+#include "redis.h"
+#include "lzf.h" /* LZF compression library */
+#include "zipmap.h"
+
#include <math.h>
#include <sys/types.h>
#include <sys/time.h>
#include <sys/wait.h>
#include <arpa/inet.h>
#include <sys/stat.h>
-#include "rdb.h"
-#include "lzf.h" /* LZF compression library */
static int rdbWriteRaw(rio *rdb, void *p, size_t len) {
- if (rioWrite(rdb,p,len) == 0)
+ if (rdb && rioWrite(rdb,p,len) == 0)
return -1;
- return 1;
+ return len;
}
int rdbSaveType(rio *rdb, unsigned char type) {
return (time_t)t32;
}
+int rdbSaveMillisecondTime(rio *rdb, long long t) {
+ int64_t t64 = (int64_t) t;
+ return rdbWriteRaw(rdb,&t64,8);
+}
+
+long long rdbLoadMillisecondTime(rio *rdb) {
+ int64_t t64;
+ if (rioRead(rdb,&t64,8) == 0) return -1;
+ return (long long)t64;
+}
+
/* Saves an encoded length. The first two bits in the first byte are used to
* hold the encoding type. See the REDIS_RDB_* definitions for more information
* on the types of encoding. */
/* Try LZF compression - under 20 bytes it's unable to compress even
* aaaaaaaaaaaaaaaaaa so skip it */
- if (server.rdbcompression && len > 20) {
+ if (server.rdb_compression && len > 20) {
n = rdbSaveLzfStringObject(rdb,s,len);
if (n == -1) return -1;
if (n > 0) return n;
if (obj->encoding == REDIS_ENCODING_INT) {
return rdbSaveLongLongAsStringObject(rdb,(long)obj->ptr);
} else {
- redisAssert(obj->encoding == REDIS_ENCODING_RAW);
+ redisAssertWithInfo(NULL,obj,obj->encoding == REDIS_ENCODING_RAW);
return rdbSaveRawString(rdb,obj->ptr,sdslen(obj->ptr));
}
}
else
redisPanic("Unknown sorted set encoding");
case REDIS_HASH:
- if (o->encoding == REDIS_ENCODING_ZIPMAP)
- return rdbSaveType(rdb,REDIS_RDB_TYPE_HASH_ZIPMAP);
+ if (o->encoding == REDIS_ENCODING_ZIPLIST)
+ return rdbSaveType(rdb,REDIS_RDB_TYPE_HASH_ZIPLIST);
else if (o->encoding == REDIS_ENCODING_HT)
return rdbSaveType(rdb,REDIS_RDB_TYPE_HASH);
else
nwritten += n;
while((de = dictNext(di)) != NULL) {
- robj *eleobj = dictGetEntryKey(de);
+ robj *eleobj = dictGetKey(de);
if ((n = rdbSaveStringObject(rdb,eleobj)) == -1) return -1;
nwritten += n;
}
nwritten += n;
while((de = dictNext(di)) != NULL) {
- robj *eleobj = dictGetEntryKey(de);
- double *score = dictGetEntryVal(de);
+ robj *eleobj = dictGetKey(de);
+ double *score = dictGetVal(de);
if ((n = rdbSaveStringObject(rdb,eleobj)) == -1) return -1;
nwritten += n;
}
} else if (o->type == REDIS_HASH) {
/* Save a hash value */
- if (o->encoding == REDIS_ENCODING_ZIPMAP) {
- size_t l = zipmapBlobLen((unsigned char*)o->ptr);
+ if (o->encoding == REDIS_ENCODING_ZIPLIST) {
+ size_t l = ziplistBlobLen((unsigned char*)o->ptr);
if ((n = rdbSaveRawString(rdb,o->ptr,l)) == -1) return -1;
nwritten += n;
- } else {
+
+ } else if (o->encoding == REDIS_ENCODING_HT) {
dictIterator *di = dictGetIterator(o->ptr);
dictEntry *de;
nwritten += n;
while((de = dictNext(di)) != NULL) {
- robj *key = dictGetEntryKey(de);
- robj *val = dictGetEntryVal(de);
+ robj *key = dictGetKey(de);
+ robj *val = dictGetVal(de);
if ((n = rdbSaveStringObject(rdb,key)) == -1) return -1;
nwritten += n;
nwritten += n;
}
dictReleaseIterator(di);
+
+ } else {
+ redisPanic("Unknown hash encoding");
}
+
} else {
redisPanic("Unknown object type");
}
* we could switch to a faster solution. */
off_t rdbSavedObjectLen(robj *o) {
int len = rdbSaveObject(NULL,o);
- redisAssert(len != -1);
+ redisAssertWithInfo(NULL,o,len != -1);
return len;
}
* On success if the key was actaully saved 1 is returned, otherwise 0
* is returned (the key was already expired). */
int rdbSaveKeyValuePair(rio *rdb, robj *key, robj *val,
- time_t expiretime, time_t now)
+ long long expiretime, long long now)
{
/* Save the expire time */
if (expiretime != -1) {
/* If this key is already expired skip it */
if (expiretime < now) return 0;
- if (rdbSaveType(rdb,REDIS_RDB_OPCODE_EXPIRETIME) == -1) return -1;
- if (rdbSaveTime(rdb,expiretime) == -1) return -1;
+ if (rdbSaveType(rdb,REDIS_RDB_OPCODE_EXPIRETIME_MS) == -1) return -1;
+ if (rdbSaveMillisecondTime(rdb,expiretime) == -1) return -1;
}
/* Save type, key, value */
dictEntry *de;
char tmpfile[256];
int j;
- time_t now = time(NULL);
+ long long now = mstime();
FILE *fp;
rio rdb;
- if (server.ds_enabled) {
- cacheForcePointInTime();
- return dsRdbSave(filename);
- }
-
snprintf(tmpfile,256,"temp-%d.rdb", (int) getpid());
fp = fopen(tmpfile,"w");
if (!fp) {
return REDIS_ERR;
}
- rdb = rioInitWithFile(fp);
- if (rdbWriteRaw(&rdb,"REDIS0002",9) == -1) goto werr;
+ rioInitWithFile(&rdb,fp);
+ if (rdbWriteRaw(&rdb,"REDIS0004",9) == -1) goto werr;
for (j = 0; j < server.dbnum; j++) {
redisDb *db = server.db+j;
dict *d = db->dict;
if (dictSize(d) == 0) continue;
- di = dictGetIterator(d);
+ di = dictGetSafeIterator(d);
if (!di) {
fclose(fp);
return REDIS_ERR;
/* Iterate this DB writing every entry */
while((de = dictNext(di)) != NULL) {
- sds keystr = dictGetEntryKey(de);
- robj key, *o = dictGetEntryVal(de);
- time_t expire;
+ sds keystr = dictGetKey(de);
+ robj key, *o = dictGetVal(de);
+ long long expire;
initStaticStringObject(key,keystr);
expire = getExpire(db,&key);
redisLog(REDIS_NOTICE,"DB saved on disk");
server.dirty = 0;
server.lastsave = time(NULL);
+ server.lastbgsave_status = REDIS_OK;
return REDIS_OK;
werr:
int rdbSaveBackground(char *filename) {
pid_t childpid;
+ long long start;
- if (server.bgsavechildpid != -1 ||
- server.bgsavethread != (pthread_t) -1) return REDIS_ERR;
+ if (server.rdb_child_pid != -1) return REDIS_ERR;
server.dirty_before_bgsave = server.dirty;
- if (server.ds_enabled) {
- cacheForcePointInTime();
- return dsRdbSaveBackground(filename);
- }
-
+ start = ustime();
if ((childpid = fork()) == 0) {
int retval;
_exit((retval == REDIS_OK) ? 0 : 1);
} else {
/* Parent */
+ server.stat_fork_time = ustime()-start;
if (childpid == -1) {
redisLog(REDIS_WARNING,"Can't save in background: fork: %s",
strerror(errno));
return REDIS_ERR;
}
redisLog(REDIS_NOTICE,"Background saving started by pid %d",childpid);
- server.bgsavechildpid = childpid;
+ server.rdb_child_pid = childpid;
updateDictResizePolicy();
return REDIS_OK;
}
maxelelen <= server.zset_max_ziplist_value)
zsetConvert(o,REDIS_ENCODING_ZIPLIST);
} else if (rdbtype == REDIS_RDB_TYPE_HASH) {
- size_t hashlen;
+ size_t len;
+ int ret;
+
+ len = rdbLoadLen(rdb, NULL);
+ if (len == REDIS_RDB_LENERR) return NULL;
- if ((hashlen = rdbLoadLen(rdb,NULL)) == REDIS_RDB_LENERR) return NULL;
o = createHashObject();
+
/* Too many entries? Use an hash table. */
- if (hashlen > server.hash_max_zipmap_entries)
- convertToRealHash(o);
- /* Load every key/value, then set it into the zipmap or hash
- * table, as needed. */
- while(hashlen--) {
- robj *key, *val;
-
- if ((key = rdbLoadEncodedStringObject(rdb)) == NULL) return NULL;
- if ((val = rdbLoadEncodedStringObject(rdb)) == NULL) return NULL;
- /* If we are using a zipmap and there are too big values
- * the object is converted to real hash table encoding. */
- if (o->encoding != REDIS_ENCODING_HT &&
- ((key->encoding == REDIS_ENCODING_RAW &&
- sdslen(key->ptr) > server.hash_max_zipmap_value) ||
- (val->encoding == REDIS_ENCODING_RAW &&
- sdslen(val->ptr) > server.hash_max_zipmap_value)))
+ if (len > server.hash_max_ziplist_entries)
+ hashTypeConvert(o, REDIS_ENCODING_HT);
+
+ /* Load every field and value into the ziplist */
+ while (o->encoding == REDIS_ENCODING_ZIPLIST && len > 0) {
+ robj *field, *value;
+
+ len--;
+ /* Load raw strings */
+ field = rdbLoadStringObject(rdb);
+ if (field == NULL) return NULL;
+ redisAssert(field->encoding == REDIS_ENCODING_RAW);
+ value = rdbLoadStringObject(rdb);
+ if (value == NULL) return NULL;
+ redisAssert(field->encoding == REDIS_ENCODING_RAW);
+
+ /* Add pair to ziplist */
+ o->ptr = ziplistPush(o->ptr, field->ptr, sdslen(field->ptr), ZIPLIST_TAIL);
+ o->ptr = ziplistPush(o->ptr, value->ptr, sdslen(value->ptr), ZIPLIST_TAIL);
+
+ /* Convert to hash table if size threshold is exceeded */
+ if (sdslen(field->ptr) > server.hash_max_ziplist_value ||
+ sdslen(value->ptr) > server.hash_max_ziplist_value)
{
- convertToRealHash(o);
+ hashTypeConvert(o, REDIS_ENCODING_HT);
+ break;
}
+ }
- if (o->encoding == REDIS_ENCODING_ZIPMAP) {
- unsigned char *zm = o->ptr;
- robj *deckey, *decval;
-
- /* We need raw string objects to add them to the zipmap */
- deckey = getDecodedObject(key);
- decval = getDecodedObject(val);
- zm = zipmapSet(zm,deckey->ptr,sdslen(deckey->ptr),
- decval->ptr,sdslen(decval->ptr),NULL);
- o->ptr = zm;
- decrRefCount(deckey);
- decrRefCount(decval);
- decrRefCount(key);
- decrRefCount(val);
- } else {
- key = tryObjectEncoding(key);
- val = tryObjectEncoding(val);
- dictAdd((dict*)o->ptr,key,val);
- }
+ /* Load remaining fields and values into the hash table */
+ while (o->encoding == REDIS_ENCODING_HT && len > 0) {
+ robj *field, *value;
+
+ len--;
+ /* Load encoded strings */
+ field = rdbLoadEncodedStringObject(rdb);
+ if (field == NULL) return NULL;
+ value = rdbLoadEncodedStringObject(rdb);
+ if (value == NULL) return NULL;
+
+ field = tryObjectEncoding(field);
+ value = tryObjectEncoding(value);
+
+ /* Add pair to hash table */
+ ret = dictAdd((dict*)o->ptr, field, value);
+ redisAssert(ret == REDIS_OK);
}
+
+ /* All pairs should be read by now */
+ redisAssert(len == 0);
+
} else if (rdbtype == REDIS_RDB_TYPE_HASH_ZIPMAP ||
rdbtype == REDIS_RDB_TYPE_LIST_ZIPLIST ||
rdbtype == REDIS_RDB_TYPE_SET_INTSET ||
- rdbtype == REDIS_RDB_TYPE_ZSET_ZIPLIST)
+ rdbtype == REDIS_RDB_TYPE_ZSET_ZIPLIST ||
+ rdbtype == REDIS_RDB_TYPE_HASH_ZIPLIST)
{
robj *aux = rdbLoadStringObject(rdb);
* converted. */
switch(rdbtype) {
case REDIS_RDB_TYPE_HASH_ZIPMAP:
- o->type = REDIS_HASH;
- o->encoding = REDIS_ENCODING_ZIPMAP;
- if (zipmapLen(o->ptr) > server.hash_max_zipmap_entries)
- convertToRealHash(o);
+ /* Convert to ziplist encoded hash. This must be deprecated
+ * when loading dumps created by Redis 2.4 gets deprecated. */
+ {
+ unsigned char *zl = ziplistNew();
+ unsigned char *zi = zipmapRewind(o->ptr);
+ unsigned char *fstr, *vstr;
+ unsigned int flen, vlen;
+ unsigned int maxlen = 0;
+
+ while ((zi = zipmapNext(zi, &fstr, &flen, &vstr, &vlen)) != NULL) {
+ if (flen > maxlen) maxlen = flen;
+ if (vlen > maxlen) maxlen = vlen;
+ zl = ziplistPush(zl, fstr, flen, ZIPLIST_TAIL);
+ zl = ziplistPush(zl, vstr, vlen, ZIPLIST_TAIL);
+ }
+
+ zfree(o->ptr);
+ o->ptr = zl;
+ o->type = REDIS_HASH;
+ o->encoding = REDIS_ENCODING_ZIPLIST;
+
+ if (hashTypeLength(o) > server.hash_max_ziplist_entries ||
+ maxlen > server.hash_max_ziplist_value)
+ {
+ hashTypeConvert(o, REDIS_ENCODING_HT);
+ }
+ }
break;
case REDIS_RDB_TYPE_LIST_ZIPLIST:
o->type = REDIS_LIST;
if (zsetLength(o) > server.zset_max_ziplist_entries)
zsetConvert(o,REDIS_ENCODING_SKIPLIST);
break;
+ case REDIS_RDB_TYPE_HASH_ZIPLIST:
+ o->type = REDIS_HASH;
+ o->encoding = REDIS_ENCODING_ZIPLIST;
+ if (hashTypeLength(o) > server.hash_max_ziplist_entries)
+ hashTypeConvert(o, REDIS_ENCODING_HT);
+ break;
default:
redisPanic("Unknown encoding");
break;
int rdbLoad(char *filename) {
uint32_t dbid;
- int type, retval, rdbver;
+ int type, rdbver;
redisDb *db = server.db+0;
char buf[1024];
- time_t expiretime, now = time(NULL);
+ long long expiretime, now = mstime();
long loops = 0;
FILE *fp;
rio rdb;
fp = fopen(filename,"r");
- if (!fp) return REDIS_ERR;
- rdb = rioInitWithFile(fp);
+ if (!fp) {
+ errno = ENOENT;
+ return REDIS_ERR;
+ }
+ rioInitWithFile(&rdb,fp);
if (rioRead(&rdb,buf,9) == 0) goto eoferr;
buf[9] = '\0';
if (memcmp(buf,"REDIS",5) != 0) {
fclose(fp);
redisLog(REDIS_WARNING,"Wrong signature trying to load DB from file");
+ errno = EINVAL;
return REDIS_ERR;
}
rdbver = atoi(buf+5);
- if (rdbver < 1 || rdbver > 2) {
+ if (rdbver < 1 || rdbver > 4) {
fclose(fp);
redisLog(REDIS_WARNING,"Can't handle RDB format version %d",rdbver);
+ errno = EINVAL;
return REDIS_ERR;
}
if ((expiretime = rdbLoadTime(&rdb)) == -1) goto eoferr;
/* We read the time so we need to read the object type again. */
if ((type = rdbLoadType(&rdb)) == -1) goto eoferr;
+ /* the EXPIRETIME opcode specifies time in seconds, so convert
+ * into milliesconds. */
+ expiretime *= 1000;
+ } else if (type == REDIS_RDB_OPCODE_EXPIRETIME_MS) {
+ /* Milliseconds precision expire times introduced with RDB
+ * version 3. */
+ if ((expiretime = rdbLoadMillisecondTime(&rdb)) == -1) goto eoferr;
+ /* We read the time so we need to read the object type again. */
+ if ((type = rdbLoadType(&rdb)) == -1) goto eoferr;
}
if (type == REDIS_RDB_OPCODE_EOF)
if ((key = rdbLoadStringObject(&rdb)) == NULL) goto eoferr;
/* Read value */
if ((val = rdbLoadObject(type,&rdb)) == NULL) goto eoferr;
- /* Check if the key already expired */
- if (expiretime != -1 && expiretime < now) {
+ /* Check if the key already expired. This function is used when loading
+ * an RDB file from disk, either at startup, or when an RDB was
+ * received from the master. In the latter case, the master is
+ * responsible for key expiry. If we would expire keys here, the
+ * snapshot taken by the master may not be reflected on the slave. */
+ if (server.masterhost == NULL && expiretime != -1 && expiretime < now) {
decrRefCount(key);
decrRefCount(val);
continue;
}
/* Add the new object in the hash table */
- retval = dbAdd(db,key,val);
- if (retval == REDIS_ERR) {
- redisLog(REDIS_WARNING,"Loading DB, duplicated key (%s) found! Unrecoverable error, exiting now.", key->ptr);
- exit(1);
- }
+ dbAdd(db,key,val);
+
/* Set the expire time if needed */
if (expiretime != -1) setExpire(db,key,expiretime);
"Background saving terminated with success");
server.dirty = server.dirty - server.dirty_before_bgsave;
server.lastsave = time(NULL);
+ server.lastbgsave_status = REDIS_OK;
} else if (!bysignal && exitcode != 0) {
redisLog(REDIS_WARNING, "Background saving error");
+ server.lastbgsave_status = REDIS_ERR;
} else {
redisLog(REDIS_WARNING,
"Background saving terminated by signal %d", bysignal);
- rdbRemoveTempFile(server.bgsavechildpid);
+ rdbRemoveTempFile(server.rdb_child_pid);
+ server.lastbgsave_status = REDIS_ERR;
}
- server.bgsavechildpid = -1;
- server.bgsavethread = (pthread_t) -1;
- server.bgsavethread_state = REDIS_BGSAVE_THREAD_UNACTIVE;
+ server.rdb_child_pid = -1;
/* 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) {
+ if (server.rdb_child_pid != -1) {
addReplyError(c,"Background save already in progress");
return;
}
- if (rdbSave(server.dbfilename) == REDIS_OK) {
+ if (rdbSave(server.rdb_filename) == REDIS_OK) {
addReply(c,shared.ok);
} else {
addReply(c,shared.err);
}
void bgsaveCommand(redisClient *c) {
- if (server.bgsavechildpid != -1 || server.bgsavethread != (pthread_t)-1) {
+ if (server.rdb_child_pid != -1) {
addReplyError(c,"Background save already in progress");
- return;
- }
- if (rdbSaveBackground(server.dbfilename) == REDIS_OK) {
+ } else if (server.aof_child_pid != -1) {
+ addReplyError(c,"Can't BGSAVE while AOF log rewriting is in progress");
+ } else if (rdbSaveBackground(server.rdb_filename) == REDIS_OK) {
addReplyStatus(c,"Background saving started");
} else {
addReply(c,shared.err);