X-Git-Url: https://git.saurik.com/redis.git/blobdiff_plain/8a623a98c31bda69c05f2c7c94bad2d131060b02..419e1cca74102440d5ce174fbbf36d81a9b8956a:/src/rdb.c diff --git a/src/rdb.c b/src/rdb.c index f89ce270..02317fda 100644 --- a/src/rdb.c +++ b/src/rdb.c @@ -9,15 +9,22 @@ #include #include +/* Convenience wrapper around fwrite, that returns the number of bytes written + * to the file instead of the number of objects (see fwrite(3)) and -1 in the + * case of an error. It also supports a NULL *fp to skip writing altogether + * instead of writing to /dev/null. */ +static int rdbWriteRaw(FILE *fp, void *p, size_t len) { + if (fp != NULL && fwrite(p,len,1,fp) == 0) return -1; + return len; +} + int rdbSaveType(FILE *fp, unsigned char type) { - if (fwrite(&type,1,1,fp) == 0) return -1; - return 1; /* bytes written */ + return rdbWriteRaw(fp,&type,1); } int rdbSaveTime(FILE *fp, time_t t) { int32_t t32 = (int32_t) t; - if (fwrite(&t32,4,1,fp) == 0) return -1; - return 4; /* bytes written */ + return rdbWriteRaw(fp,&t32,4); } /* check rdbLoadLen() comments for more info */ @@ -28,20 +35,20 @@ int rdbSaveLen(FILE *fp, uint32_t len) { if (len < (1<<6)) { /* Save a 6 bit len */ buf[0] = (len&0xFF)|(REDIS_RDB_6BITLEN<<6); - if (fwrite(buf,1,1,fp) == 0) return -1; + if (rdbWriteRaw(fp,buf,1) == -1) return -1; nwritten = 1; } else if (len < (1<<14)) { /* Save a 14 bit len */ buf[0] = ((len>>8)&0xFF)|(REDIS_RDB_14BITLEN<<6); buf[1] = len&0xFF; - if (fwrite(buf,2,1,fp) == 0) return -1; + if (rdbWriteRaw(fp,buf,2) == -1) return -1; nwritten = 2; } else { /* Save a 32 bit len */ buf[0] = (REDIS_RDB_32BITLEN<<6); - if (fwrite(buf,1,1,fp) == 0) return -1; + if (rdbWriteRaw(fp,buf,1) == -1) return -1; len = htonl(len); - if (fwrite(&len,4,1,fp) == 0) return -1; + if (rdbWriteRaw(fp,&len,4) == -1) return -1; nwritten = 1+4; } return nwritten; @@ -111,8 +118,8 @@ int rdbSaveLzfStringObject(FILE *fp, unsigned char *s, size_t len) { } /* Data compressed! Let's save it on disk */ byte = (REDIS_RDB_ENCVAL<<6)|REDIS_RDB_ENC_LZF; - if (fwrite(&byte,1,1,fp) == 0) goto writeerr; - nwritten += 1; + if ((n = rdbWriteRaw(fp,&byte,1)) == -1) goto writeerr; + nwritten += n; if ((n = rdbSaveLen(fp,comprlen)) == -1) goto writeerr; nwritten += n; @@ -120,8 +127,8 @@ int rdbSaveLzfStringObject(FILE *fp, unsigned char *s, size_t len) { if ((n = rdbSaveLen(fp,len)) == -1) goto writeerr; nwritten += n; - if (fwrite(out,comprlen,1,fp) == 0) goto writeerr; - nwritten += comprlen; + if ((n = rdbWriteRaw(fp,out,comprlen)) == -1) goto writeerr; + nwritten += n; zfree(out); return nwritten; @@ -141,7 +148,7 @@ int rdbSaveRawString(FILE *fp, unsigned char *s, size_t len) { if (len <= 11) { unsigned char buf[5]; if ((enclen = rdbTryIntegerEncoding((char*)s,len,buf)) > 0) { - if (fwrite(buf,enclen,1,fp) == 0) return -1; + if (rdbWriteRaw(fp,buf,enclen) == -1) return -1; return enclen; } } @@ -159,7 +166,7 @@ int rdbSaveRawString(FILE *fp, unsigned char *s, size_t len) { if ((n = rdbSaveLen(fp,len)) == -1) return -1; nwritten += n; if (len > 0) { - if (fwrite(s,len,1,fp) == 0) return -1; + if (rdbWriteRaw(fp,s,len) == -1) return -1; nwritten += len; } return nwritten; @@ -171,16 +178,15 @@ int rdbSaveLongLongAsStringObject(FILE *fp, long long value) { int n, nwritten = 0; int enclen = rdbEncodeInteger(value,buf); if (enclen > 0) { - if (fwrite(buf,enclen,1,fp) == 0) return -1; - nwritten = enclen; + return rdbWriteRaw(fp,buf,enclen); } else { /* Encode as string */ enclen = ll2string((char*)buf,32,value); redisAssert(enclen < 32); if ((n = rdbSaveLen(fp,enclen)) == -1) return -1; nwritten += n; - if (fwrite(buf,enclen,1,fp) == 0) return -1; - nwritten += enclen; + if ((n = rdbWriteRaw(fp,buf,enclen)) == -1) return -1; + nwritten += n; } return nwritten; } @@ -236,8 +242,7 @@ int rdbSaveDoubleValue(FILE *fp, double val) { buf[0] = strlen((char*)buf+1); len = buf[0]+1; } - if (fwrite(buf,len,1,fp) == 0) return -1; - return len; + return rdbWriteRaw(fp,buf,len); } /* Save a Redis object. */ @@ -384,21 +389,31 @@ int rdbSaveObject(FILE *fp, robj *o) { * the rdbSaveObject() function. Currently we use a trick to get * this length with very little changes to the code. In the future * we could switch to a faster solution. */ -off_t rdbSavedObjectLen(robj *o, FILE *fp) { - int nwritten; - if (fp == NULL) fp = server.devnull; - rewind(fp); - - /* Determining the saved length of an object should never return -1 */ - redisAssert((nwritten = rdbSaveObject(fp,o)) != -1); - return nwritten; +off_t rdbSavedObjectLen(robj *o) { + int len = rdbSaveObject(NULL,o); + redisAssert(len != -1); + return len; } -/* Return the number of pages required to save this object in the swap file */ -off_t rdbSavedObjectPages(robj *o, FILE *fp) { - off_t bytes = rdbSavedObjectLen(o,fp); - - return (bytes+(server.vm_page_size-1))/server.vm_page_size; +/* Save a key-value pair, with expire time, type, key, value. + * On error -1 is returned. + * On success if the key was actaully saved 1 is returned, otherwise 0 + * is returned (the key was already expired). */ +int rdbSaveKeyValuePair(FILE *fp, robj *key, robj *val, + time_t expiretime, time_t now) +{ + /* Save the expire time */ + if (expiretime != -1) { + /* If this key is already expired skip it */ + if (expiretime < now) return 0; + if (rdbSaveType(fp,REDIS_EXPIRETIME) == -1) return -1; + if (rdbSaveTime(fp,expiretime) == -1) return -1; + } + /* Save type, key, value */ + if (rdbSaveType(fp,val->type) == -1) return -1; + if (rdbSaveStringObject(fp,key) == -1) return -1; + if (rdbSaveObject(fp,val) == -1) return -1; + return 1; } /* Save the DB on disk. Return REDIS_ERR on error, REDIS_OK on success */ @@ -410,16 +425,16 @@ int rdbSave(char *filename) { int j; time_t now = time(NULL); - /* Wait for I/O therads to terminate, just in case this is a - * foreground-saving, to avoid seeking the swap file descriptor at the - * same time. */ - if (server.vm_enabled) - waitEmptyIOJobsQueue(); + if (server.ds_enabled) { + cacheForcePointInTime(); + return dsRdbSave(filename); + } snprintf(tmpfile,256,"temp-%d.rdb", (int) getpid()); fp = fopen(tmpfile,"w"); if (!fp) { - redisLog(REDIS_WARNING, "Failed saving the DB: %s", strerror(errno)); + redisLog(REDIS_WARNING, "Failed opening .rdb for saving: %s", + strerror(errno)); return REDIS_ERR; } if (fwrite("REDIS0001",9,1,fp) == 0) goto werr; @@ -441,38 +456,11 @@ int rdbSave(char *filename) { while((de = dictNext(di)) != NULL) { sds keystr = dictGetEntryKey(de); robj key, *o = dictGetEntryVal(de); - time_t expiretime; + time_t expire; initStaticStringObject(key,keystr); - expiretime = getExpire(db,&key); - - /* Save the expire time */ - if (expiretime != -1) { - /* If this key is already expired skip it */ - if (expiretime < now) continue; - if (rdbSaveType(fp,REDIS_EXPIRETIME) == -1) goto werr; - if (rdbSaveTime(fp,expiretime) == -1) goto werr; - } - /* Save the key and associated value. This requires special - * handling if the value is swapped out. */ - if (!server.vm_enabled || o->storage == REDIS_VM_MEMORY || - o->storage == REDIS_VM_SWAPPING) { - /* Save type, key, value */ - if (rdbSaveType(fp,o->type) == -1) goto werr; - if (rdbSaveStringObject(fp,&key) == -1) goto werr; - if (rdbSaveObject(fp,o) == -1) goto werr; - } else { - /* REDIS_VM_SWAPPED or REDIS_VM_LOADING */ - robj *po; - /* Get a preview of the object in memory */ - po = vmPreviewObject(o); - /* Save type, key, value */ - if (rdbSaveType(fp,po->type) == -1) goto werr; - if (rdbSaveStringObject(fp,&key) == -1) goto werr; - if (rdbSaveObject(fp,po) == -1) goto werr; - /* Remove the loaded object from memory */ - decrRefCount(po); - } + expire = getExpire(db,&key); + if (rdbSaveKeyValuePair(fp,&key,o,expire,now) == -1) goto werr; } dictReleaseIterator(di); } @@ -507,19 +495,24 @@ werr: int rdbSaveBackground(char *filename) { pid_t childpid; - if (server.bgsavechildpid != -1) return REDIS_ERR; - if (server.vm_enabled) waitEmptyIOJobsQueue(); + if (server.bgsavechildpid != -1 || + server.bgsavethread != (pthread_t) -1) return REDIS_ERR; + server.dirty_before_bgsave = server.dirty; + + if (server.ds_enabled) { + cacheForcePointInTime(); + return dsRdbSaveBackground(filename); + } + if ((childpid = fork()) == 0) { + int retval; + /* Child */ - if (server.vm_enabled) vmReopenSwapFile(); if (server.ipfd > 0) close(server.ipfd); if (server.sofd > 0) close(server.sofd); - if (rdbSave(filename) == REDIS_OK) { - _exit(0); - } else { - _exit(1); - } + retval = rdbSave(filename); + _exit((retval == REDIS_OK) ? 0 : 1); } else { /* Parent */ if (childpid == -1) { @@ -874,7 +867,6 @@ int rdbLoad(char *filename) { FILE *fp; uint32_t dbid; int type, retval, rdbver; - int swap_all_values = 0; redisDb *db = server.db+0; char buf[1024]; time_t expiretime, now = time(NULL); @@ -899,8 +891,6 @@ int rdbLoad(char *filename) { startLoading(fp); while(1) { robj *key, *val; - int force_swapout; - expiretime = -1; /* Serve the clients from time to time */ @@ -947,44 +937,7 @@ int rdbLoad(char *filename) { /* Set the expire time if needed */ if (expiretime != -1) setExpire(db,key,expiretime); - /* Handle swapping while loading big datasets when VM is on */ - - /* If we detecter we are hopeless about fitting something in memory - * we just swap every new key on disk. Directly... - * Note that's important to check for this condition before resorting - * to random sampling, otherwise we may try to swap already - * swapped keys. */ - if (swap_all_values) { - dictEntry *de = dictFind(db->dict,key->ptr); - - /* de may be NULL since the key already expired */ - if (de) { - vmpointer *vp; - val = dictGetEntryVal(de); - - if (val->refcount == 1 && - (vp = vmSwapObjectBlocking(val)) != NULL) - dictGetEntryVal(de) = vp; - } - decrRefCount(key); - continue; - } decrRefCount(key); - - /* Flush data on disk once 32 MB of additional RAM are used... */ - force_swapout = 0; - if ((zmalloc_used_memory() - server.vm_max_memory) > 1024*1024*32) - force_swapout = 1; - - /* If we have still some hope of having some value fitting memory - * then we try random sampling. */ - if (!swap_all_values && server.vm_enabled && force_swapout) { - while (zmalloc_used_memory() > server.vm_max_memory) { - if (vmSwapOneObjectBlocking() == REDIS_ERR) break; - } - if (zmalloc_used_memory() > server.vm_max_memory) - swap_all_values = 1; /* We are already using too much mem */ - } } fclose(fp); stopLoading(); @@ -997,10 +950,7 @@ eoferr: /* unexpected end of file is handled here with a fatal exit */ } /* A background saving child (BGSAVE) terminated its work. Handle this. */ -void backgroundSaveDoneHandler(int statloc) { - int exitcode = WEXITSTATUS(statloc); - int bysignal = WIFSIGNALED(statloc); - +void backgroundSaveDoneHandler(int exitcode, int bysignal) { if (!bysignal && exitcode == 0) { redisLog(REDIS_NOTICE, "Background saving terminated with success"); @@ -1010,11 +960,37 @@ void backgroundSaveDoneHandler(int statloc) { redisLog(REDIS_WARNING, "Background saving error"); } else { redisLog(REDIS_WARNING, - "Background saving terminated by signal %d", WTERMSIG(statloc)); + "Background saving terminated by signal %d", bysignal); rdbRemoveTempFile(server.bgsavechildpid); } server.bgsavechildpid = -1; + server.bgsavethread = (pthread_t) -1; + server.bgsavethread_state = REDIS_BGSAVE_THREAD_UNACTIVE; /* Possibly there are slaves waiting for a BGSAVE in order to be served * (the first stage of SYNC is a bulk transfer of dump.rdb) */ updateSlavesWaitingBgsave(exitcode == 0 ? REDIS_OK : REDIS_ERR); } + +void saveCommand(redisClient *c) { + if (server.bgsavechildpid != -1 || server.bgsavethread != (pthread_t)-1) { + addReplyError(c,"Background save already in progress"); + return; + } + if (rdbSave(server.dbfilename) == REDIS_OK) { + addReply(c,shared.ok); + } else { + addReply(c,shared.err); + } +} + +void bgsaveCommand(redisClient *c) { + if (server.bgsavechildpid != -1 || server.bgsavethread != (pthread_t)-1) { + addReplyError(c,"Background save already in progress"); + return; + } + if (rdbSaveBackground(server.dbfilename) == REDIS_OK) { + addReplyStatus(c,"Background saving started"); + } else { + addReply(c,shared.err); + } +}