]> git.saurik.com Git - redis.git/blob - src/redis.c
Two small fixes to maxclients handling.
[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 /* Alternate stack for SIGSEGV/etc handlers */
66 char altstack[SIGSTKSZ];
67
68 /* Global vars */
69 struct redisServer server; /* server global state */
70 struct redisCommand *commandTable;
71
72 /* Our command table.
73 *
74 * Every entry is composed of the following fields:
75 *
76 * name: a string representing the command name.
77 * function: pointer to the C function implementing the command.
78 * arity: number of arguments, it is possible to use -N to say >= N
79 * sflags: command flags as string. See below for a table of flags.
80 * flags: flags as bitmask. Computed by Redis using the 'sflags' field.
81 * get_keys_proc: an optional function to get key arguments from a command.
82 * This is only used when the following three fields are not
83 * enough to specify what arguments are keys.
84 * first_key_index: first argument that is a key
85 * last_key_index: last argument that is a key
86 * key_step: step to get all the keys from first to last argument. For instance
87 * in MSET the step is two since arguments are key,val,key,val,...
88 * microseconds: microseconds of total execution time for this command.
89 * calls: total number of calls of this command.
90 *
91 * The flags, microseconds and calls fields are computed by Redis and should
92 * always be set to zero.
93 *
94 * Command flags are expressed using strings where every character represents
95 * a flag. Later the populateCommandTable() function will take care of
96 * populating the real 'flags' field using this characters.
97 *
98 * This is the meaning of the flags:
99 *
100 * w: write command (may modify the key space).
101 * r: read command (will never modify the key space).
102 * m: may increase memory usage once called. Don't allow if out of memory.
103 * a: admin command, like SAVE or SHUTDOWN.
104 * p: Pub/Sub related command.
105 * f: force replication of this command, regarless of server.dirty.
106 * s: command not allowed in scripts.
107 * R: random command. Command is not deterministic, that is, the same command
108 * with the same arguments, with the same key space, may have different
109 * results. For instance SPOP and RANDOMKEY are two random commands.
110 * S: Sort command output array if called from script, so that the output
111 * is deterministic.
112 */
113 struct redisCommand redisCommandTable[] = {
114 {"get",getCommand,2,"r",0,NULL,1,1,1,0,0},
115 {"set",setCommand,3,"wm",0,noPreloadGetKeys,1,1,1,0,0},
116 {"setnx",setnxCommand,3,"wm",0,noPreloadGetKeys,1,1,1,0,0},
117 {"setex",setexCommand,4,"wm",0,noPreloadGetKeys,1,1,1,0,0},
118 {"psetex",psetexCommand,4,"wm",0,noPreloadGetKeys,1,1,1,0,0},
119 {"append",appendCommand,3,"wm",0,NULL,1,1,1,0,0},
120 {"strlen",strlenCommand,2,"r",0,NULL,1,1,1,0,0},
121 {"del",delCommand,-2,"w",0,noPreloadGetKeys,1,-1,1,0,0},
122 {"exists",existsCommand,2,"r",0,NULL,1,1,1,0,0},
123 {"setbit",setbitCommand,4,"wm",0,NULL,1,1,1,0,0},
124 {"getbit",getbitCommand,3,"r",0,NULL,1,1,1,0,0},
125 {"setrange",setrangeCommand,4,"wm",0,NULL,1,1,1,0,0},
126 {"getrange",getrangeCommand,4,"r",0,NULL,1,1,1,0,0},
127 {"substr",getrangeCommand,4,"r",0,NULL,1,1,1,0,0},
128 {"incr",incrCommand,2,"wm",0,NULL,1,1,1,0,0},
129 {"decr",decrCommand,2,"wm",0,NULL,1,1,1,0,0},
130 {"mget",mgetCommand,-2,"r",0,NULL,1,-1,1,0,0},
131 {"rpush",rpushCommand,-3,"wm",0,NULL,1,1,1,0,0},
132 {"lpush",lpushCommand,-3,"wm",0,NULL,1,1,1,0,0},
133 {"rpushx",rpushxCommand,3,"wm",0,NULL,1,1,1,0,0},
134 {"lpushx",lpushxCommand,3,"wm",0,NULL,1,1,1,0,0},
135 {"linsert",linsertCommand,5,"wm",0,NULL,1,1,1,0,0},
136 {"rpop",rpopCommand,2,"w",0,NULL,1,1,1,0,0},
137 {"lpop",lpopCommand,2,"w",0,NULL,1,1,1,0,0},
138 {"brpop",brpopCommand,-3,"ws",0,NULL,1,1,1,0,0},
139 {"brpoplpush",brpoplpushCommand,4,"wms",0,NULL,1,2,1,0,0},
140 {"blpop",blpopCommand,-3,"ws",0,NULL,1,-2,1,0,0},
141 {"llen",llenCommand,2,"r",0,NULL,1,1,1,0,0},
142 {"lindex",lindexCommand,3,"r",0,NULL,1,1,1,0,0},
143 {"lset",lsetCommand,4,"wm",0,NULL,1,1,1,0,0},
144 {"lrange",lrangeCommand,4,"r",0,NULL,1,1,1,0,0},
145 {"ltrim",ltrimCommand,4,"w",0,NULL,1,1,1,0,0},
146 {"lrem",lremCommand,4,"w",0,NULL,1,1,1,0,0},
147 {"rpoplpush",rpoplpushCommand,3,"wm",0,NULL,1,2,1,0,0},
148 {"sadd",saddCommand,-3,"wm",0,NULL,1,1,1,0,0},
149 {"srem",sremCommand,-3,"w",0,NULL,1,1,1,0,0},
150 {"smove",smoveCommand,4,"w",0,NULL,1,2,1,0,0},
151 {"sismember",sismemberCommand,3,"r",0,NULL,1,1,1,0,0},
152 {"scard",scardCommand,2,"r",0,NULL,1,1,1,0,0},
153 {"spop",spopCommand,2,"wRs",0,NULL,1,1,1,0,0},
154 {"srandmember",srandmemberCommand,2,"rR",0,NULL,1,1,1,0,0},
155 {"sinter",sinterCommand,-2,"rS",0,NULL,1,-1,1,0,0},
156 {"sinterstore",sinterstoreCommand,-3,"wm",0,NULL,1,-1,1,0,0},
157 {"sunion",sunionCommand,-2,"rS",0,NULL,1,-1,1,0,0},
158 {"sunionstore",sunionstoreCommand,-3,"wm",0,NULL,1,-1,1,0,0},
159 {"sdiff",sdiffCommand,-2,"rS",0,NULL,1,-1,1,0,0},
160 {"sdiffstore",sdiffstoreCommand,-3,"wm",0,NULL,1,-1,1,0,0},
161 {"smembers",sinterCommand,2,"rS",0,NULL,1,1,1,0,0},
162 {"zadd",zaddCommand,-4,"wm",0,NULL,1,1,1,0,0},
163 {"zincrby",zincrbyCommand,4,"wm",0,NULL,1,1,1,0,0},
164 {"zrem",zremCommand,-3,"w",0,NULL,1,1,1,0,0},
165 {"zremrangebyscore",zremrangebyscoreCommand,4,"w",0,NULL,1,1,1,0,0},
166 {"zremrangebyrank",zremrangebyrankCommand,4,"w",0,NULL,1,1,1,0,0},
167 {"zunionstore",zunionstoreCommand,-4,"wm",0,zunionInterGetKeys,0,0,0,0,0},
168 {"zinterstore",zinterstoreCommand,-4,"wm",0,zunionInterGetKeys,0,0,0,0,0},
169 {"zrange",zrangeCommand,-4,"r",0,NULL,1,1,1,0,0},
170 {"zrangebyscore",zrangebyscoreCommand,-4,"r",0,NULL,1,1,1,0,0},
171 {"zrevrangebyscore",zrevrangebyscoreCommand,-4,"r",0,NULL,1,1,1,0,0},
172 {"zcount",zcountCommand,4,"r",0,NULL,1,1,1,0,0},
173 {"zrevrange",zrevrangeCommand,-4,"r",0,NULL,1,1,1,0,0},
174 {"zcard",zcardCommand,2,"r",0,NULL,1,1,1,0,0},
175 {"zscore",zscoreCommand,3,"r",0,NULL,1,1,1,0,0},
176 {"zrank",zrankCommand,3,"r",0,NULL,1,1,1,0,0},
177 {"zrevrank",zrevrankCommand,3,"r",0,NULL,1,1,1,0,0},
178 {"hset",hsetCommand,4,"wm",0,NULL,1,1,1,0,0},
179 {"hsetnx",hsetnxCommand,4,"wm",0,NULL,1,1,1,0,0},
180 {"hget",hgetCommand,3,"r",0,NULL,1,1,1,0,0},
181 {"hmset",hmsetCommand,-4,"wm",0,NULL,1,1,1,0,0},
182 {"hmget",hmgetCommand,-3,"r",0,NULL,1,1,1,0,0},
183 {"hincrby",hincrbyCommand,4,"wm",0,NULL,1,1,1,0,0},
184 {"hincrbyfloat",hincrbyfloatCommand,4,"wm",0,NULL,1,1,1,0,0},
185 {"hdel",hdelCommand,-3,"w",0,NULL,1,1,1,0,0},
186 {"hlen",hlenCommand,2,"r",0,NULL,1,1,1,0,0},
187 {"hkeys",hkeysCommand,2,"rS",0,NULL,1,1,1,0,0},
188 {"hvals",hvalsCommand,2,"rS",0,NULL,1,1,1,0,0},
189 {"hgetall",hgetallCommand,2,"r",0,NULL,1,1,1,0,0},
190 {"hexists",hexistsCommand,3,"r",0,NULL,1,1,1,0,0},
191 {"incrby",incrbyCommand,3,"wm",0,NULL,1,1,1,0,0},
192 {"decrby",decrbyCommand,3,"wm",0,NULL,1,1,1,0,0},
193 {"incrbyfloat",incrbyfloatCommand,3,"wm",0,NULL,1,1,1,0,0},
194 {"getset",getsetCommand,3,"wm",0,NULL,1,1,1,0,0},
195 {"mset",msetCommand,-3,"wm",0,NULL,1,-1,2,0,0},
196 {"msetnx",msetnxCommand,-3,"wm",0,NULL,1,-1,2,0,0},
197 {"randomkey",randomkeyCommand,1,"rR",0,NULL,0,0,0,0,0},
198 {"select",selectCommand,2,"r",0,NULL,0,0,0,0,0},
199 {"move",moveCommand,3,"w",0,NULL,1,1,1,0,0},
200 {"rename",renameCommand,3,"w",0,renameGetKeys,1,2,1,0,0},
201 {"renamenx",renamenxCommand,3,"w",0,renameGetKeys,1,2,1,0,0},
202 {"expire",expireCommand,3,"w",0,NULL,1,1,1,0,0},
203 {"expireat",expireatCommand,3,"w",0,NULL,1,1,1,0,0},
204 {"pexpire",pexpireCommand,3,"w",0,NULL,1,1,1,0,0},
205 {"pexpireat",pexpireatCommand,3,"w",0,NULL,1,1,1,0,0},
206 {"keys",keysCommand,2,"rS",0,NULL,0,0,0,0,0},
207 {"dbsize",dbsizeCommand,1,"r",0,NULL,0,0,0,0,0},
208 {"auth",authCommand,2,"rs",0,NULL,0,0,0,0,0},
209 {"ping",pingCommand,1,"r",0,NULL,0,0,0,0,0},
210 {"echo",echoCommand,2,"r",0,NULL,0,0,0,0,0},
211 {"save",saveCommand,1,"ars",0,NULL,0,0,0,0,0},
212 {"bgsave",bgsaveCommand,1,"ar",0,NULL,0,0,0,0,0},
213 {"bgrewriteaof",bgrewriteaofCommand,1,"ar",0,NULL,0,0,0,0,0},
214 {"shutdown",shutdownCommand,-1,"ar",0,NULL,0,0,0,0,0},
215 {"lastsave",lastsaveCommand,1,"r",0,NULL,0,0,0,0,0},
216 {"type",typeCommand,2,"r",0,NULL,1,1,1,0,0},
217 {"multi",multiCommand,1,"rs",0,NULL,0,0,0,0,0},
218 {"exec",execCommand,1,"s",0,NULL,0,0,0,0,0},
219 {"discard",discardCommand,1,"rs",0,NULL,0,0,0,0,0},
220 {"sync",syncCommand,1,"ars",0,NULL,0,0,0,0,0},
221 {"flushdb",flushdbCommand,1,"w",0,NULL,0,0,0,0,0},
222 {"flushall",flushallCommand,1,"w",0,NULL,0,0,0,0,0},
223 {"sort",sortCommand,-2,"wmS",0,NULL,1,1,1,0,0},
224 {"info",infoCommand,-1,"r",0,NULL,0,0,0,0,0},
225 {"monitor",monitorCommand,1,"ars",0,NULL,0,0,0,0,0},
226 {"ttl",ttlCommand,2,"r",0,NULL,1,1,1,0,0},
227 {"pttl",pttlCommand,2,"r",0,NULL,1,1,1,0,0},
228 {"persist",persistCommand,2,"w",0,NULL,1,1,1,0,0},
229 {"slaveof",slaveofCommand,3,"as",0,NULL,0,0,0,0,0},
230 {"debug",debugCommand,-2,"as",0,NULL,0,0,0,0,0},
231 {"config",configCommand,-2,"ar",0,NULL,0,0,0,0,0},
232 {"subscribe",subscribeCommand,-2,"rps",0,NULL,0,0,0,0,0},
233 {"unsubscribe",unsubscribeCommand,-1,"rps",0,NULL,0,0,0,0,0},
234 {"psubscribe",psubscribeCommand,-2,"rps",0,NULL,0,0,0,0,0},
235 {"punsubscribe",punsubscribeCommand,-1,"rps",0,NULL,0,0,0,0,0},
236 {"publish",publishCommand,3,"pf",0,NULL,0,0,0,0,0},
237 {"watch",watchCommand,-2,"rs",0,noPreloadGetKeys,1,-1,1,0,0},
238 {"unwatch",unwatchCommand,1,"rs",0,NULL,0,0,0,0,0},
239 {"restore",restoreCommand,4,"awm",0,NULL,1,1,1,0,0},
240 {"migrate",migrateCommand,6,"aw",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,"s",0,zunionInterGetKeys,0,0,0,0,0},
245 {"evalsha",evalShaCommand,-3,"s",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 {"time",timeCommand,1,"rR",0,NULL,0,0,0,0,0}
249 };
250
251 /*============================ Utility functions ============================ */
252
253 /* Low level logging. To use only for very big messages, otherwise
254 * redisLog() is to prefer. */
255 void redisLogRaw(int level, const char *msg) {
256 const int syslogLevelMap[] = { LOG_DEBUG, LOG_INFO, LOG_NOTICE, LOG_WARNING };
257 const char *c = ".-*#";
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 int off;
272 struct timeval tv;
273
274 gettimeofday(&tv,NULL);
275 off = strftime(buf,sizeof(buf),"%d %b %H:%M:%S.",localtime(&tv.tv_sec));
276 snprintf(buf+off,sizeof(buf)-off,"%03d",(int)tv.tv_usec/1000);
277 fprintf(fp,"[%d] %s %c %s\n",(int)getpid(),buf,c[level],msg);
278 }
279 fflush(fp);
280
281 if (server.logfile) fclose(fp);
282
283 if (server.syslog_enabled) syslog(syslogLevelMap[level], "%s", msg);
284 }
285
286 /* Like redisLogRaw() but with printf-alike support. This is the funciton that
287 * is used across the code. The raw version is only used in order to dump
288 * the INFO output on crash. */
289 void redisLog(int level, const char *fmt, ...) {
290 va_list ap;
291 char msg[REDIS_MAX_LOGMSG_LEN];
292
293 if ((level&0xff) < server.verbosity) return;
294
295 va_start(ap, fmt);
296 vsnprintf(msg, sizeof(msg), fmt, ap);
297 va_end(ap);
298
299 redisLogRaw(level,msg);
300 }
301
302 /* Log a fixed message without printf-alike capabilities, in a way that is
303 * safe to call from a signal handler.
304 *
305 * We actually use this only for signals that are not fatal from the point
306 * of view of Redis. Signals that are going to kill the server anyway and
307 * where we need printf-alike features are served by redisLog(). */
308 void redisLogFromHandler(int level, const char *msg) {
309 int fd;
310 char buf[64];
311
312 if ((level&0xff) < server.verbosity ||
313 (server.logfile == NULL && server.daemonize)) return;
314 fd = server.logfile ?
315 open(server.logfile, O_APPEND|O_CREAT|O_WRONLY, 0644) :
316 STDOUT_FILENO;
317 if (fd == -1) return;
318 ll2string(buf,sizeof(buf),getpid());
319 if (write(fd,"[",1) == -1) goto err;
320 if (write(fd,buf,strlen(buf)) == -1) goto err;
321 if (write(fd," | signal handler] (",20) == -1) goto err;
322 ll2string(buf,sizeof(buf),time(NULL));
323 if (write(fd,buf,strlen(buf)) == -1) goto err;
324 if (write(fd,") ",2) == -1) goto err;
325 if (write(fd,msg,strlen(msg)) == -1) goto err;
326 if (write(fd,"\n",1) == -1) goto err;
327 err:
328 if (server.logfile) close(fd);
329 }
330
331 /* Redis generally does not try to recover from out of memory conditions
332 * when allocating objects or strings, it is not clear if it will be possible
333 * to report this condition to the client since the networking layer itself
334 * is based on heap allocation for send buffers, so we simply abort.
335 * At least the code will be simpler to read... */
336 void oom(const char *msg) {
337 redisLog(REDIS_WARNING, "%s: Out of memory\n",msg);
338 sleep(1);
339 abort();
340 }
341
342 /* Return the UNIX time in microseconds */
343 long long ustime(void) {
344 struct timeval tv;
345 long long ust;
346
347 gettimeofday(&tv, NULL);
348 ust = ((long long)tv.tv_sec)*1000000;
349 ust += tv.tv_usec;
350 return ust;
351 }
352
353 /* Return the UNIX time in milliseconds */
354 long long mstime(void) {
355 return ustime()/1000;
356 }
357
358 /* After an RDB dump or AOF rewrite we exit from children using _exit() instead of
359 * exit(), because the latter may interact with the same file objects used by
360 * the parent process. However if we are testing the coverage normal exit() is
361 * used in order to obtain the right coverage information. */
362 void exitFromChild(int retcode) {
363 #ifdef COVERAGE_TEST
364 exit(retcode);
365 #else
366 _exit(retcode);
367 #endif
368 }
369
370 /*====================== Hash table type implementation ==================== */
371
372 /* This is an hash table type that uses the SDS dynamic strings libary as
373 * keys and radis objects as values (objects can hold SDS strings,
374 * lists, sets). */
375
376 void dictVanillaFree(void *privdata, void *val)
377 {
378 DICT_NOTUSED(privdata);
379 zfree(val);
380 }
381
382 void dictListDestructor(void *privdata, void *val)
383 {
384 DICT_NOTUSED(privdata);
385 listRelease((list*)val);
386 }
387
388 int dictSdsKeyCompare(void *privdata, const void *key1,
389 const void *key2)
390 {
391 int l1,l2;
392 DICT_NOTUSED(privdata);
393
394 l1 = sdslen((sds)key1);
395 l2 = sdslen((sds)key2);
396 if (l1 != l2) return 0;
397 return memcmp(key1, key2, l1) == 0;
398 }
399
400 /* A case insensitive version used for the command lookup table. */
401 int dictSdsKeyCaseCompare(void *privdata, const void *key1,
402 const void *key2)
403 {
404 DICT_NOTUSED(privdata);
405
406 return strcasecmp(key1, key2) == 0;
407 }
408
409 void dictRedisObjectDestructor(void *privdata, void *val)
410 {
411 DICT_NOTUSED(privdata);
412
413 if (val == NULL) return; /* Values of swapped out keys as set to NULL */
414 decrRefCount(val);
415 }
416
417 void dictSdsDestructor(void *privdata, void *val)
418 {
419 DICT_NOTUSED(privdata);
420
421 sdsfree(val);
422 }
423
424 int dictObjKeyCompare(void *privdata, const void *key1,
425 const void *key2)
426 {
427 const robj *o1 = key1, *o2 = key2;
428 return dictSdsKeyCompare(privdata,o1->ptr,o2->ptr);
429 }
430
431 unsigned int dictObjHash(const void *key) {
432 const robj *o = key;
433 return dictGenHashFunction(o->ptr, sdslen((sds)o->ptr));
434 }
435
436 unsigned int dictSdsHash(const void *key) {
437 return dictGenHashFunction((unsigned char*)key, sdslen((char*)key));
438 }
439
440 unsigned int dictSdsCaseHash(const void *key) {
441 return dictGenCaseHashFunction((unsigned char*)key, sdslen((char*)key));
442 }
443
444 int dictEncObjKeyCompare(void *privdata, const void *key1,
445 const void *key2)
446 {
447 robj *o1 = (robj*) key1, *o2 = (robj*) key2;
448 int cmp;
449
450 if (o1->encoding == REDIS_ENCODING_INT &&
451 o2->encoding == REDIS_ENCODING_INT)
452 return o1->ptr == o2->ptr;
453
454 o1 = getDecodedObject(o1);
455 o2 = getDecodedObject(o2);
456 cmp = dictSdsKeyCompare(privdata,o1->ptr,o2->ptr);
457 decrRefCount(o1);
458 decrRefCount(o2);
459 return cmp;
460 }
461
462 unsigned int dictEncObjHash(const void *key) {
463 robj *o = (robj*) key;
464
465 if (o->encoding == REDIS_ENCODING_RAW) {
466 return dictGenHashFunction(o->ptr, sdslen((sds)o->ptr));
467 } else {
468 if (o->encoding == REDIS_ENCODING_INT) {
469 char buf[32];
470 int len;
471
472 len = ll2string(buf,32,(long)o->ptr);
473 return dictGenHashFunction((unsigned char*)buf, len);
474 } else {
475 unsigned int hash;
476
477 o = getDecodedObject(o);
478 hash = dictGenHashFunction(o->ptr, sdslen((sds)o->ptr));
479 decrRefCount(o);
480 return hash;
481 }
482 }
483 }
484
485 /* Sets type hash table */
486 dictType setDictType = {
487 dictEncObjHash, /* hash function */
488 NULL, /* key dup */
489 NULL, /* val dup */
490 dictEncObjKeyCompare, /* key compare */
491 dictRedisObjectDestructor, /* key destructor */
492 NULL /* val destructor */
493 };
494
495 /* Sorted sets hash (note: a skiplist is used in addition to the hash table) */
496 dictType zsetDictType = {
497 dictEncObjHash, /* hash function */
498 NULL, /* key dup */
499 NULL, /* val dup */
500 dictEncObjKeyCompare, /* key compare */
501 dictRedisObjectDestructor, /* key destructor */
502 NULL /* val destructor */
503 };
504
505 /* Db->dict, keys are sds strings, vals are Redis objects. */
506 dictType dbDictType = {
507 dictSdsHash, /* hash function */
508 NULL, /* key dup */
509 NULL, /* val dup */
510 dictSdsKeyCompare, /* key compare */
511 dictSdsDestructor, /* key destructor */
512 dictRedisObjectDestructor /* val destructor */
513 };
514
515 /* Db->expires */
516 dictType keyptrDictType = {
517 dictSdsHash, /* hash function */
518 NULL, /* key dup */
519 NULL, /* val dup */
520 dictSdsKeyCompare, /* key compare */
521 NULL, /* key destructor */
522 NULL /* val destructor */
523 };
524
525 /* Command table. sds string -> command struct pointer. */
526 dictType commandTableDictType = {
527 dictSdsCaseHash, /* hash function */
528 NULL, /* key dup */
529 NULL, /* val dup */
530 dictSdsKeyCaseCompare, /* key compare */
531 dictSdsDestructor, /* key destructor */
532 NULL /* val destructor */
533 };
534
535 /* Hash type hash table (note that small hashes are represented with zimpaps) */
536 dictType hashDictType = {
537 dictEncObjHash, /* hash function */
538 NULL, /* key dup */
539 NULL, /* val dup */
540 dictEncObjKeyCompare, /* key compare */
541 dictRedisObjectDestructor, /* key destructor */
542 dictRedisObjectDestructor /* val destructor */
543 };
544
545 /* Keylist hash table type has unencoded redis objects as keys and
546 * lists as values. It's used for blocking operations (BLPOP) and to
547 * map swapped keys to a list of clients waiting for this keys to be loaded. */
548 dictType keylistDictType = {
549 dictObjHash, /* hash function */
550 NULL, /* key dup */
551 NULL, /* val dup */
552 dictObjKeyCompare, /* key compare */
553 dictRedisObjectDestructor, /* key destructor */
554 dictListDestructor /* val destructor */
555 };
556
557 int htNeedsResize(dict *dict) {
558 long long size, used;
559
560 size = dictSlots(dict);
561 used = dictSize(dict);
562 return (size && used && size > DICT_HT_INITIAL_SIZE &&
563 (used*100/size < REDIS_HT_MINFILL));
564 }
565
566 /* If the percentage of used slots in the HT reaches REDIS_HT_MINFILL
567 * we resize the hash table to save memory */
568 void tryResizeHashTables(void) {
569 int j;
570
571 for (j = 0; j < server.dbnum; j++) {
572 if (htNeedsResize(server.db[j].dict))
573 dictResize(server.db[j].dict);
574 if (htNeedsResize(server.db[j].expires))
575 dictResize(server.db[j].expires);
576 }
577 }
578
579 /* Our hash table implementation performs rehashing incrementally while
580 * we write/read from the hash table. Still if the server is idle, the hash
581 * table will use two tables for a long time. So we try to use 1 millisecond
582 * of CPU time at every serverCron() loop in order to rehash some key. */
583 void incrementallyRehash(void) {
584 int j;
585
586 for (j = 0; j < server.dbnum; j++) {
587 if (dictIsRehashing(server.db[j].dict)) {
588 dictRehashMilliseconds(server.db[j].dict,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;
615
616 for (j = 0; j < server.dbnum; j++) {
617 int expired;
618 redisDb *db = server.db+j;
619
620 /* Continue to expire if at the end of the cycle more than 25%
621 * of the keys were expired. */
622 do {
623 long num = dictSize(db->expires);
624 long long now = mstime();
625
626 expired = 0;
627 if (num > REDIS_EXPIRELOOKUPS_PER_CRON)
628 num = REDIS_EXPIRELOOKUPS_PER_CRON;
629 while (num--) {
630 dictEntry *de;
631 long long t;
632
633 if ((de = dictGetRandomKey(db->expires)) == NULL) break;
634 t = dictGetSignedIntegerVal(de);
635 if (now > t) {
636 sds key = dictGetKey(de);
637 robj *keyobj = createStringObject(key,sdslen(key));
638
639 propagateExpire(db,keyobj);
640 dbDelete(db,keyobj);
641 decrRefCount(keyobj);
642 expired++;
643 server.stat_expiredkeys++;
644 }
645 }
646 } while (expired > REDIS_EXPIRELOOKUPS_PER_CRON/4);
647 }
648 }
649
650 void updateLRUClock(void) {
651 server.lruclock = (server.unixtime/REDIS_LRU_CLOCK_RESOLUTION) &
652 REDIS_LRU_CLOCK_MAX;
653 }
654
655
656 /* Add a sample to the operations per second array of samples. */
657 void trackOperationsPerSecond(void) {
658 long long t = mstime() - server.ops_sec_last_sample_time;
659 long long ops = server.stat_numcommands - server.ops_sec_last_sample_ops;
660 long long ops_sec;
661
662 ops_sec = t > 0 ? (ops*1000/t) : 0;
663
664 server.ops_sec_samples[server.ops_sec_idx] = ops_sec;
665 server.ops_sec_idx = (server.ops_sec_idx+1) % REDIS_OPS_SEC_SAMPLES;
666 server.ops_sec_last_sample_time = mstime();
667 server.ops_sec_last_sample_ops = server.stat_numcommands;
668 }
669
670 /* Return the mean of all the samples. */
671 long long getOperationsPerSecond(void) {
672 int j;
673 long long sum = 0;
674
675 for (j = 0; j < REDIS_OPS_SEC_SAMPLES; j++)
676 sum += server.ops_sec_samples[j];
677 return sum / REDIS_OPS_SEC_SAMPLES;
678 }
679
680 /* Check for timeouts. Returns non-zero if the client was terminated */
681 int clientsCronHandleTimeout(redisClient *c) {
682 time_t now = server.unixtime;
683
684 if (server.maxidletime &&
685 !(c->flags & REDIS_SLAVE) && /* no timeout for slaves */
686 !(c->flags & REDIS_MASTER) && /* no timeout for masters */
687 !(c->flags & REDIS_BLOCKED) && /* no timeout for BLPOP */
688 dictSize(c->pubsub_channels) == 0 && /* no timeout for pubsub */
689 listLength(c->pubsub_patterns) == 0 &&
690 (now - c->lastinteraction > server.maxidletime))
691 {
692 redisLog(REDIS_VERBOSE,"Closing idle client");
693 freeClient(c);
694 return 1;
695 } else if (c->flags & REDIS_BLOCKED) {
696 if (c->bpop.timeout != 0 && c->bpop.timeout < now) {
697 addReply(c,shared.nullmultibulk);
698 unblockClientWaitingData(c);
699 }
700 }
701 return 0;
702 }
703
704 /* The client query buffer is an sds.c string that can end with a lot of
705 * free space not used, this function reclaims space if needed.
706 *
707 * The funciton always returns 0 as it never terminates the client. */
708 int clientsCronResizeQueryBuffer(redisClient *c) {
709 size_t querybuf_size = sdsAllocSize(c->querybuf);
710 time_t idletime = server.unixtime - c->lastinteraction;
711
712 /* There are two conditions to resize the query buffer:
713 * 1) Query buffer is > BIG_ARG and too big for latest peak.
714 * 2) Client is inactive and the buffer is bigger than 1k. */
715 if (((querybuf_size > REDIS_MBULK_BIG_ARG) &&
716 (querybuf_size/(c->querybuf_peak+1)) > 2) ||
717 (querybuf_size > 1024 && idletime > 2))
718 {
719 /* Only resize the query buffer if it is actually wasting space. */
720 if (sdsavail(c->querybuf) > 1024) {
721 c->querybuf = sdsRemoveFreeSpace(c->querybuf);
722 }
723 }
724 /* Reset the peak again to capture the peak memory usage in the next
725 * cycle. */
726 c->querybuf_peak = 0;
727 return 0;
728 }
729
730 void clientsCron(void) {
731 /* Make sure to process at least 1/100 of clients per call.
732 * Since this function is called 10 times per second we are sure that
733 * in the worst case we process all the clients in 10 seconds.
734 * In normal conditions (a reasonable number of clients) we process
735 * all the clients in a shorter time. */
736 int numclients = listLength(server.clients);
737 int iterations = numclients/100;
738
739 if (iterations < 50)
740 iterations = (numclients < 50) ? numclients : 50;
741 while(listLength(server.clients) && iterations--) {
742 redisClient *c;
743 listNode *head;
744
745 /* Rotate the list, take the current head, process.
746 * This way if the client must be removed from the list it's the
747 * first element and we don't incur into O(N) computation. */
748 listRotate(server.clients);
749 head = listFirst(server.clients);
750 c = listNodeValue(head);
751 /* The following functions do different service checks on the client.
752 * The protocol is that they return non-zero if the client was
753 * terminated. */
754 if (clientsCronHandleTimeout(c)) continue;
755 if (clientsCronResizeQueryBuffer(c)) continue;
756 }
757 }
758
759 int serverCron(struct aeEventLoop *eventLoop, long long id, void *clientData) {
760 int j, loops = server.cronloops;
761 REDIS_NOTUSED(eventLoop);
762 REDIS_NOTUSED(id);
763 REDIS_NOTUSED(clientData);
764
765 /* Software watchdog: deliver the SIGALRM that will reach the signal
766 * handler if we don't return here fast enough. */
767 if (server.watchdog_period) watchdogScheduleSignal(server.watchdog_period);
768
769 /* We take a cached value of the unix time in the global state because
770 * with virtual memory and aging there is to store the current time
771 * in objects at every object access, and accuracy is not needed.
772 * To access a global var is faster than calling time(NULL) */
773 server.unixtime = time(NULL);
774
775 trackOperationsPerSecond();
776
777 /* We have just 22 bits per object for LRU information.
778 * So we use an (eventually wrapping) LRU clock with 10 seconds resolution.
779 * 2^22 bits with 10 seconds resoluton is more or less 1.5 years.
780 *
781 * Note that even if this will wrap after 1.5 years it's not a problem,
782 * everything will still work but just some object will appear younger
783 * to Redis. But for this to happen a given object should never be touched
784 * for 1.5 years.
785 *
786 * Note that you can change the resolution altering the
787 * REDIS_LRU_CLOCK_RESOLUTION define.
788 */
789 updateLRUClock();
790
791 /* Record the max memory used since the server was started. */
792 if (zmalloc_used_memory() > server.stat_peak_memory)
793 server.stat_peak_memory = zmalloc_used_memory();
794
795 /* We received a SIGTERM, shutting down here in a safe way, as it is
796 * not ok doing so inside the signal handler. */
797 if (server.shutdown_asap) {
798 if (prepareForShutdown(0) == REDIS_OK) exit(0);
799 redisLog(REDIS_WARNING,"SIGTERM received but errors trying to shut down the server, check the logs for more information");
800 }
801
802 /* Show some info about non-empty databases */
803 for (j = 0; j < server.dbnum; j++) {
804 long long size, used, vkeys;
805
806 size = dictSlots(server.db[j].dict);
807 used = dictSize(server.db[j].dict);
808 vkeys = dictSize(server.db[j].expires);
809 if (!(loops % 50) && (used || vkeys)) {
810 redisLog(REDIS_VERBOSE,"DB %d: %lld keys (%lld volatile) in %lld slots HT.",j,used,vkeys,size);
811 /* dictPrintStats(server.dict); */
812 }
813 }
814
815 /* We don't want to resize the hash tables while a bacground saving
816 * is in progress: the saving child is created using fork() that is
817 * implemented with a copy-on-write semantic in most modern systems, so
818 * if we resize the HT while there is the saving child at work actually
819 * a lot of memory movements in the parent will cause a lot of pages
820 * copied. */
821 if (server.rdb_child_pid == -1 && server.aof_child_pid == -1) {
822 if (!(loops % 10)) tryResizeHashTables();
823 if (server.activerehashing) incrementallyRehash();
824 }
825
826 /* Show information about connected clients */
827 if (!(loops % 50)) {
828 redisLog(REDIS_VERBOSE,"%d clients connected (%d slaves), %zu bytes in use",
829 listLength(server.clients)-listLength(server.slaves),
830 listLength(server.slaves),
831 zmalloc_used_memory());
832 }
833
834 /* We need to do a few operations on clients asynchronously. */
835 clientsCron();
836
837 /* Start a scheduled AOF rewrite if this was requested by the user while
838 * a BGSAVE was in progress. */
839 if (server.rdb_child_pid == -1 && server.aof_child_pid == -1 &&
840 server.aof_rewrite_scheduled)
841 {
842 rewriteAppendOnlyFileBackground();
843 }
844
845 /* Check if a background saving or AOF rewrite in progress terminated. */
846 if (server.rdb_child_pid != -1 || server.aof_child_pid != -1) {
847 int statloc;
848 pid_t pid;
849
850 if ((pid = wait3(&statloc,WNOHANG,NULL)) != 0) {
851 int exitcode = WEXITSTATUS(statloc);
852 int bysignal = 0;
853
854 if (WIFSIGNALED(statloc)) bysignal = WTERMSIG(statloc);
855
856 if (pid == server.rdb_child_pid) {
857 backgroundSaveDoneHandler(exitcode,bysignal);
858 } else {
859 backgroundRewriteDoneHandler(exitcode,bysignal);
860 }
861 updateDictResizePolicy();
862 }
863 } else {
864 /* If there is not a background saving/rewrite in progress check if
865 * we have to save/rewrite now */
866 for (j = 0; j < server.saveparamslen; j++) {
867 struct saveparam *sp = server.saveparams+j;
868
869 if (server.dirty >= sp->changes &&
870 server.unixtime-server.lastsave > sp->seconds) {
871 redisLog(REDIS_NOTICE,"%d changes in %d seconds. Saving...",
872 sp->changes, sp->seconds);
873 rdbSaveBackground(server.rdb_filename);
874 break;
875 }
876 }
877
878 /* Trigger an AOF rewrite if needed */
879 if (server.rdb_child_pid == -1 &&
880 server.aof_child_pid == -1 &&
881 server.aof_rewrite_perc &&
882 server.aof_current_size > server.aof_rewrite_min_size)
883 {
884 long long base = server.aof_rewrite_base_size ?
885 server.aof_rewrite_base_size : 1;
886 long long growth = (server.aof_current_size*100/base) - 100;
887 if (growth >= server.aof_rewrite_perc) {
888 redisLog(REDIS_NOTICE,"Starting automatic rewriting of AOF on %lld%% growth",growth);
889 rewriteAppendOnlyFileBackground();
890 }
891 }
892 }
893
894
895 /* If we postponed an AOF buffer flush, let's try to do it every time the
896 * cron function is called. */
897 if (server.aof_flush_postponed_start) flushAppendOnlyFile(0);
898
899 /* Expire a few keys per cycle, only if this is a master.
900 * On slaves we wait for DEL operations synthesized by the master
901 * in order to guarantee a strict consistency. */
902 if (server.masterhost == NULL) activeExpireCycle();
903
904 /* Close clients that need to be closed asynchronous */
905 freeClientsInAsyncFreeQueue();
906
907 /* Replication cron function -- used to reconnect to master and
908 * to detect transfer failures. */
909 if (!(loops % 10)) replicationCron();
910
911 server.cronloops++;
912 return 100;
913 }
914
915 /* This function gets called every time Redis is entering the
916 * main loop of the event driven library, that is, before to sleep
917 * for ready file descriptors. */
918 void beforeSleep(struct aeEventLoop *eventLoop) {
919 REDIS_NOTUSED(eventLoop);
920 listNode *ln;
921 redisClient *c;
922
923 /* Try to process pending commands for clients that were just unblocked. */
924 while (listLength(server.unblocked_clients)) {
925 ln = listFirst(server.unblocked_clients);
926 redisAssert(ln != NULL);
927 c = ln->value;
928 listDelNode(server.unblocked_clients,ln);
929 c->flags &= ~REDIS_UNBLOCKED;
930
931 /* Process remaining data in the input buffer. */
932 if (c->querybuf && sdslen(c->querybuf) > 0) {
933 server.current_client = c;
934 processInputBuffer(c);
935 server.current_client = NULL;
936 }
937 }
938
939 /* Write the AOF buffer on disk */
940 flushAppendOnlyFile(0);
941 }
942
943 /* =========================== Server initialization ======================== */
944
945 void createSharedObjects(void) {
946 int j;
947
948 shared.crlf = createObject(REDIS_STRING,sdsnew("\r\n"));
949 shared.ok = createObject(REDIS_STRING,sdsnew("+OK\r\n"));
950 shared.err = createObject(REDIS_STRING,sdsnew("-ERR\r\n"));
951 shared.emptybulk = createObject(REDIS_STRING,sdsnew("$0\r\n\r\n"));
952 shared.czero = createObject(REDIS_STRING,sdsnew(":0\r\n"));
953 shared.cone = createObject(REDIS_STRING,sdsnew(":1\r\n"));
954 shared.cnegone = createObject(REDIS_STRING,sdsnew(":-1\r\n"));
955 shared.nullbulk = createObject(REDIS_STRING,sdsnew("$-1\r\n"));
956 shared.nullmultibulk = createObject(REDIS_STRING,sdsnew("*-1\r\n"));
957 shared.emptymultibulk = createObject(REDIS_STRING,sdsnew("*0\r\n"));
958 shared.pong = createObject(REDIS_STRING,sdsnew("+PONG\r\n"));
959 shared.queued = createObject(REDIS_STRING,sdsnew("+QUEUED\r\n"));
960 shared.wrongtypeerr = createObject(REDIS_STRING,sdsnew(
961 "-ERR Operation against a key holding the wrong kind of value\r\n"));
962 shared.nokeyerr = createObject(REDIS_STRING,sdsnew(
963 "-ERR no such key\r\n"));
964 shared.syntaxerr = createObject(REDIS_STRING,sdsnew(
965 "-ERR syntax error\r\n"));
966 shared.sameobjecterr = createObject(REDIS_STRING,sdsnew(
967 "-ERR source and destination objects are the same\r\n"));
968 shared.outofrangeerr = createObject(REDIS_STRING,sdsnew(
969 "-ERR index out of range\r\n"));
970 shared.noscripterr = createObject(REDIS_STRING,sdsnew(
971 "-NOSCRIPT No matching script. Please use EVAL.\r\n"));
972 shared.loadingerr = createObject(REDIS_STRING,sdsnew(
973 "-LOADING Redis is loading the dataset in memory\r\n"));
974 shared.slowscripterr = createObject(REDIS_STRING,sdsnew(
975 "-BUSY Redis is busy running a script. You can only call SCRIPT KILL or SHUTDOWN NOSAVE.\r\n"));
976 shared.bgsaveerr = createObject(REDIS_STRING,sdsnew(
977 "-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"));
978 shared.roslaveerr = createObject(REDIS_STRING,sdsnew(
979 "-READONLY You can't write against a read only slave.\r\n"));
980 shared.oomerr = createObject(REDIS_STRING,sdsnew(
981 "-OOM command not allowed when used memory > 'maxmemory'.\r\n"));
982 shared.space = createObject(REDIS_STRING,sdsnew(" "));
983 shared.colon = createObject(REDIS_STRING,sdsnew(":"));
984 shared.plus = createObject(REDIS_STRING,sdsnew("+"));
985
986 for (j = 0; j < REDIS_SHARED_SELECT_CMDS; j++) {
987 shared.select[j] = createObject(REDIS_STRING,
988 sdscatprintf(sdsempty(),"select %d\r\n", j));
989 }
990 shared.messagebulk = createStringObject("$7\r\nmessage\r\n",13);
991 shared.pmessagebulk = createStringObject("$8\r\npmessage\r\n",14);
992 shared.subscribebulk = createStringObject("$9\r\nsubscribe\r\n",15);
993 shared.unsubscribebulk = createStringObject("$11\r\nunsubscribe\r\n",18);
994 shared.psubscribebulk = createStringObject("$10\r\npsubscribe\r\n",17);
995 shared.punsubscribebulk = createStringObject("$12\r\npunsubscribe\r\n",19);
996 shared.del = createStringObject("DEL",3);
997 shared.rpop = createStringObject("RPOP",4);
998 shared.lpop = createStringObject("LPOP",4);
999 for (j = 0; j < REDIS_SHARED_INTEGERS; j++) {
1000 shared.integers[j] = createObject(REDIS_STRING,(void*)(long)j);
1001 shared.integers[j]->encoding = REDIS_ENCODING_INT;
1002 }
1003 for (j = 0; j < REDIS_SHARED_BULKHDR_LEN; j++) {
1004 shared.mbulkhdr[j] = createObject(REDIS_STRING,
1005 sdscatprintf(sdsempty(),"*%d\r\n",j));
1006 shared.bulkhdr[j] = createObject(REDIS_STRING,
1007 sdscatprintf(sdsempty(),"$%d\r\n",j));
1008 }
1009 }
1010
1011 void initServerConfig() {
1012 getRandomHexChars(server.runid,REDIS_RUN_ID_SIZE);
1013 server.runid[REDIS_RUN_ID_SIZE] = '\0';
1014 server.arch_bits = (sizeof(long) == 8) ? 64 : 32;
1015 server.port = REDIS_SERVERPORT;
1016 server.bindaddr = NULL;
1017 server.unixsocket = NULL;
1018 server.unixsocketperm = 0;
1019 server.ipfd = -1;
1020 server.sofd = -1;
1021 server.dbnum = REDIS_DEFAULT_DBNUM;
1022 server.verbosity = REDIS_NOTICE;
1023 server.maxidletime = REDIS_MAXIDLETIME;
1024 server.client_max_querybuf_len = REDIS_MAX_QUERYBUF_LEN;
1025 server.saveparams = NULL;
1026 server.loading = 0;
1027 server.logfile = NULL; /* NULL = log on standard output */
1028 server.syslog_enabled = 0;
1029 server.syslog_ident = zstrdup("redis");
1030 server.syslog_facility = LOG_LOCAL0;
1031 server.daemonize = 0;
1032 server.aof_state = REDIS_AOF_OFF;
1033 server.aof_fsync = AOF_FSYNC_EVERYSEC;
1034 server.aof_no_fsync_on_rewrite = 0;
1035 server.aof_rewrite_perc = REDIS_AOF_REWRITE_PERC;
1036 server.aof_rewrite_min_size = REDIS_AOF_REWRITE_MIN_SIZE;
1037 server.aof_rewrite_base_size = 0;
1038 server.aof_rewrite_scheduled = 0;
1039 server.aof_last_fsync = time(NULL);
1040 server.aof_delayed_fsync = 0;
1041 server.aof_fd = -1;
1042 server.aof_selected_db = -1; /* Make sure the first time will not match */
1043 server.aof_flush_postponed_start = 0;
1044 server.pidfile = zstrdup("/var/run/redis.pid");
1045 server.rdb_filename = zstrdup("dump.rdb");
1046 server.aof_filename = zstrdup("appendonly.aof");
1047 server.requirepass = NULL;
1048 server.rdb_compression = 1;
1049 server.rdb_checksum = 1;
1050 server.activerehashing = 1;
1051 server.maxclients = REDIS_MAX_CLIENTS;
1052 server.bpop_blocked_clients = 0;
1053 server.maxmemory = 0;
1054 server.maxmemory_policy = REDIS_MAXMEMORY_VOLATILE_LRU;
1055 server.maxmemory_samples = 3;
1056 server.hash_max_ziplist_entries = REDIS_HASH_MAX_ZIPLIST_ENTRIES;
1057 server.hash_max_ziplist_value = REDIS_HASH_MAX_ZIPLIST_VALUE;
1058 server.list_max_ziplist_entries = REDIS_LIST_MAX_ZIPLIST_ENTRIES;
1059 server.list_max_ziplist_value = REDIS_LIST_MAX_ZIPLIST_VALUE;
1060 server.set_max_intset_entries = REDIS_SET_MAX_INTSET_ENTRIES;
1061 server.zset_max_ziplist_entries = REDIS_ZSET_MAX_ZIPLIST_ENTRIES;
1062 server.zset_max_ziplist_value = REDIS_ZSET_MAX_ZIPLIST_VALUE;
1063 server.shutdown_asap = 0;
1064 server.repl_ping_slave_period = REDIS_REPL_PING_SLAVE_PERIOD;
1065 server.repl_timeout = REDIS_REPL_TIMEOUT;
1066 server.lua_caller = NULL;
1067 server.lua_time_limit = REDIS_LUA_TIME_LIMIT;
1068 server.lua_client = NULL;
1069 server.lua_timedout = 0;
1070
1071 updateLRUClock();
1072 resetServerSaveParams();
1073
1074 appendServerSaveParams(60*60,1); /* save after 1 hour and 1 change */
1075 appendServerSaveParams(300,100); /* save after 5 minutes and 100 changes */
1076 appendServerSaveParams(60,10000); /* save after 1 minute and 10000 changes */
1077 /* Replication related */
1078 server.masterauth = NULL;
1079 server.masterhost = NULL;
1080 server.masterport = 6379;
1081 server.master = NULL;
1082 server.repl_state = REDIS_REPL_NONE;
1083 server.repl_syncio_timeout = REDIS_REPL_SYNCIO_TIMEOUT;
1084 server.repl_serve_stale_data = 1;
1085 server.repl_slave_ro = 1;
1086 server.repl_down_since = time(NULL);
1087
1088 /* Client output buffer limits */
1089 server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_NORMAL].hard_limit_bytes = 0;
1090 server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_NORMAL].soft_limit_bytes = 0;
1091 server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_NORMAL].soft_limit_seconds = 0;
1092 server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_SLAVE].hard_limit_bytes = 1024*1024*256;
1093 server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_SLAVE].soft_limit_bytes = 1024*1024*64;
1094 server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_SLAVE].soft_limit_seconds = 60;
1095 server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_PUBSUB].hard_limit_bytes = 1024*1024*32;
1096 server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_PUBSUB].soft_limit_bytes = 1024*1024*8;
1097 server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_PUBSUB].soft_limit_seconds = 60;
1098
1099 /* Double constants initialization */
1100 R_Zero = 0.0;
1101 R_PosInf = 1.0/R_Zero;
1102 R_NegInf = -1.0/R_Zero;
1103 R_Nan = R_Zero/R_Zero;
1104
1105 /* Command table -- we intiialize it here as it is part of the
1106 * initial configuration, since command names may be changed via
1107 * redis.conf using the rename-command directive. */
1108 server.commands = dictCreate(&commandTableDictType,NULL);
1109 populateCommandTable();
1110 server.delCommand = lookupCommandByCString("del");
1111 server.multiCommand = lookupCommandByCString("multi");
1112 server.lpushCommand = lookupCommandByCString("lpush");
1113
1114 /* Slow log */
1115 server.slowlog_log_slower_than = REDIS_SLOWLOG_LOG_SLOWER_THAN;
1116 server.slowlog_max_len = REDIS_SLOWLOG_MAX_LEN;
1117
1118 /* Debugging */
1119 server.assert_failed = "<no assertion failed>";
1120 server.assert_file = "<no file>";
1121 server.assert_line = 0;
1122 server.bug_report_start = 0;
1123 server.watchdog_period = 0;
1124 }
1125
1126 /* This function will try to raise the max number of open files accordingly to
1127 * the configured max number of clients. It will also account for 32 additional
1128 * file descriptors as we need a few more for persistence, listening
1129 * sockets, log files and so forth.
1130 *
1131 * If it will not be possible to set the limit accordingly to the configured
1132 * max number of clients, the function will do the reverse setting
1133 * server.maxclients to the value that we can actually handle. */
1134 void adjustOpenFilesLimit(void) {
1135 rlim_t maxfiles = server.maxclients+32;
1136 struct rlimit limit;
1137
1138 if (getrlimit(RLIMIT_NOFILE,&limit) == -1) {
1139 redisLog(REDIS_WARNING,"Unable to obtain the current NOFILE limit (%s), assuming 1024 and setting the max clients configuration accordingly.",
1140 strerror(errno));
1141 server.maxclients = 1024-32;
1142 } else {
1143 rlim_t oldlimit = limit.rlim_cur;
1144
1145 /* Set the max number of files if the current limit is not enough
1146 * for our needs. */
1147 if (oldlimit < maxfiles) {
1148 rlim_t f;
1149
1150 f = maxfiles;
1151 while(f > oldlimit) {
1152 limit.rlim_cur = f;
1153 limit.rlim_max = f;
1154 if (setrlimit(RLIMIT_NOFILE,&limit) != -1) break;
1155 f -= 128;
1156 }
1157 if (f < oldlimit) f = oldlimit;
1158 if (f != maxfiles) {
1159 server.maxclients = f-32;
1160 redisLog(REDIS_WARNING,"Unable to set the max number of files limit to %d (%s), setting the max clients configuration to %d.",
1161 (int) maxfiles, strerror(errno), (int) server.maxclients);
1162 } else {
1163 redisLog(REDIS_NOTICE,"Max number of open files set to %d",
1164 (int) maxfiles);
1165 }
1166 }
1167 }
1168 }
1169
1170 void initServer() {
1171 int j;
1172
1173 signal(SIGHUP, SIG_IGN);
1174 signal(SIGPIPE, SIG_IGN);
1175 setupSignalHandlers();
1176
1177 if (server.syslog_enabled) {
1178 openlog(server.syslog_ident, LOG_PID | LOG_NDELAY | LOG_NOWAIT,
1179 server.syslog_facility);
1180 }
1181
1182 server.current_client = NULL;
1183 server.clients = listCreate();
1184 server.clients_to_close = listCreate();
1185 server.slaves = listCreate();
1186 server.monitors = listCreate();
1187 server.unblocked_clients = listCreate();
1188
1189 createSharedObjects();
1190 adjustOpenFilesLimit();
1191 server.el = aeCreateEventLoop(server.maxclients+1024);
1192 server.db = zmalloc(sizeof(redisDb)*server.dbnum);
1193
1194 if (server.port != 0) {
1195 server.ipfd = anetTcpServer(server.neterr,server.port,server.bindaddr);
1196 if (server.ipfd == ANET_ERR) {
1197 redisLog(REDIS_WARNING, "Opening port %d: %s",
1198 server.port, server.neterr);
1199 exit(1);
1200 }
1201 }
1202 if (server.unixsocket != NULL) {
1203 unlink(server.unixsocket); /* don't care if this fails */
1204 server.sofd = anetUnixServer(server.neterr,server.unixsocket,server.unixsocketperm);
1205 if (server.sofd == ANET_ERR) {
1206 redisLog(REDIS_WARNING, "Opening socket: %s", server.neterr);
1207 exit(1);
1208 }
1209 }
1210 if (server.ipfd < 0 && server.sofd < 0) {
1211 redisLog(REDIS_WARNING, "Configured to not listen anywhere, exiting.");
1212 exit(1);
1213 }
1214 for (j = 0; j < server.dbnum; j++) {
1215 server.db[j].dict = dictCreate(&dbDictType,NULL);
1216 server.db[j].expires = dictCreate(&keyptrDictType,NULL);
1217 server.db[j].blocking_keys = dictCreate(&keylistDictType,NULL);
1218 server.db[j].watched_keys = dictCreate(&keylistDictType,NULL);
1219 server.db[j].id = j;
1220 }
1221 server.pubsub_channels = dictCreate(&keylistDictType,NULL);
1222 server.pubsub_patterns = listCreate();
1223 listSetFreeMethod(server.pubsub_patterns,freePubsubPattern);
1224 listSetMatchMethod(server.pubsub_patterns,listMatchPubsubPattern);
1225 server.cronloops = 0;
1226 server.rdb_child_pid = -1;
1227 server.aof_child_pid = -1;
1228 server.aof_rewrite_buf = sdsempty();
1229 server.aof_buf = sdsempty();
1230 server.lastsave = time(NULL);
1231 server.dirty = 0;
1232 server.stat_numcommands = 0;
1233 server.stat_numconnections = 0;
1234 server.stat_expiredkeys = 0;
1235 server.stat_evictedkeys = 0;
1236 server.stat_starttime = time(NULL);
1237 server.stat_keyspace_misses = 0;
1238 server.stat_keyspace_hits = 0;
1239 server.stat_peak_memory = 0;
1240 server.stat_fork_time = 0;
1241 server.stat_rejected_conn = 0;
1242 memset(server.ops_sec_samples,0,sizeof(server.ops_sec_samples));
1243 server.ops_sec_idx = 0;
1244 server.ops_sec_last_sample_time = mstime();
1245 server.ops_sec_last_sample_ops = 0;
1246 server.unixtime = time(NULL);
1247 server.lastbgsave_status = REDIS_OK;
1248 server.stop_writes_on_bgsave_err = 1;
1249 aeCreateTimeEvent(server.el, 1, serverCron, NULL, NULL);
1250 if (server.ipfd > 0 && aeCreateFileEvent(server.el,server.ipfd,AE_READABLE,
1251 acceptTcpHandler,NULL) == AE_ERR) oom("creating file event");
1252 if (server.sofd > 0 && aeCreateFileEvent(server.el,server.sofd,AE_READABLE,
1253 acceptUnixHandler,NULL) == AE_ERR) oom("creating file event");
1254
1255 if (server.aof_state == REDIS_AOF_ON) {
1256 server.aof_fd = open(server.aof_filename,
1257 O_WRONLY|O_APPEND|O_CREAT,0644);
1258 if (server.aof_fd == -1) {
1259 redisLog(REDIS_WARNING, "Can't open the append-only file: %s",
1260 strerror(errno));
1261 exit(1);
1262 }
1263 }
1264
1265 /* 32 bit instances are limited to 4GB of address space, so if there is
1266 * no explicit limit in the user provided configuration we set a limit
1267 * at 3.5GB using maxmemory with 'noeviction' policy'. This saves
1268 * useless crashes of the Redis instance. */
1269 if (server.arch_bits == 32 && server.maxmemory == 0) {
1270 redisLog(REDIS_WARNING,"Warning: 32 bit instance detected but no memory limit set. Setting 3.5 GB maxmemory limit with 'noeviction' policy now.");
1271 server.maxmemory = 3584LL*(1024*1024); /* 3584 MB = 3.5 GB */
1272 server.maxmemory_policy = REDIS_MAXMEMORY_NO_EVICTION;
1273 }
1274
1275 scriptingInit();
1276 slowlogInit();
1277 bioInit();
1278 }
1279
1280 /* Populates the Redis Command Table starting from the hard coded list
1281 * we have on top of redis.c file. */
1282 void populateCommandTable(void) {
1283 int j;
1284 int numcommands = sizeof(redisCommandTable)/sizeof(struct redisCommand);
1285
1286 for (j = 0; j < numcommands; j++) {
1287 struct redisCommand *c = redisCommandTable+j;
1288 char *f = c->sflags;
1289 int retval;
1290
1291 while(*f != '\0') {
1292 switch(*f) {
1293 case 'w': c->flags |= REDIS_CMD_WRITE; break;
1294 case 'r': c->flags |= REDIS_CMD_READONLY; break;
1295 case 'm': c->flags |= REDIS_CMD_DENYOOM; break;
1296 case 'a': c->flags |= REDIS_CMD_ADMIN; break;
1297 case 'p': c->flags |= REDIS_CMD_PUBSUB; break;
1298 case 'f': c->flags |= REDIS_CMD_FORCE_REPLICATION; break;
1299 case 's': c->flags |= REDIS_CMD_NOSCRIPT; break;
1300 case 'R': c->flags |= REDIS_CMD_RANDOM; break;
1301 case 'S': c->flags |= REDIS_CMD_SORT_FOR_SCRIPT; break;
1302 default: redisPanic("Unsupported command flag"); break;
1303 }
1304 f++;
1305 }
1306
1307 retval = dictAdd(server.commands, sdsnew(c->name), c);
1308 assert(retval == DICT_OK);
1309 }
1310 }
1311
1312 void resetCommandTableStats(void) {
1313 int numcommands = sizeof(redisCommandTable)/sizeof(struct redisCommand);
1314 int j;
1315
1316 for (j = 0; j < numcommands; j++) {
1317 struct redisCommand *c = redisCommandTable+j;
1318
1319 c->microseconds = 0;
1320 c->calls = 0;
1321 }
1322 }
1323
1324 /* ========================== Redis OP Array API ============================ */
1325
1326 void redisOpArrayInit(redisOpArray *oa) {
1327 oa->ops = NULL;
1328 oa->numops = 0;
1329 }
1330
1331 int redisOpArrayAppend(redisOpArray *oa, struct redisCommand *cmd, int dbid,
1332 robj **argv, int argc, int target)
1333 {
1334 redisOp *op;
1335
1336 oa->ops = zrealloc(oa->ops,sizeof(redisOp)*(oa->numops+1));
1337 op = oa->ops+oa->numops;
1338 op->cmd = cmd;
1339 op->dbid = dbid;
1340 op->argv = argv;
1341 op->argc = argc;
1342 op->target = target;
1343 oa->numops++;
1344 return oa->numops;
1345 }
1346
1347 void redisOpArrayFree(redisOpArray *oa) {
1348 while(oa->numops) {
1349 int j;
1350 redisOp *op;
1351
1352 oa->numops--;
1353 op = oa->ops+oa->numops;
1354 for (j = 0; j < op->argc; j++)
1355 decrRefCount(op->argv[j]);
1356 zfree(op->argv);
1357 }
1358 zfree(oa->ops);
1359 }
1360
1361 /* ====================== Commands lookup and execution ===================== */
1362
1363 struct redisCommand *lookupCommand(sds name) {
1364 return dictFetchValue(server.commands, name);
1365 }
1366
1367 struct redisCommand *lookupCommandByCString(char *s) {
1368 struct redisCommand *cmd;
1369 sds name = sdsnew(s);
1370
1371 cmd = dictFetchValue(server.commands, name);
1372 sdsfree(name);
1373 return cmd;
1374 }
1375
1376 /* Propagate the specified command (in the context of the specified database id)
1377 * to AOF, Slaves and Monitors.
1378 *
1379 * flags are an xor between:
1380 * + REDIS_PROPAGATE_NONE (no propagation of command at all)
1381 * + REDIS_PROPAGATE_AOF (propagate into the AOF file if is enabled)
1382 * + REDIS_PROPAGATE_REPL (propagate into the replication link)
1383 */
1384 void propagate(struct redisCommand *cmd, int dbid, robj **argv, int argc,
1385 int flags)
1386 {
1387 if (server.aof_state != REDIS_AOF_OFF && flags & REDIS_PROPAGATE_AOF)
1388 feedAppendOnlyFile(cmd,dbid,argv,argc);
1389 if (flags & REDIS_PROPAGATE_REPL && listLength(server.slaves))
1390 replicationFeedSlaves(server.slaves,dbid,argv,argc);
1391 }
1392
1393 /* Used inside commands to schedule the propagation of additional commands
1394 * after the current command is propagated to AOF / Replication. */
1395 void alsoPropagate(struct redisCommand *cmd, int dbid, robj **argv, int argc,
1396 int target)
1397 {
1398 redisOpArrayAppend(&server.also_propagate,cmd,dbid,argv,argc,target);
1399 }
1400
1401 /* Call() is the core of Redis execution of a command */
1402 void call(redisClient *c, int flags) {
1403 long long dirty, start = ustime(), duration;
1404
1405 /* Sent the command to clients in MONITOR mode, only if the commands are
1406 * not geneated from reading an AOF. */
1407 if (listLength(server.monitors) && !server.loading)
1408 replicationFeedMonitors(c,server.monitors,c->db->id,c->argv,c->argc);
1409
1410 /* Call the command. */
1411 redisOpArrayInit(&server.also_propagate);
1412 dirty = server.dirty;
1413 c->cmd->proc(c);
1414 dirty = server.dirty-dirty;
1415 duration = ustime()-start;
1416
1417 /* When EVAL is called loading the AOF we don't want commands called
1418 * from Lua to go into the slowlog or to populate statistics. */
1419 if (server.loading && c->flags & REDIS_LUA_CLIENT)
1420 flags &= ~(REDIS_CALL_SLOWLOG | REDIS_CALL_STATS);
1421
1422 /* Log the command into the Slow log if needed, and populate the
1423 * per-command statistics that we show in INFO commandstats. */
1424 if (flags & REDIS_CALL_SLOWLOG)
1425 slowlogPushEntryIfNeeded(c->argv,c->argc,duration);
1426 if (flags & REDIS_CALL_STATS) {
1427 c->cmd->microseconds += duration;
1428 c->cmd->calls++;
1429 }
1430
1431 /* Propagate the command into the AOF and replication link */
1432 if (flags & REDIS_CALL_PROPAGATE) {
1433 int flags = REDIS_PROPAGATE_NONE;
1434
1435 if (c->cmd->flags & REDIS_CMD_FORCE_REPLICATION)
1436 flags |= REDIS_PROPAGATE_REPL;
1437 if (dirty)
1438 flags |= (REDIS_PROPAGATE_REPL | REDIS_PROPAGATE_AOF);
1439 if (flags != REDIS_PROPAGATE_NONE)
1440 propagate(c->cmd,c->db->id,c->argv,c->argc,flags);
1441 }
1442 /* Commands such as LPUSH or BRPOPLPUSH may propagate an additional
1443 * PUSH command. */
1444 if (server.also_propagate.numops) {
1445 int j;
1446 redisOp *rop;
1447
1448 for (j = 0; j < server.also_propagate.numops; j++) {
1449 rop = &server.also_propagate.ops[j];
1450 propagate(rop->cmd, rop->dbid, rop->argv, rop->argc, rop->target);
1451 }
1452 redisOpArrayFree(&server.also_propagate);
1453 }
1454 server.stat_numcommands++;
1455 }
1456
1457 /* If this function gets called we already read a whole
1458 * command, argments are in the client argv/argc fields.
1459 * processCommand() execute the command or prepare the
1460 * server for a bulk read from the client.
1461 *
1462 * If 1 is returned the client is still alive and valid and
1463 * and other operations can be performed by the caller. Otherwise
1464 * if 0 is returned the client was destroied (i.e. after QUIT). */
1465 int processCommand(redisClient *c) {
1466 /* The QUIT command is handled separately. Normal command procs will
1467 * go through checking for replication and QUIT will cause trouble
1468 * when FORCE_REPLICATION is enabled and would be implemented in
1469 * a regular command proc. */
1470 if (!strcasecmp(c->argv[0]->ptr,"quit")) {
1471 addReply(c,shared.ok);
1472 c->flags |= REDIS_CLOSE_AFTER_REPLY;
1473 return REDIS_ERR;
1474 }
1475
1476 /* Now lookup the command and check ASAP about trivial error conditions
1477 * such as wrong arity, bad command name and so forth. */
1478 c->cmd = c->lastcmd = lookupCommand(c->argv[0]->ptr);
1479 if (!c->cmd) {
1480 addReplyErrorFormat(c,"unknown command '%s'",
1481 (char*)c->argv[0]->ptr);
1482 return REDIS_OK;
1483 } else if ((c->cmd->arity > 0 && c->cmd->arity != c->argc) ||
1484 (c->argc < -c->cmd->arity)) {
1485 addReplyErrorFormat(c,"wrong number of arguments for '%s' command",
1486 c->cmd->name);
1487 return REDIS_OK;
1488 }
1489
1490 /* Check if the user is authenticated */
1491 if (server.requirepass && !c->authenticated && c->cmd->proc != authCommand)
1492 {
1493 addReplyError(c,"operation not permitted");
1494 return REDIS_OK;
1495 }
1496
1497 /* Handle the maxmemory directive.
1498 *
1499 * First we try to free some memory if possible (if there are volatile
1500 * keys in the dataset). If there are not the only thing we can do
1501 * is returning an error. */
1502 if (server.maxmemory) {
1503 int retval = freeMemoryIfNeeded();
1504 if ((c->cmd->flags & REDIS_CMD_DENYOOM) && retval == REDIS_ERR) {
1505 addReply(c, shared.oomerr);
1506 return REDIS_OK;
1507 }
1508 }
1509
1510 /* Don't accept write commands if there are problems persisting on disk. */
1511 if (server.stop_writes_on_bgsave_err &&
1512 server.saveparamslen > 0
1513 && server.lastbgsave_status == REDIS_ERR &&
1514 c->cmd->flags & REDIS_CMD_WRITE)
1515 {
1516 addReply(c, shared.bgsaveerr);
1517 return REDIS_OK;
1518 }
1519
1520 /* Don't accept wirte commands if this is a read only slave. But
1521 * accept write commands if this is our master. */
1522 if (server.masterhost && server.repl_slave_ro &&
1523 !(c->flags & REDIS_MASTER) &&
1524 c->cmd->flags & REDIS_CMD_WRITE)
1525 {
1526 addReply(c, shared.roslaveerr);
1527 return REDIS_OK;
1528 }
1529
1530 /* Only allow SUBSCRIBE and UNSUBSCRIBE in the context of Pub/Sub */
1531 if ((dictSize(c->pubsub_channels) > 0 || listLength(c->pubsub_patterns) > 0)
1532 &&
1533 c->cmd->proc != subscribeCommand &&
1534 c->cmd->proc != unsubscribeCommand &&
1535 c->cmd->proc != psubscribeCommand &&
1536 c->cmd->proc != punsubscribeCommand) {
1537 addReplyError(c,"only (P)SUBSCRIBE / (P)UNSUBSCRIBE / QUIT allowed in this context");
1538 return REDIS_OK;
1539 }
1540
1541 /* Only allow INFO and SLAVEOF when slave-serve-stale-data is no and
1542 * we are a slave with a broken link with master. */
1543 if (server.masterhost && server.repl_state != REDIS_REPL_CONNECTED &&
1544 server.repl_serve_stale_data == 0 &&
1545 c->cmd->proc != infoCommand && c->cmd->proc != slaveofCommand)
1546 {
1547 addReplyError(c,
1548 "link with MASTER is down and slave-serve-stale-data is set to no");
1549 return REDIS_OK;
1550 }
1551
1552 /* Loading DB? Return an error if the command is not INFO */
1553 if (server.loading && c->cmd->proc != infoCommand) {
1554 addReply(c, shared.loadingerr);
1555 return REDIS_OK;
1556 }
1557
1558 /* Lua script too slow? Only allow SHUTDOWN NOSAVE and SCRIPT KILL. */
1559 if (server.lua_timedout &&
1560 !(c->cmd->proc != shutdownCommand &&
1561 c->argc == 2 &&
1562 tolower(((char*)c->argv[1]->ptr)[0]) == 'n') &&
1563 !(c->cmd->proc == scriptCommand &&
1564 c->argc == 2 &&
1565 tolower(((char*)c->argv[1]->ptr)[0]) == 'k'))
1566 {
1567 addReply(c, shared.slowscripterr);
1568 return REDIS_OK;
1569 }
1570
1571 /* Exec the command */
1572 if (c->flags & REDIS_MULTI &&
1573 c->cmd->proc != execCommand && c->cmd->proc != discardCommand &&
1574 c->cmd->proc != multiCommand && c->cmd->proc != watchCommand)
1575 {
1576 queueMultiCommand(c);
1577 addReply(c,shared.queued);
1578 } else {
1579 call(c,REDIS_CALL_FULL);
1580 }
1581 return REDIS_OK;
1582 }
1583
1584 /*================================== Shutdown =============================== */
1585
1586 int prepareForShutdown(int flags) {
1587 int save = flags & REDIS_SHUTDOWN_SAVE;
1588 int nosave = flags & REDIS_SHUTDOWN_NOSAVE;
1589
1590 redisLog(REDIS_WARNING,"User requested shutdown...");
1591 /* Kill the saving child if there is a background saving in progress.
1592 We want to avoid race conditions, for instance our saving child may
1593 overwrite the synchronous saving did by SHUTDOWN. */
1594 if (server.rdb_child_pid != -1) {
1595 redisLog(REDIS_WARNING,"There is a child saving an .rdb. Killing it!");
1596 kill(server.rdb_child_pid,SIGKILL);
1597 rdbRemoveTempFile(server.rdb_child_pid);
1598 }
1599 if (server.aof_state != REDIS_AOF_OFF) {
1600 /* Kill the AOF saving child as the AOF we already have may be longer
1601 * but contains the full dataset anyway. */
1602 if (server.aof_child_pid != -1) {
1603 redisLog(REDIS_WARNING,
1604 "There is a child rewriting the AOF. Killing it!");
1605 kill(server.aof_child_pid,SIGKILL);
1606 }
1607 /* Append only file: fsync() the AOF and exit */
1608 redisLog(REDIS_NOTICE,"Calling fsync() on the AOF file.");
1609 aof_fsync(server.aof_fd);
1610 }
1611 if ((server.saveparamslen > 0 && !nosave) || save) {
1612 redisLog(REDIS_NOTICE,"Saving the final RDB snapshot before exiting.");
1613 /* Snapshotting. Perform a SYNC SAVE and exit */
1614 if (rdbSave(server.rdb_filename) != REDIS_OK) {
1615 /* Ooops.. error saving! The best we can do is to continue
1616 * operating. Note that if there was a background saving process,
1617 * in the next cron() Redis will be notified that the background
1618 * saving aborted, handling special stuff like slaves pending for
1619 * synchronization... */
1620 redisLog(REDIS_WARNING,"Error trying to save the DB, can't exit.");
1621 return REDIS_ERR;
1622 }
1623 }
1624 if (server.daemonize) {
1625 redisLog(REDIS_NOTICE,"Removing the pid file.");
1626 unlink(server.pidfile);
1627 }
1628 /* Close the listening sockets. Apparently this allows faster restarts. */
1629 if (server.ipfd != -1) close(server.ipfd);
1630 if (server.sofd != -1) close(server.sofd);
1631 if (server.unixsocket) {
1632 redisLog(REDIS_NOTICE,"Removing the unix socket file.");
1633 unlink(server.unixsocket); /* don't care if this fails */
1634 }
1635
1636 redisLog(REDIS_WARNING,"Redis is now ready to exit, bye bye...");
1637 return REDIS_OK;
1638 }
1639
1640 /*================================== Commands =============================== */
1641
1642 void authCommand(redisClient *c) {
1643 if (!server.requirepass) {
1644 addReplyError(c,"Client sent AUTH, but no password is set");
1645 } else if (!strcmp(c->argv[1]->ptr, server.requirepass)) {
1646 c->authenticated = 1;
1647 addReply(c,shared.ok);
1648 } else {
1649 c->authenticated = 0;
1650 addReplyError(c,"invalid password");
1651 }
1652 }
1653
1654 void pingCommand(redisClient *c) {
1655 addReply(c,shared.pong);
1656 }
1657
1658 void echoCommand(redisClient *c) {
1659 addReplyBulk(c,c->argv[1]);
1660 }
1661
1662 void timeCommand(redisClient *c) {
1663 struct timeval tv;
1664
1665 /* gettimeofday() can only fail if &tv is a bad addresss so we
1666 * don't check for errors. */
1667 gettimeofday(&tv,NULL);
1668 addReplyMultiBulkLen(c,2);
1669 addReplyBulkLongLong(c,tv.tv_sec);
1670 addReplyBulkLongLong(c,tv.tv_usec);
1671 }
1672
1673 /* Convert an amount of bytes into a human readable string in the form
1674 * of 100B, 2G, 100M, 4K, and so forth. */
1675 void bytesToHuman(char *s, unsigned long long n) {
1676 double d;
1677
1678 if (n < 1024) {
1679 /* Bytes */
1680 sprintf(s,"%lluB",n);
1681 return;
1682 } else if (n < (1024*1024)) {
1683 d = (double)n/(1024);
1684 sprintf(s,"%.2fK",d);
1685 } else if (n < (1024LL*1024*1024)) {
1686 d = (double)n/(1024*1024);
1687 sprintf(s,"%.2fM",d);
1688 } else if (n < (1024LL*1024*1024*1024)) {
1689 d = (double)n/(1024LL*1024*1024);
1690 sprintf(s,"%.2fG",d);
1691 }
1692 }
1693
1694 /* Create the string returned by the INFO command. This is decoupled
1695 * by the INFO command itself as we need to report the same information
1696 * on memory corruption problems. */
1697 sds genRedisInfoString(char *section) {
1698 sds info = sdsempty();
1699 time_t uptime = server.unixtime-server.stat_starttime;
1700 int j, numcommands;
1701 struct rusage self_ru, c_ru;
1702 unsigned long lol, bib;
1703 int allsections = 0, defsections = 0;
1704 int sections = 0;
1705
1706 if (section) {
1707 allsections = strcasecmp(section,"all") == 0;
1708 defsections = strcasecmp(section,"default") == 0;
1709 }
1710
1711 getrusage(RUSAGE_SELF, &self_ru);
1712 getrusage(RUSAGE_CHILDREN, &c_ru);
1713 getClientsMaxBuffers(&lol,&bib);
1714
1715 /* Server */
1716 if (allsections || defsections || !strcasecmp(section,"server")) {
1717 struct utsname name;
1718
1719 if (sections++) info = sdscat(info,"\r\n");
1720 uname(&name);
1721 info = sdscatprintf(info,
1722 "# Server\r\n"
1723 "redis_version:%s\r\n"
1724 "redis_git_sha1:%s\r\n"
1725 "redis_git_dirty:%d\r\n"
1726 "os:%s %s %s\r\n"
1727 "arch_bits:%d\r\n"
1728 "multiplexing_api:%s\r\n"
1729 "gcc_version:%d.%d.%d\r\n"
1730 "process_id:%ld\r\n"
1731 "run_id:%s\r\n"
1732 "tcp_port:%d\r\n"
1733 "uptime_in_seconds:%ld\r\n"
1734 "uptime_in_days:%ld\r\n"
1735 "lru_clock:%ld\r\n",
1736 REDIS_VERSION,
1737 redisGitSHA1(),
1738 strtol(redisGitDirty(),NULL,10) > 0,
1739 name.sysname, name.release, name.machine,
1740 server.arch_bits,
1741 aeGetApiName(),
1742 #ifdef __GNUC__
1743 __GNUC__,__GNUC_MINOR__,__GNUC_PATCHLEVEL__,
1744 #else
1745 0,0,0,
1746 #endif
1747 (long) getpid(),
1748 server.runid,
1749 server.port,
1750 uptime,
1751 uptime/(3600*24),
1752 (unsigned long) server.lruclock);
1753 }
1754
1755 /* Clients */
1756 if (allsections || defsections || !strcasecmp(section,"clients")) {
1757 if (sections++) info = sdscat(info,"\r\n");
1758 info = sdscatprintf(info,
1759 "# Clients\r\n"
1760 "connected_clients:%lu\r\n"
1761 "client_longest_output_list:%lu\r\n"
1762 "client_biggest_input_buf:%lu\r\n"
1763 "blocked_clients:%d\r\n",
1764 listLength(server.clients)-listLength(server.slaves),
1765 lol, bib,
1766 server.bpop_blocked_clients);
1767 }
1768
1769 /* Memory */
1770 if (allsections || defsections || !strcasecmp(section,"memory")) {
1771 char hmem[64];
1772 char peak_hmem[64];
1773
1774 bytesToHuman(hmem,zmalloc_used_memory());
1775 bytesToHuman(peak_hmem,server.stat_peak_memory);
1776 if (sections++) info = sdscat(info,"\r\n");
1777 info = sdscatprintf(info,
1778 "# Memory\r\n"
1779 "used_memory:%zu\r\n"
1780 "used_memory_human:%s\r\n"
1781 "used_memory_rss:%zu\r\n"
1782 "used_memory_peak:%zu\r\n"
1783 "used_memory_peak_human:%s\r\n"
1784 "used_memory_lua:%lld\r\n"
1785 "mem_fragmentation_ratio:%.2f\r\n"
1786 "mem_allocator:%s\r\n",
1787 zmalloc_used_memory(),
1788 hmem,
1789 zmalloc_get_rss(),
1790 server.stat_peak_memory,
1791 peak_hmem,
1792 ((long long)lua_gc(server.lua,LUA_GCCOUNT,0))*1024LL,
1793 zmalloc_get_fragmentation_ratio(),
1794 ZMALLOC_LIB
1795 );
1796 }
1797
1798 /* Persistence */
1799 if (allsections || defsections || !strcasecmp(section,"persistence")) {
1800 if (sections++) info = sdscat(info,"\r\n");
1801 info = sdscatprintf(info,
1802 "# Persistence\r\n"
1803 "loading:%d\r\n"
1804 "aof_enabled:%d\r\n"
1805 "changes_since_last_save:%lld\r\n"
1806 "bgsave_in_progress:%d\r\n"
1807 "last_save_time:%ld\r\n"
1808 "last_bgsave_status:%s\r\n"
1809 "bgrewriteaof_in_progress:%d\r\n"
1810 "bgrewriteaof_scheduled:%d\r\n",
1811 server.loading,
1812 server.aof_state != REDIS_AOF_OFF,
1813 server.dirty,
1814 server.rdb_child_pid != -1,
1815 server.lastsave,
1816 server.lastbgsave_status == REDIS_OK ? "ok" : "err",
1817 server.aof_child_pid != -1,
1818 server.aof_rewrite_scheduled);
1819
1820 if (server.aof_state != REDIS_AOF_OFF) {
1821 info = sdscatprintf(info,
1822 "aof_current_size:%lld\r\n"
1823 "aof_base_size:%lld\r\n"
1824 "aof_pending_rewrite:%d\r\n"
1825 "aof_buffer_length:%zu\r\n"
1826 "aof_pending_bio_fsync:%llu\r\n"
1827 "aof_delayed_fsync:%lu\r\n",
1828 (long long) server.aof_current_size,
1829 (long long) server.aof_rewrite_base_size,
1830 server.aof_rewrite_scheduled,
1831 sdslen(server.aof_buf),
1832 bioPendingJobsOfType(REDIS_BIO_AOF_FSYNC),
1833 server.aof_delayed_fsync);
1834 }
1835
1836 if (server.loading) {
1837 double perc;
1838 time_t eta, elapsed;
1839 off_t remaining_bytes = server.loading_total_bytes-
1840 server.loading_loaded_bytes;
1841
1842 perc = ((double)server.loading_loaded_bytes /
1843 server.loading_total_bytes) * 100;
1844
1845 elapsed = server.unixtime-server.loading_start_time;
1846 if (elapsed == 0) {
1847 eta = 1; /* A fake 1 second figure if we don't have
1848 enough info */
1849 } else {
1850 eta = (elapsed*remaining_bytes)/server.loading_loaded_bytes;
1851 }
1852
1853 info = sdscatprintf(info,
1854 "loading_start_time:%ld\r\n"
1855 "loading_total_bytes:%llu\r\n"
1856 "loading_loaded_bytes:%llu\r\n"
1857 "loading_loaded_perc:%.2f\r\n"
1858 "loading_eta_seconds:%ld\r\n"
1859 ,(unsigned long) server.loading_start_time,
1860 (unsigned long long) server.loading_total_bytes,
1861 (unsigned long long) server.loading_loaded_bytes,
1862 perc,
1863 eta
1864 );
1865 }
1866 }
1867
1868 /* Stats */
1869 if (allsections || defsections || !strcasecmp(section,"stats")) {
1870 if (sections++) info = sdscat(info,"\r\n");
1871 info = sdscatprintf(info,
1872 "# Stats\r\n"
1873 "total_connections_received:%lld\r\n"
1874 "total_commands_processed:%lld\r\n"
1875 "instantaneous_ops_per_sec:%lld\r\n"
1876 "rejected_connections:%lld\r\n"
1877 "expired_keys:%lld\r\n"
1878 "evicted_keys:%lld\r\n"
1879 "keyspace_hits:%lld\r\n"
1880 "keyspace_misses:%lld\r\n"
1881 "pubsub_channels:%ld\r\n"
1882 "pubsub_patterns:%lu\r\n"
1883 "latest_fork_usec:%lld\r\n",
1884 server.stat_numconnections,
1885 server.stat_numcommands,
1886 getOperationsPerSecond(),
1887 server.stat_rejected_conn,
1888 server.stat_expiredkeys,
1889 server.stat_evictedkeys,
1890 server.stat_keyspace_hits,
1891 server.stat_keyspace_misses,
1892 dictSize(server.pubsub_channels),
1893 listLength(server.pubsub_patterns),
1894 server.stat_fork_time);
1895 }
1896
1897 /* Replication */
1898 if (allsections || defsections || !strcasecmp(section,"replication")) {
1899 if (sections++) info = sdscat(info,"\r\n");
1900 info = sdscatprintf(info,
1901 "# Replication\r\n"
1902 "role:%s\r\n",
1903 server.masterhost == NULL ? "master" : "slave");
1904 if (server.masterhost) {
1905 info = sdscatprintf(info,
1906 "master_host:%s\r\n"
1907 "master_port:%d\r\n"
1908 "master_link_status:%s\r\n"
1909 "master_last_io_seconds_ago:%d\r\n"
1910 "master_sync_in_progress:%d\r\n"
1911 ,server.masterhost,
1912 server.masterport,
1913 (server.repl_state == REDIS_REPL_CONNECTED) ?
1914 "up" : "down",
1915 server.master ?
1916 ((int)(server.unixtime-server.master->lastinteraction)) : -1,
1917 server.repl_state == REDIS_REPL_TRANSFER
1918 );
1919
1920 if (server.repl_state == REDIS_REPL_TRANSFER) {
1921 info = sdscatprintf(info,
1922 "master_sync_left_bytes:%ld\r\n"
1923 "master_sync_last_io_seconds_ago:%d\r\n"
1924 ,(long)server.repl_transfer_left,
1925 (int)(server.unixtime-server.repl_transfer_lastio)
1926 );
1927 }
1928
1929 if (server.repl_state != REDIS_REPL_CONNECTED) {
1930 info = sdscatprintf(info,
1931 "master_link_down_since_seconds:%ld\r\n",
1932 (long)server.unixtime-server.repl_down_since);
1933 }
1934 }
1935 info = sdscatprintf(info,
1936 "connected_slaves:%lu\r\n",
1937 listLength(server.slaves));
1938 if (listLength(server.slaves)) {
1939 int slaveid = 0;
1940 listNode *ln;
1941 listIter li;
1942
1943 listRewind(server.slaves,&li);
1944 while((ln = listNext(&li))) {
1945 redisClient *slave = listNodeValue(ln);
1946 char *state = NULL;
1947 char ip[32];
1948 int port;
1949
1950 if (anetPeerToString(slave->fd,ip,&port) == -1) continue;
1951 switch(slave->replstate) {
1952 case REDIS_REPL_WAIT_BGSAVE_START:
1953 case REDIS_REPL_WAIT_BGSAVE_END:
1954 state = "wait_bgsave";
1955 break;
1956 case REDIS_REPL_SEND_BULK:
1957 state = "send_bulk";
1958 break;
1959 case REDIS_REPL_ONLINE:
1960 state = "online";
1961 break;
1962 }
1963 if (state == NULL) continue;
1964 info = sdscatprintf(info,"slave%d:%s,%d,%s\r\n",
1965 slaveid,ip,port,state);
1966 slaveid++;
1967 }
1968 }
1969 }
1970
1971 /* CPU */
1972 if (allsections || defsections || !strcasecmp(section,"cpu")) {
1973 if (sections++) info = sdscat(info,"\r\n");
1974 info = sdscatprintf(info,
1975 "# CPU\r\n"
1976 "used_cpu_sys:%.2f\r\n"
1977 "used_cpu_user:%.2f\r\n"
1978 "used_cpu_sys_children:%.2f\r\n"
1979 "used_cpu_user_children:%.2f\r\n",
1980 (float)self_ru.ru_stime.tv_sec+(float)self_ru.ru_stime.tv_usec/1000000,
1981 (float)self_ru.ru_utime.tv_sec+(float)self_ru.ru_utime.tv_usec/1000000,
1982 (float)c_ru.ru_stime.tv_sec+(float)c_ru.ru_stime.tv_usec/1000000,
1983 (float)c_ru.ru_utime.tv_sec+(float)c_ru.ru_utime.tv_usec/1000000);
1984 }
1985
1986 /* cmdtime */
1987 if (allsections || !strcasecmp(section,"commandstats")) {
1988 if (sections++) info = sdscat(info,"\r\n");
1989 info = sdscatprintf(info, "# Commandstats\r\n");
1990 numcommands = sizeof(redisCommandTable)/sizeof(struct redisCommand);
1991 for (j = 0; j < numcommands; j++) {
1992 struct redisCommand *c = redisCommandTable+j;
1993
1994 if (!c->calls) continue;
1995 info = sdscatprintf(info,
1996 "cmdstat_%s:calls=%lld,usec=%lld,usec_per_call=%.2f\r\n",
1997 c->name, c->calls, c->microseconds,
1998 (c->calls == 0) ? 0 : ((float)c->microseconds/c->calls));
1999 }
2000 }
2001
2002 /* Key space */
2003 if (allsections || defsections || !strcasecmp(section,"keyspace")) {
2004 if (sections++) info = sdscat(info,"\r\n");
2005 info = sdscatprintf(info, "# Keyspace\r\n");
2006 for (j = 0; j < server.dbnum; j++) {
2007 long long keys, vkeys;
2008
2009 keys = dictSize(server.db[j].dict);
2010 vkeys = dictSize(server.db[j].expires);
2011 if (keys || vkeys) {
2012 info = sdscatprintf(info, "db%d:keys=%lld,expires=%lld\r\n",
2013 j, keys, vkeys);
2014 }
2015 }
2016 }
2017 return info;
2018 }
2019
2020 void infoCommand(redisClient *c) {
2021 char *section = c->argc == 2 ? c->argv[1]->ptr : "default";
2022
2023 if (c->argc > 2) {
2024 addReply(c,shared.syntaxerr);
2025 return;
2026 }
2027 sds info = genRedisInfoString(section);
2028 addReplySds(c,sdscatprintf(sdsempty(),"$%lu\r\n",
2029 (unsigned long)sdslen(info)));
2030 addReplySds(c,info);
2031 addReply(c,shared.crlf);
2032 }
2033
2034 void monitorCommand(redisClient *c) {
2035 /* ignore MONITOR if aleady slave or in monitor mode */
2036 if (c->flags & REDIS_SLAVE) return;
2037
2038 c->flags |= (REDIS_SLAVE|REDIS_MONITOR);
2039 c->slaveseldb = 0;
2040 listAddNodeTail(server.monitors,c);
2041 addReply(c,shared.ok);
2042 }
2043
2044 /* ============================ Maxmemory directive ======================== */
2045
2046 /* This function gets called when 'maxmemory' is set on the config file to limit
2047 * the max memory used by the server, before processing a command.
2048 *
2049 * The goal of the function is to free enough memory to keep Redis under the
2050 * configured memory limit.
2051 *
2052 * The function starts calculating how many bytes should be freed to keep
2053 * Redis under the limit, and enters a loop selecting the best keys to
2054 * evict accordingly to the configured policy.
2055 *
2056 * If all the bytes needed to return back under the limit were freed the
2057 * function returns REDIS_OK, otherwise REDIS_ERR is returned, and the caller
2058 * should block the execution of commands that will result in more memory
2059 * used by the server.
2060 */
2061 int freeMemoryIfNeeded(void) {
2062 size_t mem_used, mem_tofree, mem_freed;
2063 int slaves = listLength(server.slaves);
2064
2065 /* Remove the size of slaves output buffers and AOF buffer from the
2066 * count of used memory. */
2067 mem_used = zmalloc_used_memory();
2068 if (slaves) {
2069 listIter li;
2070 listNode *ln;
2071
2072 listRewind(server.slaves,&li);
2073 while((ln = listNext(&li))) {
2074 redisClient *slave = listNodeValue(ln);
2075 unsigned long obuf_bytes = getClientOutputBufferMemoryUsage(slave);
2076 if (obuf_bytes > mem_used)
2077 mem_used = 0;
2078 else
2079 mem_used -= obuf_bytes;
2080 }
2081 }
2082 if (server.aof_state != REDIS_AOF_OFF) {
2083 mem_used -= sdslen(server.aof_buf);
2084 mem_used -= sdslen(server.aof_rewrite_buf);
2085 }
2086
2087 /* Check if we are over the memory limit. */
2088 if (mem_used <= server.maxmemory) return REDIS_OK;
2089
2090 if (server.maxmemory_policy == REDIS_MAXMEMORY_NO_EVICTION)
2091 return REDIS_ERR; /* We need to free memory, but policy forbids. */
2092
2093 /* Compute how much memory we need to free. */
2094 mem_tofree = mem_used - server.maxmemory;
2095 mem_freed = 0;
2096 while (mem_freed < mem_tofree) {
2097 int j, k, keys_freed = 0;
2098
2099 for (j = 0; j < server.dbnum; j++) {
2100 long bestval = 0; /* just to prevent warning */
2101 sds bestkey = NULL;
2102 struct dictEntry *de;
2103 redisDb *db = server.db+j;
2104 dict *dict;
2105
2106 if (server.maxmemory_policy == REDIS_MAXMEMORY_ALLKEYS_LRU ||
2107 server.maxmemory_policy == REDIS_MAXMEMORY_ALLKEYS_RANDOM)
2108 {
2109 dict = server.db[j].dict;
2110 } else {
2111 dict = server.db[j].expires;
2112 }
2113 if (dictSize(dict) == 0) continue;
2114
2115 /* volatile-random and allkeys-random policy */
2116 if (server.maxmemory_policy == REDIS_MAXMEMORY_ALLKEYS_RANDOM ||
2117 server.maxmemory_policy == REDIS_MAXMEMORY_VOLATILE_RANDOM)
2118 {
2119 de = dictGetRandomKey(dict);
2120 bestkey = dictGetKey(de);
2121 }
2122
2123 /* volatile-lru and allkeys-lru policy */
2124 else if (server.maxmemory_policy == REDIS_MAXMEMORY_ALLKEYS_LRU ||
2125 server.maxmemory_policy == REDIS_MAXMEMORY_VOLATILE_LRU)
2126 {
2127 for (k = 0; k < server.maxmemory_samples; k++) {
2128 sds thiskey;
2129 long thisval;
2130 robj *o;
2131
2132 de = dictGetRandomKey(dict);
2133 thiskey = dictGetKey(de);
2134 /* When policy is volatile-lru we need an additonal lookup
2135 * to locate the real key, as dict is set to db->expires. */
2136 if (server.maxmemory_policy == REDIS_MAXMEMORY_VOLATILE_LRU)
2137 de = dictFind(db->dict, thiskey);
2138 o = dictGetVal(de);
2139 thisval = estimateObjectIdleTime(o);
2140
2141 /* Higher idle time is better candidate for deletion */
2142 if (bestkey == NULL || thisval > bestval) {
2143 bestkey = thiskey;
2144 bestval = thisval;
2145 }
2146 }
2147 }
2148
2149 /* volatile-ttl */
2150 else if (server.maxmemory_policy == REDIS_MAXMEMORY_VOLATILE_TTL) {
2151 for (k = 0; k < server.maxmemory_samples; k++) {
2152 sds thiskey;
2153 long thisval;
2154
2155 de = dictGetRandomKey(dict);
2156 thiskey = dictGetKey(de);
2157 thisval = (long) dictGetVal(de);
2158
2159 /* Expire sooner (minor expire unix timestamp) is better
2160 * candidate for deletion */
2161 if (bestkey == NULL || thisval < bestval) {
2162 bestkey = thiskey;
2163 bestval = thisval;
2164 }
2165 }
2166 }
2167
2168 /* Finally remove the selected key. */
2169 if (bestkey) {
2170 long long delta;
2171
2172 robj *keyobj = createStringObject(bestkey,sdslen(bestkey));
2173 propagateExpire(db,keyobj);
2174 /* We compute the amount of memory freed by dbDelete() alone.
2175 * It is possible that actually the memory needed to propagate
2176 * the DEL in AOF and replication link is greater than the one
2177 * we are freeing removing the key, but we can't account for
2178 * that otherwise we would never exit the loop.
2179 *
2180 * AOF and Output buffer memory will be freed eventually so
2181 * we only care about memory used by the key space. */
2182 delta = (long long) zmalloc_used_memory();
2183 dbDelete(db,keyobj);
2184 delta -= (long long) zmalloc_used_memory();
2185 mem_freed += delta;
2186 server.stat_evictedkeys++;
2187 decrRefCount(keyobj);
2188 keys_freed++;
2189
2190 /* When the memory to free starts to be big enough, we may
2191 * start spending so much time here that is impossible to
2192 * deliver data to the slaves fast enough, so we force the
2193 * transmission here inside the loop. */
2194 if (slaves) flushSlavesOutputBuffers();
2195 }
2196 }
2197 if (!keys_freed) return REDIS_ERR; /* nothing to free... */
2198 }
2199 return REDIS_OK;
2200 }
2201
2202 /* =================================== Main! ================================ */
2203
2204 #ifdef __linux__
2205 int linuxOvercommitMemoryValue(void) {
2206 FILE *fp = fopen("/proc/sys/vm/overcommit_memory","r");
2207 char buf[64];
2208
2209 if (!fp) return -1;
2210 if (fgets(buf,64,fp) == NULL) {
2211 fclose(fp);
2212 return -1;
2213 }
2214 fclose(fp);
2215
2216 return atoi(buf);
2217 }
2218
2219 void linuxOvercommitMemoryWarning(void) {
2220 if (linuxOvercommitMemoryValue() == 0) {
2221 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.");
2222 }
2223 }
2224 #endif /* __linux__ */
2225
2226 void createPidFile(void) {
2227 /* Try to write the pid file in a best-effort way. */
2228 FILE *fp = fopen(server.pidfile,"w");
2229 if (fp) {
2230 fprintf(fp,"%d\n",(int)getpid());
2231 fclose(fp);
2232 }
2233 }
2234
2235 void daemonize(void) {
2236 int fd;
2237
2238 if (fork() != 0) exit(0); /* parent exits */
2239 setsid(); /* create a new session */
2240
2241 /* Every output goes to /dev/null. If Redis is daemonized but
2242 * the 'logfile' is set to 'stdout' in the configuration file
2243 * it will not log at all. */
2244 if ((fd = open("/dev/null", O_RDWR, 0)) != -1) {
2245 dup2(fd, STDIN_FILENO);
2246 dup2(fd, STDOUT_FILENO);
2247 dup2(fd, STDERR_FILENO);
2248 if (fd > STDERR_FILENO) close(fd);
2249 }
2250 }
2251
2252 void version() {
2253 printf("Redis server v=%s sha=%s:%d malloc=%s bits=%d\n",
2254 REDIS_VERSION,
2255 redisGitSHA1(),
2256 atoi(redisGitDirty()) > 0,
2257 ZMALLOC_LIB,
2258 sizeof(long) == 4 ? 32 : 64);
2259 exit(0);
2260 }
2261
2262 void usage() {
2263 fprintf(stderr,"Usage: ./redis-server [/path/to/redis.conf] [options]\n");
2264 fprintf(stderr," ./redis-server - (read config from stdin)\n");
2265 fprintf(stderr," ./redis-server -v or --version\n");
2266 fprintf(stderr," ./redis-server -h or --help\n");
2267 fprintf(stderr," ./redis-server --test-memory <megabytes>\n\n");
2268 fprintf(stderr,"Examples:\n");
2269 fprintf(stderr," ./redis-server (run the server with default conf)\n");
2270 fprintf(stderr," ./redis-server /etc/redis/6379.conf\n");
2271 fprintf(stderr," ./redis-server --port 7777\n");
2272 fprintf(stderr," ./redis-server --port 7777 --slaveof 127.0.0.1 8888\n");
2273 fprintf(stderr," ./redis-server /etc/myredis.conf --loglevel verbose\n");
2274 exit(1);
2275 }
2276
2277 void redisAsciiArt(void) {
2278 #include "asciilogo.h"
2279 char *buf = zmalloc(1024*16);
2280
2281 snprintf(buf,1024*16,ascii_logo,
2282 REDIS_VERSION,
2283 redisGitSHA1(),
2284 strtol(redisGitDirty(),NULL,10) > 0,
2285 (sizeof(long) == 8) ? "64" : "32",
2286 "stand alone",
2287 server.port,
2288 (long) getpid()
2289 );
2290 redisLogRaw(REDIS_NOTICE|REDIS_LOG_RAW,buf);
2291 zfree(buf);
2292 }
2293
2294 static void sigtermHandler(int sig) {
2295 REDIS_NOTUSED(sig);
2296
2297 redisLogFromHandler(REDIS_WARNING,"Received SIGTERM, scheduling shutdown...");
2298 server.shutdown_asap = 1;
2299 }
2300
2301 void setupSignalHandlers(void) {
2302 struct sigaction act;
2303 stack_t stack;
2304
2305 stack.ss_sp = altstack;
2306 stack.ss_flags = 0;
2307 stack.ss_size = SIGSTKSZ;
2308
2309 sigaltstack(&stack, NULL);
2310
2311 /* When the SA_SIGINFO flag is set in sa_flags then sa_sigaction is used.
2312 * Otherwise, sa_handler is used. */
2313 sigemptyset(&act.sa_mask);
2314 act.sa_flags = 0;
2315 act.sa_handler = sigtermHandler;
2316 sigaction(SIGTERM, &act, NULL);
2317
2318 #ifdef HAVE_BACKTRACE
2319 /* Use alternate stack so we don't clobber stack in case of segv, or when we run out of stack ..
2320 * also resethand & nodefer so we can get interrupted (and killed) if we cause SEGV during SEGV handler */
2321 sigemptyset(&act.sa_mask);
2322 act.sa_flags = SA_NODEFER | SA_ONSTACK | SA_RESETHAND | SA_SIGINFO;
2323 act.sa_sigaction = sigsegvHandler;
2324 sigaction(SIGSEGV, &act, NULL);
2325 sigaction(SIGBUS, &act, NULL);
2326 sigaction(SIGFPE, &act, NULL);
2327 sigaction(SIGILL, &act, NULL);
2328 #endif
2329 return;
2330 }
2331
2332 void memtest(size_t megabytes, int passes);
2333
2334 int main(int argc, char **argv) {
2335 long long start;
2336 struct timeval tv;
2337
2338 /* We need to initialize our libraries, and the server configuration. */
2339 zmalloc_enable_thread_safeness();
2340 srand(time(NULL)^getpid());
2341 gettimeofday(&tv,NULL);
2342 dictSetHashFunctionSeed(tv.tv_sec^tv.tv_usec^getpid());
2343 initServerConfig();
2344
2345 if (argc >= 2) {
2346 int j = 1; /* First option to parse in argv[] */
2347 sds options = sdsempty();
2348 char *configfile = NULL;
2349
2350 /* Handle special options --help and --version */
2351 if (strcmp(argv[1], "-v") == 0 ||
2352 strcmp(argv[1], "--version") == 0) version();
2353 if (strcmp(argv[1], "--help") == 0 ||
2354 strcmp(argv[1], "-h") == 0) usage();
2355 if (strcmp(argv[1], "--test-memory") == 0) {
2356 if (argc == 3) {
2357 memtest(atoi(argv[2]),50);
2358 exit(0);
2359 } else {
2360 fprintf(stderr,"Please specify the amount of memory to test in megabytes.\n");
2361 fprintf(stderr,"Example: ./redis-server --test-memory 4096\n\n");
2362 exit(1);
2363 }
2364 }
2365
2366 /* First argument is the config file name? */
2367 if (argv[j][0] != '-' || argv[j][1] != '-')
2368 configfile = argv[j++];
2369 /* All the other options are parsed and conceptually appended to the
2370 * configuration file. For instance --port 6380 will generate the
2371 * string "port 6380\n" to be parsed after the actual file name
2372 * is parsed, if any. */
2373 while(j != argc) {
2374 if (argv[j][0] == '-' && argv[j][1] == '-') {
2375 /* Option name */
2376 if (sdslen(options)) options = sdscat(options,"\n");
2377 options = sdscat(options,argv[j]+2);
2378 options = sdscat(options," ");
2379 } else {
2380 /* Option argument */
2381 options = sdscatrepr(options,argv[j],strlen(argv[j]));
2382 options = sdscat(options," ");
2383 }
2384 j++;
2385 }
2386 resetServerSaveParams();
2387 loadServerConfig(configfile,options);
2388 sdsfree(options);
2389 } else {
2390 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'");
2391 }
2392 if (server.daemonize) daemonize();
2393 initServer();
2394 if (server.daemonize) createPidFile();
2395 redisAsciiArt();
2396 redisLog(REDIS_WARNING,"Server started, Redis version " REDIS_VERSION);
2397 #ifdef __linux__
2398 linuxOvercommitMemoryWarning();
2399 #endif
2400 start = ustime();
2401 if (server.aof_state == REDIS_AOF_ON) {
2402 if (loadAppendOnlyFile(server.aof_filename) == REDIS_OK)
2403 redisLog(REDIS_NOTICE,"DB loaded from append only file: %.3f seconds",(float)(ustime()-start)/1000000);
2404 } else {
2405 if (rdbLoad(server.rdb_filename) == REDIS_OK) {
2406 redisLog(REDIS_NOTICE,"DB loaded from disk: %.3f seconds",
2407 (float)(ustime()-start)/1000000);
2408 } else if (errno != ENOENT) {
2409 redisLog(REDIS_WARNING,"Fatal error loading the DB. Exiting.");
2410 exit(1);
2411 }
2412 }
2413 if (server.ipfd > 0)
2414 redisLog(REDIS_NOTICE,"The server is now ready to accept connections on port %d", server.port);
2415 if (server.sofd > 0)
2416 redisLog(REDIS_NOTICE,"The server is now ready to accept connections at %s", server.unixsocket);
2417 aeSetBeforeSleepProc(server.el,beforeSleep);
2418 aeMain(server.el);
2419 aeDeleteEventLoop(server.el);
2420 return 0;
2421 }
2422
2423 /* The End */