X-Git-Url: https://git.saurik.com/redis.git/blobdiff_plain/c1ef6ffe8a53e573b6b9e8d1a02c35e2c62cdad8..3e6a4463e02dc4a39cf4f73704dbcabacd7db8ff:/src/redis.c diff --git a/src/redis.c b/src/redis.c index 4b9ba618..9966f945 100644 --- a/src/redis.c +++ b/src/redis.c @@ -48,6 +48,7 @@ #include #include #include +#include /* Our shared "common" objects */ @@ -61,6 +62,9 @@ double R_Zero, R_PosInf, R_NegInf, R_Nan; /*================================= Globals ================================= */ +/* Alternate stack for SIGSEGV/etc handlers */ +char altstack[SIGSTKSZ]; + /* Global vars */ struct redisServer server; /* server global state */ struct redisCommand *commandTable; @@ -211,7 +215,7 @@ struct redisCommand redisCommandTable[] = { {"lastsave",lastsaveCommand,1,"r",0,NULL,0,0,0,0,0}, {"type",typeCommand,2,"r",0,NULL,1,1,1,0,0}, {"multi",multiCommand,1,"rs",0,NULL,0,0,0,0,0}, - {"exec",execCommand,1,"wms",0,NULL,0,0,0,0,0}, + {"exec",execCommand,1,"s",0,NULL,0,0,0,0,0}, {"discard",discardCommand,1,"rs",0,NULL,0,0,0,0,0}, {"sync",syncCommand,1,"ars",0,NULL,0,0,0,0,0}, {"flushdb",flushdbCommand,1,"w",0,NULL,0,0,0,0,0}, @@ -222,27 +226,26 @@ struct redisCommand redisCommandTable[] = { {"ttl",ttlCommand,2,"r",0,NULL,1,1,1,0,0}, {"pttl",pttlCommand,2,"r",0,NULL,1,1,1,0,0}, {"persist",persistCommand,2,"w",0,NULL,1,1,1,0,0}, - {"slaveof",slaveofCommand,3,"aws",0,NULL,0,0,0,0,0}, - {"debug",debugCommand,-2,"aws",0,NULL,0,0,0,0,0}, + {"slaveof",slaveofCommand,3,"as",0,NULL,0,0,0,0,0}, + {"debug",debugCommand,-2,"as",0,NULL,0,0,0,0,0}, {"config",configCommand,-2,"ar",0,NULL,0,0,0,0,0}, {"subscribe",subscribeCommand,-2,"rps",0,NULL,0,0,0,0,0}, {"unsubscribe",unsubscribeCommand,-1,"rps",0,NULL,0,0,0,0,0}, {"psubscribe",psubscribeCommand,-2,"rps",0,NULL,0,0,0,0,0}, {"punsubscribe",punsubscribeCommand,-1,"rps",0,NULL,0,0,0,0,0}, - {"publish",publishCommand,3,"rpf",0,NULL,0,0,0,0,0}, + {"publish",publishCommand,3,"pf",0,NULL,0,0,0,0,0}, {"watch",watchCommand,-2,"rs",0,noPreloadGetKeys,1,-1,1,0,0}, {"unwatch",unwatchCommand,1,"rs",0,NULL,0,0,0,0,0}, - {"cluster",clusterCommand,-2,"ar",0,NULL,0,0,0,0,0}, {"restore",restoreCommand,4,"awm",0,NULL,1,1,1,0,0}, {"migrate",migrateCommand,6,"aw",0,NULL,0,0,0,0,0}, - {"asking",askingCommand,1,"r",0,NULL,0,0,0,0,0}, {"dump",dumpCommand,2,"ar",0,NULL,1,1,1,0,0}, {"object",objectCommand,-2,"r",0,NULL,2,2,2,0,0}, {"client",clientCommand,-2,"ar",0,NULL,0,0,0,0,0}, - {"eval",evalCommand,-3,"wms",0,zunionInterGetKeys,0,0,0,0,0}, - {"evalsha",evalShaCommand,-3,"wms",0,zunionInterGetKeys,0,0,0,0,0}, + {"eval",evalCommand,-3,"s",0,zunionInterGetKeys,0,0,0,0,0}, + {"evalsha",evalShaCommand,-3,"s",0,zunionInterGetKeys,0,0,0,0,0}, {"slowlog",slowlogCommand,-2,"r",0,NULL,0,0,0,0,0}, - {"script",scriptCommand,-2,"ras",0,NULL,0,0,0,0,0} + {"script",scriptCommand,-2,"ras",0,NULL,0,0,0,0,0}, + {"time",timeCommand,1,"rR",0,NULL,0,0,0,0,0} }; /*============================ Utility functions ============================ */ @@ -252,7 +255,6 @@ struct redisCommand redisCommandTable[] = { void redisLogRaw(int level, const char *msg) { const int syslogLevelMap[] = { LOG_DEBUG, LOG_INFO, LOG_NOTICE, LOG_WARNING }; const char *c = ".-*#"; - time_t now = time(NULL); FILE *fp; char buf[64]; int rawmode = (level & REDIS_LOG_RAW); @@ -266,7 +268,12 @@ void redisLogRaw(int level, const char *msg) { if (rawmode) { fprintf(fp,"%s",msg); } else { - strftime(buf,sizeof(buf),"%d %b %H:%M:%S",localtime(&now)); + int off; + struct timeval tv; + + gettimeofday(&tv,NULL); + off = strftime(buf,sizeof(buf),"%d %b %H:%M:%S.",localtime(&tv.tv_sec)); + snprintf(buf+off,sizeof(buf)-off,"%03d",(int)tv.tv_usec/1000); fprintf(fp,"[%d] %s %c %s\n",(int)getpid(),buf,c[level],msg); } fflush(fp); @@ -292,6 +299,35 @@ void redisLog(int level, const char *fmt, ...) { redisLogRaw(level,msg); } +/* Log a fixed message without printf-alike capabilities, in a way that is + * safe to call from a signal handler. + * + * We actually use this only for signals that are not fatal from the point + * of view of Redis. Signals that are going to kill the server anyway and + * where we need printf-alike features are served by redisLog(). */ +void redisLogFromHandler(int level, const char *msg) { + int fd; + char buf[64]; + + if ((level&0xff) < server.verbosity || + (server.logfile == NULL && server.daemonize)) return; + fd = server.logfile ? + open(server.logfile, O_APPEND|O_CREAT|O_WRONLY, 0644) : + STDOUT_FILENO; + if (fd == -1) return; + ll2string(buf,sizeof(buf),getpid()); + if (write(fd,"[",1) == -1) goto err; + if (write(fd,buf,strlen(buf)) == -1) goto err; + if (write(fd," | signal handler] (",20) == -1) goto err; + ll2string(buf,sizeof(buf),time(NULL)); + if (write(fd,buf,strlen(buf)) == -1) goto err; + if (write(fd,") ",2) == -1) goto err; + if (write(fd,msg,strlen(msg)) == -1) goto err; + if (write(fd,"\n",1) == -1) goto err; +err: + if (server.logfile) close(fd); +} + /* Redis generally does not try to recover from out of memory conditions * when allocating objects or strings, it is not clear if it will be possible * to report this condition to the client since the networking layer itself @@ -319,6 +355,18 @@ long long mstime(void) { return ustime()/1000; } +/* After an RDB dump or AOF rewrite we exit from children using _exit() instead of + * exit(), because the latter may interact with the same file objects used by + * the parent process. However if we are testing the coverage normal exit() is + * used in order to obtain the right coverage information. */ +void exitFromChild(int retcode) { +#ifdef COVERAGE_TEST + exit(retcode); +#else + _exit(retcode); +#endif +} + /*====================== Hash table type implementation ==================== */ /* This is an hash table type that uses the SDS dynamic strings libary as @@ -506,17 +554,6 @@ dictType keylistDictType = { dictListDestructor /* val destructor */ }; -/* Cluster nodes hash table, mapping nodes addresses 1.2.3.4:6379 to - * clusterNode structures. */ -dictType clusterNodesDictType = { - dictSdsHash, /* hash function */ - NULL, /* key dup */ - NULL, /* val dup */ - dictSdsKeyCompare, /* key compare */ - dictSdsDestructor, /* key destructor */ - NULL /* val destructor */ -}; - int htNeedsResize(dict *dict) { long long size, used; @@ -611,22 +648,132 @@ void activeExpireCycle(void) { } void updateLRUClock(void) { - server.lruclock = (time(NULL)/REDIS_LRU_CLOCK_RESOLUTION) & + server.lruclock = (server.unixtime/REDIS_LRU_CLOCK_RESOLUTION) & REDIS_LRU_CLOCK_MAX; } + +/* Add a sample to the operations per second array of samples. */ +void trackOperationsPerSecond(void) { + long long t = mstime() - server.ops_sec_last_sample_time; + long long ops = server.stat_numcommands - server.ops_sec_last_sample_ops; + long long ops_sec; + + ops_sec = t > 0 ? (ops*1000/t) : 0; + + server.ops_sec_samples[server.ops_sec_idx] = ops_sec; + server.ops_sec_idx = (server.ops_sec_idx+1) % REDIS_OPS_SEC_SAMPLES; + server.ops_sec_last_sample_time = mstime(); + server.ops_sec_last_sample_ops = server.stat_numcommands; +} + +/* Return the mean of all the samples. */ +long long getOperationsPerSecond(void) { + int j; + long long sum = 0; + + for (j = 0; j < REDIS_OPS_SEC_SAMPLES; j++) + sum += server.ops_sec_samples[j]; + return sum / REDIS_OPS_SEC_SAMPLES; +} + +/* Check for timeouts. Returns non-zero if the client was terminated */ +int clientsCronHandleTimeout(redisClient *c) { + time_t now = server.unixtime; + + if (server.maxidletime && + !(c->flags & REDIS_SLAVE) && /* no timeout for slaves */ + !(c->flags & REDIS_MASTER) && /* no timeout for masters */ + !(c->flags & REDIS_BLOCKED) && /* no timeout for BLPOP */ + dictSize(c->pubsub_channels) == 0 && /* no timeout for pubsub */ + listLength(c->pubsub_patterns) == 0 && + (now - c->lastinteraction > server.maxidletime)) + { + redisLog(REDIS_VERBOSE,"Closing idle client"); + freeClient(c); + return 1; + } else if (c->flags & REDIS_BLOCKED) { + if (c->bpop.timeout != 0 && c->bpop.timeout < now) { + addReply(c,shared.nullmultibulk); + unblockClientWaitingData(c); + } + } + return 0; +} + +/* The client query buffer is an sds.c string that can end with a lot of + * free space not used, this function reclaims space if needed. + * + * The funciton always returns 0 as it never terminates the client. */ +int clientsCronResizeQueryBuffer(redisClient *c) { + size_t querybuf_size = sdsAllocSize(c->querybuf); + time_t idletime = server.unixtime - c->lastinteraction; + + /* There are two conditions to resize the query buffer: + * 1) Query buffer is > BIG_ARG and too big for latest peak. + * 2) Client is inactive and the buffer is bigger than 1k. */ + if (((querybuf_size > REDIS_MBULK_BIG_ARG) && + (querybuf_size/(c->querybuf_peak+1)) > 2) || + (querybuf_size > 1024 && idletime > 2)) + { + /* Only resize the query buffer if it is actually wasting space. */ + if (sdsavail(c->querybuf) > 1024) { + c->querybuf = sdsRemoveFreeSpace(c->querybuf); + } + } + /* Reset the peak again to capture the peak memory usage in the next + * cycle. */ + c->querybuf_peak = 0; + return 0; +} + +void clientsCron(void) { + /* Make sure to process at least 1/100 of clients per call. + * Since this function is called 10 times per second we are sure that + * in the worst case we process all the clients in 10 seconds. + * In normal conditions (a reasonable number of clients) we process + * all the clients in a shorter time. */ + int numclients = listLength(server.clients); + int iterations = numclients/100; + + if (iterations < 50) + iterations = (numclients < 50) ? numclients : 50; + while(listLength(server.clients) && iterations--) { + redisClient *c; + listNode *head; + + /* Rotate the list, take the current head, process. + * This way if the client must be removed from the list it's the + * first element and we don't incur into O(N) computation. */ + listRotate(server.clients); + head = listFirst(server.clients); + c = listNodeValue(head); + /* The following functions do different service checks on the client. + * The protocol is that they return non-zero if the client was + * terminated. */ + if (clientsCronHandleTimeout(c)) continue; + if (clientsCronResizeQueryBuffer(c)) continue; + } +} + int serverCron(struct aeEventLoop *eventLoop, long long id, void *clientData) { int j, loops = server.cronloops; REDIS_NOTUSED(eventLoop); REDIS_NOTUSED(id); REDIS_NOTUSED(clientData); + /* Software watchdog: deliver the SIGALRM that will reach the signal + * handler if we don't return here fast enough. */ + if (server.watchdog_period) watchdogScheduleSignal(server.watchdog_period); + /* We take a cached value of the unix time in the global state because * with virtual memory and aging there is to store the current time * in objects at every object access, and accuracy is not needed. * To access a global var is faster than calling time(NULL) */ server.unixtime = time(NULL); + trackOperationsPerSecond(); + /* We have just 22 bits per object for LRU information. * So we use an (eventually wrapping) LRU clock with 10 seconds resolution. * 2^22 bits with 10 seconds resoluton is more or less 1.5 years. @@ -684,9 +831,8 @@ int serverCron(struct aeEventLoop *eventLoop, long long id, void *clientData) { zmalloc_used_memory()); } - /* Close connections of timedout clients */ - if ((server.maxidletime && !(loops % 100)) || server.bpop_blocked_clients) - closeTimedoutClients(); + /* We need to do a few operations on clients asynchronously. */ + clientsCron(); /* Start a scheduled AOF rewrite if this was requested by the user while * a BGSAVE was in progress. */ @@ -715,15 +861,13 @@ int serverCron(struct aeEventLoop *eventLoop, long long id, void *clientData) { updateDictResizePolicy(); } } else { - time_t now = time(NULL); - /* If there is not a background saving/rewrite in progress check if * we have to save/rewrite now */ for (j = 0; j < server.saveparamslen; j++) { struct saveparam *sp = server.saveparams+j; if (server.dirty >= sp->changes && - now-server.lastsave > sp->seconds) { + server.unixtime-server.lastsave > sp->seconds) { redisLog(REDIS_NOTICE,"%d changes in %d seconds. Saving...", sp->changes, sp->seconds); rdbSaveBackground(server.rdb_filename); @@ -764,9 +908,6 @@ int serverCron(struct aeEventLoop *eventLoop, long long id, void *clientData) { * to detect transfer failures. */ if (!(loops % 10)) replicationCron(); - /* Run other sub-systems specific cron jobs */ - if (server.cluster_enabled && !(loops % 10)) clusterCron(); - server.cronloops++; return 100; } @@ -832,19 +973,20 @@ void createSharedObjects(void) { "-LOADING Redis is loading the dataset in memory\r\n")); shared.slowscripterr = createObject(REDIS_STRING,sdsnew( "-BUSY Redis is busy running a script. You can only call SCRIPT KILL or SHUTDOWN NOSAVE.\r\n")); + shared.bgsaveerr = createObject(REDIS_STRING,sdsnew( + "-MISCONF Redis is configured to save RDB snapshots, but is currently not able to persist on disk. Commands that may modify the data set are disabled. Please check Redis logs for details about the error.\r\n")); + shared.roslaveerr = createObject(REDIS_STRING,sdsnew( + "-READONLY You can't write against a read only slave.\r\n")); + shared.oomerr = createObject(REDIS_STRING,sdsnew( + "-OOM command not allowed when used memory > 'maxmemory'.\r\n")); shared.space = createObject(REDIS_STRING,sdsnew(" ")); shared.colon = createObject(REDIS_STRING,sdsnew(":")); shared.plus = createObject(REDIS_STRING,sdsnew("+")); - shared.select0 = createStringObject("select 0\r\n",10); - shared.select1 = createStringObject("select 1\r\n",10); - shared.select2 = createStringObject("select 2\r\n",10); - shared.select3 = createStringObject("select 3\r\n",10); - shared.select4 = createStringObject("select 4\r\n",10); - shared.select5 = createStringObject("select 5\r\n",10); - shared.select6 = createStringObject("select 6\r\n",10); - shared.select7 = createStringObject("select 7\r\n",10); - shared.select8 = createStringObject("select 8\r\n",10); - shared.select9 = createStringObject("select 9\r\n",10); + + for (j = 0; j < REDIS_SHARED_SELECT_CMDS; j++) { + shared.select[j] = createObject(REDIS_STRING, + sdscatprintf(sdsempty(),"select %d\r\n", j)); + } shared.messagebulk = createStringObject("$7\r\nmessage\r\n",13); shared.pmessagebulk = createStringObject("$8\r\npmessage\r\n",14); shared.subscribebulk = createStringObject("$9\r\nsubscribe\r\n",15); @@ -852,6 +994,8 @@ void createSharedObjects(void) { shared.psubscribebulk = createStringObject("$10\r\npsubscribe\r\n",17); shared.punsubscribebulk = createStringObject("$12\r\npunsubscribe\r\n",19); shared.del = createStringObject("DEL",3); + shared.rpop = createStringObject("RPOP",4); + shared.lpop = createStringObject("LPOP",4); for (j = 0; j < REDIS_SHARED_INTEGERS; j++) { shared.integers[j] = createObject(REDIS_STRING,(void*)(long)j); shared.integers[j]->encoding = REDIS_ENCODING_INT; @@ -865,6 +1009,8 @@ void createSharedObjects(void) { } void initServerConfig() { + getRandomHexChars(server.runid,REDIS_RUN_ID_SIZE); + server.runid[REDIS_RUN_ID_SIZE] = '\0'; server.arch_bits = (sizeof(long) == 8) ? 64 : 32; server.port = REDIS_SERVERPORT; server.bindaddr = NULL; @@ -891,6 +1037,7 @@ void initServerConfig() { server.aof_rewrite_base_size = 0; server.aof_rewrite_scheduled = 0; server.aof_last_fsync = time(NULL); + server.aof_delayed_fsync = 0; server.aof_fd = -1; server.aof_selected_db = -1; /* Make sure the first time will not match */ server.aof_flush_postponed_start = 0; @@ -899,14 +1046,15 @@ void initServerConfig() { server.aof_filename = zstrdup("appendonly.aof"); server.requirepass = NULL; server.rdb_compression = 1; + server.rdb_checksum = 1; server.activerehashing = 1; server.maxclients = REDIS_MAX_CLIENTS; server.bpop_blocked_clients = 0; server.maxmemory = 0; server.maxmemory_policy = REDIS_MAXMEMORY_VOLATILE_LRU; server.maxmemory_samples = 3; - server.hash_max_zipmap_entries = REDIS_HASH_MAX_ZIPMAP_ENTRIES; - server.hash_max_zipmap_value = REDIS_HASH_MAX_ZIPMAP_VALUE; + server.hash_max_ziplist_entries = REDIS_HASH_MAX_ZIPLIST_ENTRIES; + server.hash_max_ziplist_value = REDIS_HASH_MAX_ZIPLIST_VALUE; server.list_max_ziplist_entries = REDIS_LIST_MAX_ZIPLIST_ENTRIES; server.list_max_ziplist_value = REDIS_LIST_MAX_ZIPLIST_VALUE; server.set_max_intset_entries = REDIS_SET_MAX_INTSET_ENTRIES; @@ -915,12 +1063,11 @@ void initServerConfig() { server.shutdown_asap = 0; server.repl_ping_slave_period = REDIS_REPL_PING_SLAVE_PERIOD; server.repl_timeout = REDIS_REPL_TIMEOUT; - server.cluster_enabled = 0; - server.cluster.configfile = zstrdup("nodes.conf"); server.lua_caller = NULL; server.lua_time_limit = REDIS_LUA_TIME_LIMIT; server.lua_client = NULL; server.lua_timedout = 0; + server.lua_protect_globals = 1; updateLRUClock(); resetServerSaveParams(); @@ -936,7 +1083,8 @@ void initServerConfig() { server.repl_state = REDIS_REPL_NONE; server.repl_syncio_timeout = REDIS_REPL_SYNCIO_TIMEOUT; server.repl_serve_stale_data = 1; - server.repl_down_since = -1; + server.repl_slave_ro = 1; + server.repl_down_since = time(NULL); /* Client output buffer limits */ server.client_obuf_limits[REDIS_CLIENT_LIMIT_CLASS_NORMAL].hard_limit_bytes = 0; @@ -962,16 +1110,18 @@ void initServerConfig() { populateCommandTable(); server.delCommand = lookupCommandByCString("del"); server.multiCommand = lookupCommandByCString("multi"); + server.lpushCommand = lookupCommandByCString("lpush"); /* Slow log */ server.slowlog_log_slower_than = REDIS_SLOWLOG_LOG_SLOWER_THAN; server.slowlog_max_len = REDIS_SLOWLOG_MAX_LEN; - /* Assert */ + /* Debugging */ server.assert_failed = ""; server.assert_file = ""; server.assert_line = 0; server.bug_report_start = 0; + server.watchdog_period = 0; } /* This function will try to raise the max number of open files accordingly to @@ -997,10 +1147,18 @@ void adjustOpenFilesLimit(void) { /* Set the max number of files if the current limit is not enough * for our needs. */ if (oldlimit < maxfiles) { - limit.rlim_cur = maxfiles; - limit.rlim_max = maxfiles; - if (setrlimit(RLIMIT_NOFILE,&limit) == -1) { - server.maxclients = oldlimit-32; + rlim_t f; + + f = maxfiles; + while(f > oldlimit) { + limit.rlim_cur = f; + limit.rlim_max = f; + if (setrlimit(RLIMIT_NOFILE,&limit) != -1) break; + f -= 128; + } + if (f < oldlimit) f = oldlimit; + if (f != maxfiles) { + server.maxclients = f-32; redisLog(REDIS_WARNING,"Unable to set the max number of files limit to %d (%s), setting the max clients configuration to %d.", (int) maxfiles, strerror(errno), (int) server.maxclients); } else { @@ -1083,7 +1241,13 @@ void initServer() { server.stat_peak_memory = 0; server.stat_fork_time = 0; server.stat_rejected_conn = 0; + memset(server.ops_sec_samples,0,sizeof(server.ops_sec_samples)); + server.ops_sec_idx = 0; + server.ops_sec_last_sample_time = mstime(); + server.ops_sec_last_sample_ops = 0; server.unixtime = time(NULL); + server.lastbgsave_status = REDIS_OK; + server.stop_writes_on_bgsave_err = 1; aeCreateTimeEvent(server.el, 1, serverCron, NULL, NULL); if (server.ipfd > 0 && aeCreateFileEvent(server.el,server.ipfd,AE_READABLE, acceptTcpHandler,NULL) == AE_ERR) oom("creating file event"); @@ -1110,7 +1274,6 @@ void initServer() { server.maxmemory_policy = REDIS_MAXMEMORY_NO_EVICTION; } - if (server.cluster_enabled) clusterInit(); scriptingInit(); slowlogInit(); bioInit(); @@ -1160,6 +1323,43 @@ void resetCommandTableStats(void) { } } +/* ========================== Redis OP Array API ============================ */ + +void redisOpArrayInit(redisOpArray *oa) { + oa->ops = NULL; + oa->numops = 0; +} + +int redisOpArrayAppend(redisOpArray *oa, struct redisCommand *cmd, int dbid, + robj **argv, int argc, int target) +{ + redisOp *op; + + oa->ops = zrealloc(oa->ops,sizeof(redisOp)*(oa->numops+1)); + op = oa->ops+oa->numops; + op->cmd = cmd; + op->dbid = dbid; + op->argv = argv; + op->argc = argc; + op->target = target; + oa->numops++; + return oa->numops; +} + +void redisOpArrayFree(redisOpArray *oa) { + while(oa->numops) { + int j; + redisOp *op; + + oa->numops--; + op = oa->ops+oa->numops; + for (j = 0; j < op->argc; j++) + decrRefCount(op->argv[j]); + zfree(op->argv); + } + zfree(oa->ops); +} + /* ====================== Commands lookup and execution ===================== */ struct redisCommand *lookupCommand(sds name) { @@ -1175,10 +1375,42 @@ struct redisCommand *lookupCommandByCString(char *s) { return cmd; } +/* Propagate the specified command (in the context of the specified database id) + * to AOF, Slaves and Monitors. + * + * flags are an xor between: + * + REDIS_PROPAGATE_NONE (no propagation of command at all) + * + REDIS_PROPAGATE_AOF (propagate into the AOF file if is enabled) + * + REDIS_PROPAGATE_REPL (propagate into the replication link) + */ +void propagate(struct redisCommand *cmd, int dbid, robj **argv, int argc, + int flags) +{ + if (server.aof_state != REDIS_AOF_OFF && flags & REDIS_PROPAGATE_AOF) + feedAppendOnlyFile(cmd,dbid,argv,argc); + if (flags & REDIS_PROPAGATE_REPL && listLength(server.slaves)) + replicationFeedSlaves(server.slaves,dbid,argv,argc); +} + +/* Used inside commands to schedule the propagation of additional commands + * after the current command is propagated to AOF / Replication. */ +void alsoPropagate(struct redisCommand *cmd, int dbid, robj **argv, int argc, + int target) +{ + redisOpArrayAppend(&server.also_propagate,cmd,dbid,argv,argc,target); +} + /* Call() is the core of Redis execution of a command */ void call(redisClient *c, int flags) { long long dirty, start = ustime(), duration; + /* Sent the command to clients in MONITOR mode, only if the commands are + * not geneated from reading an AOF. */ + if (listLength(server.monitors) && !server.loading) + replicationFeedMonitors(c,server.monitors,c->db->id,c->argv,c->argc); + + /* Call the command. */ + redisOpArrayInit(&server.also_propagate); dirty = server.dirty; c->cmd->proc(c); dirty = server.dirty-dirty; @@ -1189,20 +1421,37 @@ void call(redisClient *c, int flags) { if (server.loading && c->flags & REDIS_LUA_CLIENT) flags &= ~(REDIS_CALL_SLOWLOG | REDIS_CALL_STATS); + /* Log the command into the Slow log if needed, and populate the + * per-command statistics that we show in INFO commandstats. */ if (flags & REDIS_CALL_SLOWLOG) slowlogPushEntryIfNeeded(c->argv,c->argc,duration); if (flags & REDIS_CALL_STATS) { c->cmd->microseconds += duration; c->cmd->calls++; } + + /* Propagate the command into the AOF and replication link */ if (flags & REDIS_CALL_PROPAGATE) { - if (server.aof_state != REDIS_AOF_OFF && dirty > 0) - feedAppendOnlyFile(c->cmd,c->db->id,c->argv,c->argc); - if ((dirty > 0 || c->cmd->flags & REDIS_CMD_FORCE_REPLICATION) && - listLength(server.slaves)) - replicationFeedSlaves(server.slaves,c->db->id,c->argv,c->argc); - if (listLength(server.monitors)) - replicationFeedMonitors(server.monitors,c->db->id,c->argv,c->argc); + int flags = REDIS_PROPAGATE_NONE; + + if (c->cmd->flags & REDIS_CMD_FORCE_REPLICATION) + flags |= REDIS_PROPAGATE_REPL; + if (dirty) + flags |= (REDIS_PROPAGATE_REPL | REDIS_PROPAGATE_AOF); + if (flags != REDIS_PROPAGATE_NONE) + propagate(c->cmd,c->db->id,c->argv,c->argc,flags); + } + /* Commands such as LPUSH or BRPOPLPUSH may propagate an additional + * PUSH command. */ + if (server.also_propagate.numops) { + int j; + redisOp *rop; + + for (j = 0; j < server.also_propagate.numops; j++) { + rop = &server.also_propagate.ops[j]; + propagate(rop->cmd, rop->dbid, rop->argv, rop->argc, rop->target); + } + redisOpArrayFree(&server.also_propagate); } server.stat_numcommands++; } @@ -1247,29 +1496,6 @@ int processCommand(redisClient *c) { return REDIS_OK; } - /* If cluster is enabled, redirect here */ - if (server.cluster_enabled && - !(c->cmd->getkeys_proc == NULL && c->cmd->firstkey == 0)) { - int hashslot; - - if (server.cluster.state != REDIS_CLUSTER_OK) { - addReplyError(c,"The cluster is down. Check with CLUSTER INFO for more information"); - return REDIS_OK; - } else { - int ask; - clusterNode *n = getNodeByQuery(c,c->cmd,c->argv,c->argc,&hashslot,&ask); - if (n == NULL) { - addReplyError(c,"Multi keys request invalid in cluster"); - return REDIS_OK; - } else if (n != server.cluster.myself) { - addReplySds(c,sdscatprintf(sdsempty(), - "-%s %d %s:%d\r\n", ask ? "ASK" : "MOVED", - hashslot,n->ip,n->port)); - return REDIS_OK; - } - } - } - /* Handle the maxmemory directive. * * First we try to free some memory if possible (if there are volatile @@ -1278,12 +1504,31 @@ int processCommand(redisClient *c) { if (server.maxmemory) { int retval = freeMemoryIfNeeded(); if ((c->cmd->flags & REDIS_CMD_DENYOOM) && retval == REDIS_ERR) { - addReplyError(c, - "command not allowed when used memory > 'maxmemory'"); + addReply(c, shared.oomerr); return REDIS_OK; } } + /* Don't accept write commands if there are problems persisting on disk. */ + if (server.stop_writes_on_bgsave_err && + server.saveparamslen > 0 + && server.lastbgsave_status == REDIS_ERR && + c->cmd->flags & REDIS_CMD_WRITE) + { + addReply(c, shared.bgsaveerr); + return REDIS_OK; + } + + /* Don't accept wirte commands if this is a read only slave. But + * accept write commands if this is our master. */ + if (server.masterhost && server.repl_slave_ro && + !(c->flags & REDIS_MASTER) && + c->cmd->flags & REDIS_CMD_WRITE) + { + addReply(c, shared.roslaveerr); + return REDIS_OK; + } + /* Only allow SUBSCRIBE and UNSUBSCRIBE in the context of Pub/Sub */ if ((dictSize(c->pubsub_channels) > 0 || listLength(c->pubsub_patterns) > 0) && @@ -1416,6 +1661,17 @@ void echoCommand(redisClient *c) { addReplyBulk(c,c->argv[1]); } +void timeCommand(redisClient *c) { + struct timeval tv; + + /* gettimeofday() can only fail if &tv is a bad addresss so we + * don't check for errors. */ + gettimeofday(&tv,NULL); + addReplyMultiBulkLen(c,2); + addReplyBulkLongLong(c,tv.tv_sec); + addReplyBulkLongLong(c,tv.tv_usec); +} + /* Convert an amount of bytes into a human readable string in the form * of 100B, 2G, 100M, 4K, and so forth. */ void bytesToHuman(char *s, unsigned long long n) { @@ -1442,7 +1698,7 @@ void bytesToHuman(char *s, unsigned long long n) { * on memory corruption problems. */ sds genRedisInfoString(char *section) { sds info = sdsempty(); - time_t uptime = time(NULL)-server.stat_starttime; + time_t uptime = server.unixtime-server.stat_starttime; int j, numcommands; struct rusage self_ru, c_ru; unsigned long lol, bib; @@ -1460,16 +1716,21 @@ sds genRedisInfoString(char *section) { /* Server */ if (allsections || defsections || !strcasecmp(section,"server")) { + struct utsname name; + if (sections++) info = sdscat(info,"\r\n"); + uname(&name); info = sdscatprintf(info, "# Server\r\n" "redis_version:%s\r\n" "redis_git_sha1:%s\r\n" "redis_git_dirty:%d\r\n" + "os:%s %s %s\r\n" "arch_bits:%d\r\n" "multiplexing_api:%s\r\n" "gcc_version:%d.%d.%d\r\n" "process_id:%ld\r\n" + "run_id:%s\r\n" "tcp_port:%d\r\n" "uptime_in_seconds:%ld\r\n" "uptime_in_days:%ld\r\n" @@ -1477,6 +1738,7 @@ sds genRedisInfoString(char *section) { REDIS_VERSION, redisGitSHA1(), strtol(redisGitDirty(),NULL,10) > 0, + name.sysname, name.release, name.machine, server.arch_bits, aeGetApiName(), #ifdef __GNUC__ @@ -1485,6 +1747,7 @@ sds genRedisInfoString(char *section) { 0,0,0, #endif (long) getpid(), + server.runid, server.port, uptime, uptime/(3600*24), @@ -1544,13 +1807,17 @@ sds genRedisInfoString(char *section) { "changes_since_last_save:%lld\r\n" "bgsave_in_progress:%d\r\n" "last_save_time:%ld\r\n" - "bgrewriteaof_in_progress:%d\r\n", + "last_bgsave_status:%s\r\n" + "bgrewriteaof_in_progress:%d\r\n" + "bgrewriteaof_scheduled:%d\r\n", server.loading, server.aof_state != REDIS_AOF_OFF, server.dirty, server.rdb_child_pid != -1, server.lastsave, - server.aof_child_pid != -1); + server.lastbgsave_status == REDIS_OK ? "ok" : "err", + server.aof_child_pid != -1, + server.aof_rewrite_scheduled); if (server.aof_state != REDIS_AOF_OFF) { info = sdscatprintf(info, @@ -1558,12 +1825,14 @@ sds genRedisInfoString(char *section) { "aof_base_size:%lld\r\n" "aof_pending_rewrite:%d\r\n" "aof_buffer_length:%zu\r\n" - "aof_pending_bio_fsync:%llu\r\n", + "aof_pending_bio_fsync:%llu\r\n" + "aof_delayed_fsync:%lu\r\n", (long long) server.aof_current_size, (long long) server.aof_rewrite_base_size, server.aof_rewrite_scheduled, sdslen(server.aof_buf), - bioPendingJobsOfType(REDIS_BIO_AOF_FSYNC)); + bioPendingJobsOfType(REDIS_BIO_AOF_FSYNC), + server.aof_delayed_fsync); } if (server.loading) { @@ -1575,7 +1844,7 @@ sds genRedisInfoString(char *section) { perc = ((double)server.loading_loaded_bytes / server.loading_total_bytes) * 100; - elapsed = time(NULL)-server.loading_start_time; + elapsed = server.unixtime-server.loading_start_time; if (elapsed == 0) { eta = 1; /* A fake 1 second figure if we don't have enough info */ @@ -1605,6 +1874,7 @@ sds genRedisInfoString(char *section) { "# Stats\r\n" "total_connections_received:%lld\r\n" "total_commands_processed:%lld\r\n" + "instantaneous_ops_per_sec:%lld\r\n" "rejected_connections:%lld\r\n" "expired_keys:%lld\r\n" "evicted_keys:%lld\r\n" @@ -1615,6 +1885,7 @@ sds genRedisInfoString(char *section) { "latest_fork_usec:%lld\r\n", server.stat_numconnections, server.stat_numcommands, + getOperationsPerSecond(), server.stat_rejected_conn, server.stat_expiredkeys, server.stat_evictedkeys, @@ -1644,7 +1915,7 @@ sds genRedisInfoString(char *section) { (server.repl_state == REDIS_REPL_CONNECTED) ? "up" : "down", server.master ? - ((int)(time(NULL)-server.master->lastinteraction)) : -1, + ((int)(server.unixtime-server.master->lastinteraction)) : -1, server.repl_state == REDIS_REPL_TRANSFER ); @@ -1653,14 +1924,14 @@ sds genRedisInfoString(char *section) { "master_sync_left_bytes:%ld\r\n" "master_sync_last_io_seconds_ago:%d\r\n" ,(long)server.repl_transfer_left, - (int)(time(NULL)-server.repl_transfer_lastio) + (int)(server.unixtime-server.repl_transfer_lastio) ); } if (server.repl_state != REDIS_REPL_CONNECTED) { info = sdscatprintf(info, "master_link_down_since_seconds:%ld\r\n", - (long)time(NULL)-server.repl_down_since); + (long)server.unixtime-server.repl_down_since); } } info = sdscatprintf(info, @@ -1730,15 +2001,6 @@ sds genRedisInfoString(char *section) { } } - /* Clusetr */ - if (allsections || defsections || !strcasecmp(section,"cluster")) { - if (sections++) info = sdscat(info,"\r\n"); - info = sdscatprintf(info, - "# Cluster\r\n" - "cluster_enabled:%d\r\n", - server.cluster_enabled); - } - /* Key space */ if (allsections || defsections || !strcasecmp(section,"keyspace")) { if (sections++) info = sdscat(info,"\r\n"); @@ -1832,7 +2094,6 @@ int freeMemoryIfNeeded(void) { /* Compute how much memory we need to free. */ mem_tofree = mem_used - server.maxmemory; - printf("USED: %zu, TOFREE: %zu\n", mem_used, mem_tofree); mem_freed = 0; while (mem_freed < mem_tofree) { int j, k, keys_freed = 0; @@ -1923,7 +2184,6 @@ int freeMemoryIfNeeded(void) { delta = (long long) zmalloc_used_memory(); dbDelete(db,keyobj); delta -= (long long) zmalloc_used_memory(); - // printf("%lld\n",delta); mem_freed += delta; server.stat_evictedkeys++; decrRefCount(keyobj); @@ -1933,27 +2193,7 @@ int freeMemoryIfNeeded(void) { * start spending so much time here that is impossible to * deliver data to the slaves fast enough, so we force the * transmission here inside the loop. */ - if (slaves) { - listIter li; - listNode *ln; - - listRewind(server.slaves,&li); - while((ln = listNext(&li))) { - redisClient *slave = listNodeValue(ln); - int events; - - events = aeGetFileEvents(server.el,slave->fd); - printf("EVENTS: %d\n", events); - if (events & AE_WRITABLE && - slave->replstate == REDIS_REPL_ONLINE && - listLength(slave->reply)) - { - printf("SLAVE %d -> %d\n", - slave->fd, (int) listLength(slave->reply)); - sendReplyToClient(server.el,slave->fd,slave,0); - } - } - } + if (slaves) flushSlavesOutputBuffers(); } } if (!keys_freed) return REDIS_ERR; /* nothing to free... */ @@ -2012,8 +2252,12 @@ void daemonize(void) { } void version() { - printf("Redis server version %s (%s:%d)\n", REDIS_VERSION, - redisGitSHA1(), atoi(redisGitDirty()) > 0); + printf("Redis server v=%s sha=%s:%d malloc=%s bits=%d\n", + REDIS_VERSION, + redisGitSHA1(), + atoi(redisGitDirty()) > 0, + ZMALLOC_LIB, + sizeof(long) == 4 ? 32 : 64); exit(0); } @@ -2021,7 +2265,8 @@ void usage() { fprintf(stderr,"Usage: ./redis-server [/path/to/redis.conf] [options]\n"); fprintf(stderr," ./redis-server - (read config from stdin)\n"); fprintf(stderr," ./redis-server -v or --version\n"); - fprintf(stderr," ./redis-server -h or --help\n\n"); + fprintf(stderr," ./redis-server -h or --help\n"); + fprintf(stderr," ./redis-server --test-memory \n\n"); fprintf(stderr,"Examples:\n"); fprintf(stderr," ./redis-server (run the server with default conf)\n"); fprintf(stderr," ./redis-server /etc/redis/6379.conf\n"); @@ -2040,7 +2285,7 @@ void redisAsciiArt(void) { redisGitSHA1(), strtol(redisGitDirty(),NULL,10) > 0, (sizeof(long) == 8) ? "64" : "32", - server.cluster_enabled ? "cluster" : "stand alone", + "stand alone", server.port, (long) getpid() ); @@ -2051,21 +2296,30 @@ void redisAsciiArt(void) { static void sigtermHandler(int sig) { REDIS_NOTUSED(sig); - redisLog(REDIS_WARNING,"Received SIGTERM, scheduling shutdown..."); + redisLogFromHandler(REDIS_WARNING,"Received SIGTERM, scheduling shutdown..."); server.shutdown_asap = 1; } void setupSignalHandlers(void) { struct sigaction act; + stack_t stack; + + stack.ss_sp = altstack; + stack.ss_flags = 0; + stack.ss_size = SIGSTKSZ; + + sigaltstack(&stack, NULL); /* When the SA_SIGINFO flag is set in sa_flags then sa_sigaction is used. * Otherwise, sa_handler is used. */ sigemptyset(&act.sa_mask); - act.sa_flags = SA_NODEFER | SA_ONSTACK | SA_RESETHAND; + act.sa_flags = 0; act.sa_handler = sigtermHandler; sigaction(SIGTERM, &act, NULL); #ifdef HAVE_BACKTRACE + /* Use alternate stack so we don't clobber stack in case of segv, or when we run out of stack .. + * also resethand & nodefer so we can get interrupted (and killed) if we cause SEGV during SEGV handler */ sigemptyset(&act.sa_mask); act.sa_flags = SA_NODEFER | SA_ONSTACK | SA_RESETHAND | SA_SIGINFO; act.sa_sigaction = sigsegvHandler; @@ -2077,6 +2331,8 @@ void setupSignalHandlers(void) { return; } +void memtest(size_t megabytes, int passes); + int main(int argc, char **argv) { long long start; struct timeval tv; @@ -2098,6 +2354,17 @@ int main(int argc, char **argv) { strcmp(argv[1], "--version") == 0) version(); if (strcmp(argv[1], "--help") == 0 || strcmp(argv[1], "-h") == 0) usage(); + if (strcmp(argv[1], "--test-memory") == 0) { + if (argc == 3) { + memtest(atoi(argv[2]),50); + exit(0); + } else { + fprintf(stderr,"Please specify the amount of memory to test in megabytes.\n"); + fprintf(stderr,"Example: ./redis-server --test-memory 4096\n\n"); + exit(1); + } + } + /* First argument is the config file name? */ if (argv[j][0] != '-' || argv[j][1] != '-') configfile = argv[j++];