,
1563 server.stat_keyspace_misses,
1564 dictSize(server.pubsub_channels),
1565 listLength(server.pubsub_patterns),
1566 server.stat_fork_time);
1567 }
1568
1569 /* Replication */
1570 if (allsections || defsections || !strcasecmp(section,"replication")) {
1571 if (sections++) info = sdscat(info,"\r\n");
1572 info = sdscatprintf(info,
1573 "# Replication\r\n"
1574 "role:%s\r\n",
1575 server.masterhost == NULL ? "master" : "slave");
1576 if (server.masterhost) {
1577 info = sdscatprintf(info,
1578 "master_host:%s\r\n"
1579 "master_port:%d\r\n"
1580 "master_link_status:%s\r\n"
1581 "master_last_io_seconds_ago:%d\r\n"
1582 "master_sync_in_progress:%d\r\n"
1583 ,server.masterhost,
1584 server.masterport,
1585 (server.replstate == REDIS_REPL_CONNECTED) ?
1586 "up" : "down",
1587 server.master ?
1588 ((int)(time(NULL)-server.master->lastinteraction)) : -1,
1589 server.replstate == REDIS_REPL_TRANSFER
1590 );
1591
1592 if (server.replstate == REDIS_REPL_TRANSFER) {
1593 info = sdscatprintf(info,
1594 "master_sync_left_bytes:%ld\r\n"
1595 "master_sync_last_io_seconds_ago:%d\r\n"
1596 ,(long)server.repl_transfer_left,
1597 (int)(time(NULL)-server.repl_transfer_lastio)
1598 );
1599 }
1600
1601 if (server.replstate != REDIS_REPL_CONNECTED) {
1602 info = sdscatprintf(info,
1603 "master_link_down_since_seconds:%ld\r\n",
1604 (long)time(NULL)-server.repl_down_since);
1605 }
1606 }
1607 info = sdscatprintf(info,
1608 "connected_slaves:%d\r\n",
1609 listLength(server.slaves));
1610 }
1611
1612 /* CPU */
1613 if (allsections || defsections || !strcasecmp(section,"cpu")) {
1614 if (sections++) info = sdscat(info,"\r\n");
1615 info = sdscatprintf(info,
1616 "# CPU\r\n"
1617 "used_cpu_sys:%.2f\r\n"
1618 "used_cpu_user:%.2f\r\n"
1619 "used_cpu_sys_children:%.2f\r\n"
1620 "used_cpu_user_children:%.2f\r\n",
1621 (float)self_ru.ru_stime.tv_sec+(float)self_ru.ru_stime.tv_usec/1000000,
1622 (float)self_ru.ru_utime.tv_sec+(float)self_ru.ru_utime.tv_usec/1000000,
1623 (float)c_ru.ru_stime.tv_sec+(float)c_ru.ru_stime.tv_usec/1000000,
1624 (float)c_ru.ru_utime.tv_sec+(float)c_ru.ru_utime.tv_usec/1000000);
1625 }
1626
1627 /* cmdtime */
1628 if (allsections || !strcasecmp(section,"commandstats")) {
1629 if (sections++) info = sdscat(info,"\r\n");
1630 info = sdscatprintf(info, "# Commandstats\r\n");
1631 numcommands = sizeof(redisCommandTable)/sizeof(struct redisCommand);
1632 for (j = 0; j < numcommands; j++) {
1633 struct redisCommand *c = redisCommandTable+j;
1634
1635 if (!c->calls) continue;
1636 info = sdscatprintf(info,
1637 "cmdstat_%s:calls=%lld,usec=%lld,usec_per_call=%.2f\r\n",
1638 c->name, c->calls, c->microseconds,
1639 (c->calls == 0) ? 0 : ((float)c->microseconds/c->calls));
1640 }
1641 }
1642
1643 /* Clusetr */
1644 if (allsections || defsections || !strcasecmp(section,"cluster")) {
1645 if (sections++) info = sdscat(info,"\r\n");
1646 info = sdscatprintf(info,
1647 "# Cluster\r\n"
1648 "cluster_enabled:%d\r\n",
1649 server.cluster_enabled);
1650 }
1651
1652 /* Key space */
1653 if (allsections || defsections || !strcasecmp(section,"keyspace")) {
1654 if (sections++) info = sdscat(info,"\r\n");
1655 info = sdscatprintf(info, "# Keyspace\r\n");
1656 for (j = 0; j < server.dbnum; j++) {
1657 long long keys, vkeys;
1658
1659 keys = dictSize(server.db[j].dict);
1660 vkeys = dictSize(server.db[j].expires);
1661 if (keys || vkeys) {
1662 info = sdscatprintf(info, "db%d:keys=%lld,expires=%lld\r\n",
1663 j, keys, vkeys);
1664 }
1665 }
1666 }
1667 return info;
1668 }
1669
1670 void infoCommand(redisClient *c) {
1671 char *section = c->argc == 2 ? c->argv[1]->ptr : "default";
1672
1673 if (c->argc > 2) {
1674 addReply(c,shared.syntaxerr);
1675 return;
1676 }
1677 sds info = genRedisInfoString(section);
1678 addReplySds(c,sdscatprintf(sdsempty(),"$%lu\r\n",
1679 (unsigned long)sdslen(info)));
1680 addReplySds(c,info);
1681 addReply(c,shared.crlf);
1682 }
1683
1684 void monitorCommand(redisClient *c) {
1685 /* ignore MONITOR if aleady slave or in monitor mode */
1686 if (c->flags & REDIS_SLAVE) return;
1687
1688 c->flags |= (REDIS_SLAVE|REDIS_MONITOR);
1689 c->slaveseldb = 0;
1690 listAddNodeTail(server.monitors,c);
1691 addReply(c,shared.ok);
1692 }
1693
1694 /* ============================ Maxmemory directive ======================== */
1695
1696 /* This function gets called when 'maxmemory' is set on the config file to limit
1697 * the max memory used by the server, and we are out of memory.
1698 * This function will try to, in order:
1699 *
1700 * - Free objects from the free list
1701 * - Try to remove keys with an EXPIRE set
1702 *
1703 * It is not possible to free enough memory to reach used-memory < maxmemory
1704 * the server will start refusing commands that will enlarge even more the
1705 * memory usage.
1706 */
1707 void freeMemoryIfNeeded(void) {
1708 /* Remove keys accordingly to the active policy as long as we are
1709 * over the memory limit. */
1710 if (server.maxmemory_policy == REDIS_MAXMEMORY_NO_EVICTION) return;
1711
1712 while (server.maxmemory && zmalloc_used_memory() > server.maxmemory) {
1713 int j, k, freed = 0;
1714
1715 for (j = 0; j < server.dbnum; j++) {
1716 long bestval = 0; /* just to prevent warning */
1717 sds bestkey = NULL;
1718 struct dictEntry *de;
1719 redisDb *db = server.db+j;
1720 dict *dict;
1721
1722 if (server.maxmemory_policy == REDIS_MAXMEMORY_ALLKEYS_LRU ||
1723 server.maxmemory_policy == REDIS_MAXMEMORY_ALLKEYS_RANDOM)
1724 {
1725 dict = server.db[j].dict;
1726 } else {
1727 dict = server.db[j].expires;
1728 }
1729 if (dictSize(dict) == 0) continue;
1730
1731 /* volatile-random and allkeys-random policy */
1732 if (server.maxmemory_policy == REDIS_MAXMEMORY_ALLKEYS_RANDOM ||
1733 server.maxmemory_policy == REDIS_MAXMEMORY_VOLATILE_RANDOM)
1734 {
1735 de = dictGetRandomKey(dict);
1736 bestkey = dictGetKey(de);
1737 }
1738
1739 /* volatile-lru and allkeys-lru policy */
1740 else if (server.maxmemory_policy == REDIS_MAXMEMORY_ALLKEYS_LRU ||
1741 server.maxmemory_policy == REDIS_MAXMEMORY_VOLATILE_LRU)
1742 {
1743 for (k = 0; k < server.maxmemory_samples; k++) {
1744 sds thiskey;
1745 long thisval;
1746 robj *o;
1747
1748 de = dictGetRandomKey(dict);
1749 thiskey = dictGetKey(de);
1750 /* When policy is volatile-lru we need an additonal lookup
1751 * to locate the real key, as dict is set to db->expires. */
1752 if (server.maxmemory_policy == REDIS_MAXMEMORY_VOLATILE_LRU)
1753 de = dictFind(db->dict, thiskey);
1754 o = dictGetVal(de);
1755 thisval = estimateObjectIdleTime(o);
1756
1757 /* Higher idle time is better candidate for deletion */
1758 if (bestkey == NULL || thisval > bestval) {
1759 bestkey = thiskey;
1760 bestval = thisval;
1761 }
1762 }
1763 }
1764
1765 /* volatile-ttl */
1766 else if (server.maxmemory_policy == REDIS_MAXMEMORY_VOLATILE_TTL) {
1767 for (k = 0; k < server.maxmemory_samples; k++) {
1768 sds thiskey;
1769 long thisval;
1770
1771 de = dictGetRandomKey(dict);
1772 thiskey = dictGetKey(de);
1773 thisval = (long) dictGetVal(de);
1774
1775 /* Expire sooner (minor expire unix timestamp) is better
1776 * candidate for deletion */
1777 if (bestkey == NULL || thisval < bestval) {
1778 bestkey = thiskey;
1779 bestval = thisval;
1780 }
1781 }
1782 }
1783
1784 /* Finally remove the selected key. */
1785 if (bestkey) {
1786 robj *keyobj = createStringObject(bestkey,sdslen(bestkey));
1787 propagateExpire(db,keyobj);
1788 dbDelete(db,keyobj);
1789 server.stat_evictedkeys++;
1790 decrRefCount(keyobj);
1791 freed++;
1792 }
1793 }
1794 if (!freed) return; /* nothing to free... */
1795 }
1796 }
1797
1798 /* =================================== Main! ================================ */
1799
1800 #ifdef __linux__
1801 int linuxOvercommitMemoryValue(void) {
1802 FILE *fp = fopen("/proc/sys/vm/overcommit_memory","r");
1803 char buf[64];
1804
1805 if (!fp) return -1;
1806 if (fgets(buf,64,fp) == NULL) {
1807 fclose(fp);
1808 return -1;
1809 }
1810 fclose(fp);
1811
1812 return atoi(buf);
1813 }
1814
1815 void linuxOvercommitMemoryWarning(void) {
1816 if (linuxOvercommitMemoryValue() == 0) {
1817 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.");
1818 }
1819 }
1820 #endif /* __linux__ */
1821
1822 void createPidFile(void) {
1823 /* Try to write the pid file in a best-effort way. */
1824 FILE *fp = fopen(server.pidfile,"w");
1825 if (fp) {
1826 fprintf(fp,"%d\n",(int)getpid());
1827 fclose(fp);
1828 }
1829 }
1830
1831 void daemonize(void) {
1832 int fd;
1833
1834 if (fork() != 0) exit(0); /* parent exits */
1835 setsid(); /* create a new session */
1836
1837 /* Every output goes to /dev/null. If Redis is daemonized but
1838 * the 'logfile' is set to 'stdout' in the configuration file
1839 * it will not log at all. */
1840 if ((fd = open("/dev/null", O_RDWR, 0)) != -1) {
1841 dup2(fd, STDIN_FILENO);
1842 dup2(fd, STDOUT_FILENO);
1843 dup2(fd, STDERR_FILENO);
1844 if (fd > STDERR_FILENO) close(fd);
1845 }
1846 }
1847
1848 void version() {
1849 printf("Redis server version %s (%s:%d)\n", REDIS_VERSION,
1850 redisGitSHA1(), atoi(redisGitDirty()) > 0);
1851 exit(0);
1852 }
1853
1854 void usage() {
1855 fprintf(stderr,"Usage: ./redis-server [/path/to/redis.conf]\n");
1856 fprintf(stderr," ./redis-server - (read config from stdin)\n");
1857 exit(1);
1858 }
1859
1860 void redisAsciiArt(void) {
1861 #include "asciilogo.h"
1862 char *buf = zmalloc(1024*16);
1863
1864 snprintf(buf,1024*16,ascii_logo,
1865 REDIS_VERSION,
1866 redisGitSHA1(),
1867 strtol(redisGitDirty(),NULL,10) > 0,
1868 (sizeof(long) == 8) ? "64" : "32",
1869 server.cluster_enabled ? "cluster" : "stand alone",
1870 server.port,
1871 (long) getpid()
1872 );
1873 redisLogRaw(REDIS_NOTICE|REDIS_LOG_RAW,buf);
1874 zfree(buf);
1875 }
1876
1877 int main(int argc, char **argv) {
1878 long long start;
1879
1880 zmalloc_enable_thread_safeness();
1881 initServerConfig();
1882 if (argc == 2) {
1883 if (strcmp(argv[1], "-v") == 0 ||
1884 strcmp(argv[1], "--version") == 0) version();
1885 if (strcmp(argv[1], "--help") == 0) usage();
1886 resetServerSaveParams();
1887 loadServerConfig(argv[1]);
1888 } else if ((argc > 2)) {
1889 usage();
1890 } else {
1891 redisLog(REDIS_WARNING,"Warning: no config file specified, using the default config. In order to specify a config file use 'redis-server /path/to/redis.conf'");
1892 }
1893 if (server.daemonize) daemonize();
1894 initServer();
1895 if (server.daemonize) createPidFile();
1896 redisAsciiArt();
1897 redisLog(REDIS_NOTICE,"Server started, Redis version " REDIS_VERSION);
1898 #ifdef __linux__
1899 linuxOvercommitMemoryWarning();
1900 #endif
1901 start = ustime();
1902 if (server.appendonly) {
1903 if (loadAppendOnlyFile(server.appendfilename) == REDIS_OK)
1904 redisLog(REDIS_NOTICE,"DB loaded from append only file: %.3f seconds",(float)(ustime()-start)/1000000);
1905 } else {
1906 if (rdbLoad(server.dbfilename) == REDIS_OK) {
1907 redisLog(REDIS_NOTICE,"DB loaded from disk: %.3f seconds",
1908 (float)(ustime()-start)/1000000);
1909 } else if (errno != ENOENT) {
1910 redisLog(REDIS_WARNING,"Fatal error loading the DB. Exiting.");
1911 exit(1);
1912 }
1913 }
1914 if (server.ipfd > 0)
1915 redisLog(REDIS_NOTICE,"The server is now ready to accept connections on port %d", server.port);
1916 if (server.sofd > 0)
1917 redisLog(REDIS_NOTICE,"The server is now ready to accept connections at %s", server.unixsocket);
1918 aeSetBeforeSleepProc(server.el,beforeSleep);
1919 aeMain(server.el);
1920 aeDeleteEventLoop(server.el);
1921 return 0;
1922 }
1923
1924 #ifdef HAVE_BACKTRACE
1925 static void *getMcontextEip(ucontext_t *uc) {
1926 #if defined(__FreeBSD__)
1927 return (void*) uc->uc_mcontext.mc_eip;
1928 #elif defined(__dietlibc__)
1929 return (void*) uc->uc_mcontext.eip;
1930 #elif defined(__APPLE__) && !defined(MAC_OS_X_VERSION_10_6)
1931 #if __x86_64__
1932 return (void*) uc->uc_mcontext->__ss.__rip;
1933 #elif __i386__
1934 return (void*) uc->uc_mcontext->__ss.__eip;
1935 #else
1936 return (void*) uc->uc_mcontext->__ss.__srr0;
1937 #endif
1938 #elif defined(__APPLE__) && defined(MAC_OS_X_VERSION_10_6)
1939 #if defined(_STRUCT_X86_THREAD_STATE64) && !defined(__i386__)
1940 return (void*) uc->uc_mcontext->__ss.__rip;
1941 #else
1942 return (void*) uc->uc_mcontext->__ss.__eip;
1943 #endif
1944 #elif defined(__i386__)
1945 return (void*) uc->uc_mcontext.gregs[14]; /* Linux 32 */
1946 #elif defined(__X86_64__) || defined(__x86_64__)
1947 return (void*) uc->uc_mcontext.gregs[16]; /* Linux 64 */
1948 #elif defined(__ia64__) /* Linux IA64 */
1949 return (void*) uc->uc_mcontext.sc_ip;
1950 #else
1951 return NULL;
1952 #endif
1953 }
1954
1955 void bugReportStart(void) {
1956 if (server.bug_report_start == 0) {
1957 redisLog(REDIS_WARNING,
1958 "=== REDIS BUG REPORT START: Cut & paste starting from here ===");
1959 server.bug_report_start = 1;
1960 }
1961 }
1962
1963 static void sigsegvHandler(int sig, siginfo_t *info, void *secret) {
1964 void *trace[100];
1965 char **messages = NULL;
1966 int i, trace_size = 0;
1967 ucontext_t *uc = (ucontext_t*) secret;
1968 sds infostring, clients;
1969 struct sigaction act;
1970 REDIS_NOTUSED(info);
1971
1972 bugReportStart();
1973 redisLog(REDIS_WARNING,
1974 " Redis %s crashed by signal: %d", REDIS_VERSION, sig);
1975 redisLog(REDIS_WARNING,
1976 " Failed assertion: %s (%s:%d)", server.assert_failed,
1977 server.assert_file, server.assert_line);
1978
1979 /* Generate the stack trace */
1980 trace_size = backtrace(trace, 100);
1981
1982 /* overwrite sigaction with caller's address */
1983 if (getMcontextEip(uc) != NULL) {
1984 trace[1] = getMcontextEip(uc);
1985 }
1986 messages = backtrace_symbols(trace, trace_size);
1987 redisLog(REDIS_WARNING, "--- STACK TRACE");
1988 for (i=1; i<trace_size; ++i)
1989 redisLog(REDIS_WARNING,"%s", messages[i]);
1990
1991 /* Log INFO and CLIENT LIST */
1992 redisLog(REDIS_WARNING, "--- INFO OUTPUT");
1993 infostring = genRedisInfoString("all");
1994 redisLogRaw(REDIS_WARNING, infostring);
1995 redisLog(REDIS_WARNING, "--- CLIENT LIST OUTPUT");
1996 clients = getAllClientsInfoString();
1997 redisLogRaw(REDIS_WARNING, clients);
1998 /* Don't sdsfree() strings to avoid a crash. Memory may be corrupted. */
1999
2000 redisLog(REDIS_WARNING,
2001 "=== REDIS BUG REPORT END. Make sure to include from START to END. ===\n\n"
2002 " Please report the crash opening an issue on github:\n\n"
2003 " http://github.com/antirez/redis/issues\n\n"
2004 );
2005 /* free(messages); Don't call free() with possibly corrupted memory. */
2006 if (server.daemonize) unlink(server.pidfile);
2007
2008 /* Make sure we exit with the right signal at the end. So for instance
2009 * the core will be dumped if enabled. */
2010 sigemptyset (&act.sa_mask);
2011 /* When the SA_SIGINFO flag is set in sa_flags then sa_sigaction
2012 * is used. Otherwise, sa_handler is used */
2013 act.sa_flags = SA_NODEFER | SA_ONSTACK | SA_RESETHAND;
2014 act.sa_handler = SIG_DFL;
2015 sigaction (sig, &act, NULL);
2016 kill(getpid(),sig);
2017 }
2018 #endif /* HAVE_BACKTRACE */
2019
2020 static void sigtermHandler(int sig) {
2021 REDIS_NOTUSED(sig);
2022
2023 redisLog(REDIS_WARNING,"Received SIGTERM, scheduling shutdown...");
2024 server.shutdown_asap = 1;
2025 }
2026
2027 void setupSignalHandlers(void) {
2028 struct sigaction act;
2029
2030 /* When the SA_SIGINFO flag is set in sa_flags then sa_sigaction is used.
2031 * Otherwise, sa_handler is used. */
2032 sigemptyset(&act.sa_mask);
2033 act.sa_flags = SA_NODEFER | SA_ONSTACK | SA_RESETHAND;
2034 act.sa_handler = sigtermHandler;
2035 sigaction(SIGTERM, &act, NULL);
2036
2037 #ifdef HAVE_BACKTRACE
2038 sigemptyset(&act.sa_mask);
2039 act.sa_flags = SA_NODEFER | SA_ONSTACK | SA_RESETHAND | SA_SIGINFO;
2040 act.sa_sigaction = sigsegvHandler;
2041 sigaction(SIGSEGV, &act, NULL);
2042 sigaction(SIGBUS, &act, NULL);
2043 sigaction(SIGFPE, &act, NULL);
2044 sigaction(SIGILL, &act, NULL);
2045 #endif
2046 return;
2047 }
2048
2049 /* The End */