redis源码分析之十二数据持久化存储
一、数据持久化
redis做为一种内存型数据库，做持久化，个人感觉略有鸡肋的意思。似乎有一种，别人有，自己不有也不行的感觉。以目前Redis主流的应用方式，如果仔细分析，基本上都是在内存中即可完成，对持久化没要求或者说不大。再举一个反例，如果内存中有几百G甚至更多的数据，真要是整体当机，恢复的时间基本就是灾难。
目前基本应用仍然是以关系型数据库或者其它数据库（如Hadoop，Mysql等）为持久化的方式，Redis只是发挥其内存数据库的长处，诸如高速缓存、分布式锁、分布式海量数据的高并发等。但有胜于无吧，本篇主要讨论一下其两种持久化的方式及其源码的实现。
二、持久化的形式
Redis有两种持久化实现的方式，一种是RDB数据快照方式，另外一种是AOF日志追加方式。
1、RDB快照
snapshotting，有手动和自动两种方式，可以在配置文件中进行配置，比如save m n（ save 300 1000）， 表示有一千条数据修改时，300秒刷新一次。RDB采用二进制加密方式，整体备份数据库。如果是手动的话，可以在命令行中执行SAVE（同步阻塞），BGSAVE（非阻塞）命令。自动备份会在下面几种情况发生：
修改配置文件中的配置项，如上面的Save M N的数值；
从服务器向主服务器发出复制请求，主服务器则使用bgsave生成rdb文件并传输给从服务器；
执行debug reload命令时，会使用save生成RDB文件；
未设置AOF的Redis服务使用shutdown关闭时，则会调用bgsave持久化。
2、AOF日志
日志方式其实是现在大多数数据库都采用的方式，就是将存储以日志的形式存储在缓冲区并刷入磁盘中。如果想恢复只要重新执行日志即可，比如MYSQL中就采用了这一种方式。AOF通过四个过程来实现持久化，即写入缓冲，文件同步 和文件重写，最后是重新加载。重写和重新加载的目的有点类似于KV型数据库LEVELDB等的压缩，就是减少空间占用为目的进行的。
AOF使用的是文本格式协议，所以它要比RDB的二进制占用空间大的多，但好处是，各个平台或者不同版本间的兼容性要好很多。它会在以下几种情况进行：
修改同步的always，保持数据同步，但性能较差（写命令都执行）；
修改每秒同步everysec，可能会损失一秒内的数据；
从不同步，这样就不管理同步了。
那么如何对上面的两种方式进行选择呢，一般情况下，如果数据量不是很大，同时对数据丢失的强一致性稍弱，可以选择RDB方式，或者虽然数据量很大，但是对Redis的数据访问压力不大，磁盘灾后恢复的时间要求也不紧急，仍然可以使用RDB方式。反之，如果对数据一致性要求很强，对保存数据占用磁盘空间不敏感，或者对实时访问Redis的需求和并发量都很高，就需要使用AOF的方式。
当然，你的最好的选择方式是，二者同时使用，把RDB做为灾备恢复的首选，利用AOF快速恢复RDB没有保存的数据。
三、源码分析
1、RDB
首先执行bgsave命令，Redis会判断是否已经启动此进程，如果已经启动直接返回；
然后fork一个进程通过命令info stats检查latest_fork_usec选项,检查最后一次备份的时间；
再后，子进程Fork成功后，通知Redis。
再下一步，子进程根据当前内存数据生成快照文件并替换原文件。
最后，备份完成通知Redis备份成功。
看一下源码：
//rdb.h rdb.c
typedef struct rdbSaveInfo {
/* Used saving and loading. */
int repl_stream_db; /* DB to select in server.master client. */
/* Used only loading. */
int repl_id_is_set; /* True if repl_id field is set. */
char repl_id[CONFIG_RUN_ID_SIZE+1]; /* Replication ID. */
long long repl_offset; /* Replication offset. */
} rdbSaveInfo;
void saveCommand(client *c) {
//前面提到过，如果已经启动，则直接退出
if (server.rdb_child_pid != -1) {
addReplyError(c,"Background save already in progress");
return;
}
//处理RDB存储信息相关
rdbSaveInfo rsi, * rsiptr;
rsiptr = rdbPopulateSaveInfo(&rsi);
if (rdbSave(server.rdb_filename,rsiptr) == C_OK) {
addReply(c,shared.ok);
} else {
addReply(c,shared.err);
}
}
/* BGSAVE [SCHEDULE] */
void bgsaveCommand(client * c) {
int schedule = 0;
/* The SCHEDULE option changes the behavior of BGSAVE when an AOF rewrite
* is in progress. Instead of returning an error a BGSAVE gets scheduled. */
if (c->argc > 1) {
if (c->argc == 2 && !strcasecmp(c->argv[1]->ptr,"schedule")) {
schedule = 1;
} else {
addReply(c,shared.syntaxerr);
return;
}
}
rdbSaveInfo rsi, * rsiptr;
rsiptr = rdbPopulateSaveInfo(&rsi);
//同样，已经执行，直接返回
if (server.rdb_child_pid != -1) {
addReplyError(c,"Background save already in progress");
//如果已经存在三种情况（RDB,AOF,MODULE）之一，则设置scheduled
} else if (hasActiveChildProcess()) {
if (schedule) {
server.rdb_bgsave_scheduled = 1;
addReplyStatus(c,"Background saving scheduled");
} else {
addReplyError(c,
"Another child process is active (AOF?): can't BGSAVE right now. "
"Use BGSAVE SCHEDULE in order to schedule a BGSAVE whenever "
"possible.");
}
} else if (rdbSaveBackground(server.rdb_filename,rsiptr) == C_OK) {
addReplyStatus(c,"Background saving started");
} else {
addReply(c,shared.err);
}
}
//下面两个函数分别代表着同步和后台保存
int rdbSave(char *filename, rdbSaveInfo *rsi) {
//下面的代码是普通的C文件操作文件的步骤
char tmpfile[256];
char cwd[MAXPATHLEN]; /* Current working dir path for error messages. */
FILE * fp;
rio rdb;
int error = 0;
snprintf(tmpfile,256,"temp-%d.rdb", (int) getpid());
fp = fopen(tmpfile,"w");
if (!fp) {
char *cwdp = getcwd(cwd,MAXPATHLEN);
serverLog(LL_WARNING,
"Failed opening the RDB file %s (in server root dir %s) "
"for saving: %s",
filename,
cwdp ? cwdp : "unknown",
strerror(errno));
return C_ERR;
}
//初始化相关RIO文件的参数
rioInitWithFile(&rdb,fp);
//设置事件
startSaving(RDBFLAGS_NONE);
//处理异步操作
if (server.rdb_save_incremental_fsync)
rioSetAutoSync(&rdb,REDIS_AUTOSYNC_BYTES);
// 保存RIO
if (rdbSaveRio(&rdb,&error,RDBFLAGS_NONE,rsi) == C_ERR) {
errno = error;
goto werr;
}
//其下又是C语言相关的保存刷入动作
/* Make sure data will not remain on the OS's output buffers */
if (fflush(fp) == EOF) goto werr;
if (fsync(fileno(fp)) == -1) goto werr;
if (fclose(fp) == EOF) goto werr;
/* Use RENAME to make sure the DB file is changed atomically only
* if the generate DB file is ok. */
if (rename(tmpfile,filename) == -1) {
char *cwdp = getcwd(cwd,MAXPATHLEN);
serverLog(LL_WARNING,
"Error moving temp DB file %s on the final "
"destination %s (in server root dir %s): %s",
tmpfile,
filename,
cwdp ? cwdp : "unknown",
strerror(errno));
unlink(tmpfile);
stopSaving(0);
return C_ERR;
}
serverLog(LL_NOTICE,"DB saved on disk");
server.dirty = 0;
server.lastsave = time(NULL);
server.lastbgsave_status = C_OK;
stopSaving(1);
return C_OK;
werr:
serverLog(LL_WARNING,"Write error saving DB on disk: %s", strerror(errno));
fclose(fp);
unlink(tmpfile);
stopSaving(0);
return C_ERR;
}
int rdbSaveBackground(char *filename, rdbSaveInfo * rsi) {
pid_t childpid;
if (hasActiveChildProcess()) return C_ERR;
server.dirty_before_bgsave = server.dirty;
server.lastbgsave_try = time(NULL);
openChildInfoPipe();
if ((childpid = redisFork()) == 0) {
int retval;
/* Child */
redisSetProcTitle("redis-rdb-bgsave");
retval = rdbSave(filename,rsi);
if (retval == C_OK) {
sendChildCOWInfo(CHILD_INFO_TYPE_RDB, "RDB");
}
exitFromChild((retval == C_OK) ? 0 : 1);
} else {
/* Parent */
if (childpid == -1) {
closeChildInfoPipe();
server.lastbgsave_status = C_ERR;
serverLog(LL_WARNING,"Can't save in background: fork: %s",
strerror(errno));
return C_ERR;
}
serverLog(LL_NOTICE,"Background saving started by pid %d",childpid);
server.rdb_save_time_start = time(NULL);
server.rdb_child_pid = childpid;
server.rdb_child_type = RDB_CHILD_TYPE_DISK;
return C_OK;
}
return C_OK; /* unreached * /
}
这里面涉及到了BIO和RIO，如果项目经验比较丰富的人，还是比较好理解的，其实RIO就是抽象出一层，对内存、文件和网络等的统一管理操作。看一下它的定义就明白了：
//下面的定义可以清楚的看到，SDS(内存)，标准输入输出，网络和管道
struct _rio {
/* Backend functions.
* Since this functions do not tolerate short writes or reads the return
* value is simplified to: zero on error, non zero on complete success. */
size_t (*read)(struct _rio *, void *buf, size_t len);
size_t (*write)(struct _rio *, const void *buf, size_t len);
off_t (*tell)(struct _rio *);
int (*flush)(struct _rio *);
/* The update_cksum method if not NULL is used to compute the checksum of
* all the data that was read or written so far. The method should be
* designed so that can be called with the current checksum, and the buf
* and len fields pointing to the new block of data to add to the checksum
* computation. */
void (*update_cksum)(struct _rio *, const void *buf, size_t len);
/* The current checksum and flags (see RIO_FLAG_*) */
uint64_t cksum, flags;
/* number of bytes read or written */
size_t processed_bytes;
/* maximum single read or write chunk size */
size_t max_processing_chunk;
/* Backend-specific vars. */
union {
/* In-memory buffer target. */
struct {
sds ptr;
off_t pos;
} buffer;
/* Stdio file pointer target. */
struct {
FILE *fp;
off_t buffered; /* Bytes written since last fsync. */
off_t autosync; /* fsync after 'autosync' bytes written. */
} file;
/* Connection object (used to read from socket) */
struct {
connection *conn; /* Connection */
off_t pos; /* pos in buf that was returned */
sds buf; /* buffered data */
size_t read_limit; /* don't allow to buffer/read more than that */
size_t read_so_far; /* amount of data read from the rio (not buffered) */
} conn;
/* FD target (used to write to pipe). */
struct {
int fd; /* File descriptor. */
off_t pos;
sds buf;
} fd;
} io;
};
typedef struct _rio rio;
而BIO则是后台操作IO，主要是辅助RIO的工作。也就是异步操作文件，这样的好处显而易见，就是为了防止阻塞服务进程。看一下bio.c中的代码：
void bioInit(void) {
pthread_attr_t attr;
pthread_t thread;
size_t stacksize;
int j;
/* Initialization of state vars and objects */
for (j = 0; j < BIO_NUM_OPS; j++) {
pthread_mutex_init(&bio_mutex[j],NULL);
pthread_cond_init(&bio_newjob_cond[j],NULL);
pthread_cond_init(&bio_step_cond[j],NULL);
bio_jobs[j] = listCreate();
bio_pending[j] = 0;
}
/* Set the stack size as by default it may be small in some system */
pthread_attr_init(&attr);
pthread_attr_getstacksize(&attr,&stacksize);
if (!stacksize) stacksize = 1; /* The world is full of Solaris Fixes */
while (stacksize < REDIS_THREAD_STACK_SIZE) stacksize *= 2;
pthread_attr_setstacksize(&attr, stacksize);
/* Ready to spawn our threads. We use the single argument the thread
* function accepts in order to pass the job ID the thread is
* responsible of. */
for (j = 0; j < BIO_NUM_OPS; j++) {
void *arg = (void*)(unsigned long) j;
if (pthread_create(&thread,&attr,bioProcessBackgroundJobs,arg) != 0) {
serverLog(LL_WARNING,"Fatal: Can't initialize Background Jobs.");
exit(1);
}
bio_threads[j] = thread;
}
}
void bioCreateBackgroundJob(int type, void *arg1, void *arg2, void *arg3) {
struct bio_job * job = zmalloc(sizeof(*job));
job->time = time(NULL);
job->arg1 = arg1;
job->arg2 = arg2;
job->arg3 = arg3;
pthread_mutex_lock(&bio_mutex[type]);
listAddNodeTail(bio_jobs[type],job);
bio_pending[type]++;
pthread_cond_signal(&bio_newjob_cond[type]);
pthread_mutex_unlock(&bio_mutex[type]);
}
看到的都是创建线程，要想提供异步操作，线程肯定是少不了的。这里比较麻烦的是存储的格式，但是这个不是本篇的重点，有兴趣可以查查相关资料，和什么PE,ELF文件的格式原理都一样，侧重点可能各有不同。
扯回来，两个Save函数其实都差不多，后台保存只不创建了PIPE，进行了后台的通信过程而已，实现RDB的代码基本一致,最后会调用：
int rdbSaveRio(rio *rdb, int *error, int rdbflags, rdbSaveInfo *rsi) {
dictIterator *di = NULL;
dictEntry *de;
char magic[10];
int j;
uint64_t cksum;
size_t processed = 0;
if (server.rdb_checksum)
rdb->update_cksum = rioGenericUpdateChecksum;
snprintf(magic,sizeof(magic),"REDIS%04d",RDB_VERSION);
if (rdbWriteRaw(rdb,magic,9) == -1) goto werr;
if (rdbSaveInfoAuxFields(rdb,rdbflags,rsi) == -1) goto werr;
if (rdbSaveModulesAux(rdb, REDISMODULE_AUX_BEFORE_RDB) == -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 = dictGetSafeIterator(d);
/* Write the SELECT DB opcode */
if (rdbSaveType(rdb,RDB_OPCODE_SELECTDB) == -1) goto werr;
if (rdbSaveLen(rdb,j) == -1) goto werr;
/* Write the RESIZE DB opcode. We trim the size to UINT32_MAX, which
* is currently the largest type we are able to represent in RDB sizes.
* However this does not limit the actual size of the DB to load since
* these sizes are just hints to resize the hash tables. */
uint64_t db_size, expires_size;
db_size = dictSize(db->dict);
expires_size = dictSize(db->expires);
if (rdbSaveType(rdb,RDB_OPCODE_RESIZEDB) == -1) goto werr;
if (rdbSaveLen(rdb,db_size) == -1) goto werr;
if (rdbSaveLen(rdb,expires_size) == -1) goto werr;
//下面正式保存键值对，循环进行
/* Iterate this DB writing every entry */
while((de = dictNext(di)) != NULL) {
sds keystr = dictGetKey(de);
robj key, *o = dictGetVal(de);
long long expire;
initStaticStringObject(key,keystr);
expire = getExpire(db,&key);
if (rdbSaveKeyValuePair(rdb,&key,o,expire) == -1) goto werr;
/* When this RDB is produced as part of an AOF rewrite, move
* accumulated diff from parent to child while rewriting in
* order to have a smaller final write. */
if (rdbflags & RDBFLAGS_AOF_PREAMBLE &&
rdb->processed_bytes > processed+AOF_READ_DIFF_INTERVAL_BYTES)
{
processed = rdb->processed_bytes;
aofReadDiffFromParent();
}
}
dictReleaseIterator(di);
di = NULL; /* So that we don't release it again on error. */
}
/* If we are storing the replication information on disk, persist
* the script cache as well: on successful PSYNC after a restart, we need
* to be able to process any EVALSHA inside the replication backlog the
* master will send us. */
if (rsi && dictSize(server.lua_scripts)) {
di = dictGetIterator(server.lua_scripts);
while((de = dictNext(di)) != NULL) {
robj *body = dictGetVal(de);
if (rdbSaveAuxField(rdb,"lua",3,body->ptr,sdslen(body->ptr)) == -1)
goto werr;
}
dictReleaseIterator(di);
di = NULL; /* So that we don't release it again on error. */
}
if (rdbSaveModulesAux(rdb, REDISMODULE_AUX_AFTER_RDB) == -1) goto werr;
/* EOF opcode */
if (rdbSaveType(rdb,RDB_OPCODE_EOF) == -1) goto werr;
/* CRC64 checksum. It will be zero if checksum computation is disabled, the
* loading code skips the check in this case. */
cksum = rdb->cksum;
memrev64ifbe(&cksum);
//写入校验数据
if (rioWrite(rdb,&cksum,8) == 0) goto werr;
return C_OK;
werr:
if (error) * error = errno;
if (di) dictReleaseIterator(di);
return C_ERR;
}
int rdbSaveKeyValuePair(rio *rdb, robj *key, robj *val, long long expiretime) {
int savelru = server.maxmemory_policy & MAXMEMORY_FLAG_LRU;
int savelfu = server.maxmemory_policy & MAXMEMORY_FLAG_LFU;
/* Save the expire time */
if (expiretime != -1) {
if (rdbSaveType(rdb,RDB_OPCODE_EXPIRETIME_MS) == -1) return -1;
if (rdbSaveMillisecondTime(rdb,expiretime) == -1) return -1;
}
/* Save the LRU info. */
if (savelru) {
uint64_t idletime = estimateObjectIdleTime(val);
idletime /= 1000; /* Using seconds is enough and requires less space.*/
if (rdbSaveType(rdb,RDB_OPCODE_IDLE) == -1) return -1;
if (rdbSaveLen(rdb,idletime) == -1) return -1;
}
/* Save the LFU info. */
if (savelfu) {
uint8_t buf[1];
buf[0] = LFUDecrAndReturn(val);
/* We can encode this in exactly two bytes: the opcode and an 8
* bit counter, since the frequency is logarithmic with a 0-255 range.
* Note that we do not store the halving time because to reset it
* a single time when loading does not affect the frequency much. */
if (rdbSaveType(rdb,RDB_OPCODE_FREQ) == -1) return -1;
if (rdbWriteRaw(rdb,buf,1) == -1) return -1;
}
/* Save type, key, value */
if (rdbSaveObjectType(rdb,val) == -1) return -1;
if (rdbSaveStringObject(rdb,key) == -1) return -1;
if (rdbSaveObject(rdb,val,key) == -1) return -1;
/* Delay return if required (for testing) */
if (server.rdb_key_save_delay)
usleep(server.rdb_key_save_delay);
return 1;
}
ssize_t rdbSaveObject(rio *rdb, robj *o, robj *key) {
ssize_t n = 0, nwritten = 0;
if (o->type == OBJ_STRING) {
/* Save a string value */
if ((n = rdbSaveStringObject(rdb,o)) == -1) return -1;
nwritten += n;
} else if (o->type == OBJ_LIST) {
/* Save a list value */
if (o->encoding == OBJ_ENCODING_QUICKLIST) {
quicklist *ql = o->ptr;
quicklistNode *node = ql->head;
if ((n = rdbSaveLen(rdb,ql->len)) == -1) return -1;
nwritten += n;
while(node) {
if (quicklistNodeIsCompressed(node)) {
void *data;
size_t compress_len = quicklistGetLzf(node, &data);
if ((n = rdbSaveLzfBlob(rdb,data,compress_len,node->sz)) == -1) return -1;
nwritten += n;
} else {
if ((n = rdbSaveRawString(rdb,node->zl,node->sz)) == -1) return -1;
nwritten += n;
}
node = node->next;
}
} else {
serverPanic("Unknown list encoding");
}
} else if (o->type == OBJ_SET) {
/* Save a set value */
if (o->encoding == OBJ_ENCODING_HT) {
dict *set = o->ptr;
dictIterator *di = dictGetIterator(set);
dictEntry *de;
if ((n = rdbSaveLen(rdb,dictSize(set))) == -1) {
dictReleaseIterator(di);
return -1;
}
nwritten += n;
while((de = dictNext(di)) != NULL) {
sds ele = dictGetKey(de);
if ((n = rdbSaveRawString(rdb,(unsigned char*)ele,sdslen(ele)))
== -1)
{
dictReleaseIterator(di);
return -1;
}
nwritten += n;
}
dictReleaseIterator(di);
} else if (o->encoding == OBJ_ENCODING_INTSET) {
size_t l = intsetBlobLen((intset*)o->ptr);
if ((n = rdbSaveRawString(rdb,o->ptr,l)) == -1) return -1;
nwritten += n;
} else {
serverPanic("Unknown set encoding");
}
} else if (o->type == OBJ_ZSET) {
/* Save a sorted set value */
if (o->encoding == OBJ_ENCODING_ZIPLIST) {
size_t l = ziplistBlobLen((unsigned char*)o->ptr);
if ((n = rdbSaveRawString(rdb,o->ptr,l)) == -1) return -1;
nwritten += n;
} else if (o->encoding == OBJ_ENCODING_SKIPLIST) {
zset *zs = o->ptr;
zskiplist *zsl = zs->zsl;
if ((n = rdbSaveLen(rdb,zsl->length)) == -1) return -1;
nwritten += n;
/* We save the skiplist elements from the greatest to the smallest
* (that's trivial since the elements are already ordered in the
* skiplist): this improves the load process, since the next loaded
* element will always be the smaller, so adding to the skiplist
* will always immediately stop at the head, making the insertion
* O(1) instead of O(log(N)). */
zskiplistNode *zn = zsl->tail;
while (zn != NULL) {
if ((n = rdbSaveRawString(rdb,
(unsigned char*)zn->ele,sdslen(zn->ele))) == -1)
{
return -1;
}
nwritten += n;
if ((n = rdbSaveBinaryDoubleValue(rdb,zn->score)) == -1)
return -1;
nwritten += n;
zn = zn->backward;
}
} else {
serverPanic("Unknown sorted set encoding");
}
} else if (o->type == OBJ_HASH) {
/* Save a hash value */
if (o->encoding == OBJ_ENCODING_ZIPLIST) {
size_t l = ziplistBlobLen((unsigned char*)o->ptr);
if ((n = rdbSaveRawString(rdb,o->ptr,l)) == -1) return -1;
nwritten += n;
} else if (o->encoding == OBJ_ENCODING_HT) {
dictIterator *di = dictGetIterator(o->ptr);
dictEntry *de;
if ((n = rdbSaveLen(rdb,dictSize((dict*)o->ptr))) == -1) {
dictReleaseIterator(di);
return -1;
}
nwritten += n;
while((de = dictNext(di)) != NULL) {
sds field = dictGetKey(de);
sds value = dictGetVal(de);
if ((n = rdbSaveRawString(rdb,(unsigned char*)field,
sdslen(field))) == -1)
{
dictReleaseIterator(di);
return -1;
}
nwritten += n;
if ((n = rdbSaveRawString(rdb,(unsigned char*)value,
sdslen(value))) == -1)
{
dictReleaseIterator(di);
return -1;
}
nwritten += n;
}
dictReleaseIterator(di);
} else {
serverPanic("Unknown hash encoding");
}
} else if (o->type == OBJ_STREAM) {
/* Store how many listpacks we have inside the radix tree. */
stream *s = o->ptr;
rax *rax = s->rax;
if ((n = rdbSaveLen(rdb,raxSize(rax))) == -1) return -1;
nwritten += n;
/* Serialize all the listpacks inside the radix tree as they are,
* when loading back, we'll use the first entry of each listpack
* to insert it back into the radix tree. */
raxIterator ri;
raxStart(&ri,rax);
raxSeek(&ri,"^",NULL,0);
while (raxNext(&ri)) {
unsigned char *lp = ri.data;
size_t lp_bytes = lpBytes(lp);
if ((n = rdbSaveRawString(rdb,ri.key,ri.key_len)) == -1) return -1;
nwritten += n;
if ((n = rdbSaveRawString(rdb,lp,lp_bytes)) == -1) return -1;
nwritten += n;
}
raxStop(&ri);
/* Save the number of elements inside the stream. We cannot obtain
* this easily later, since our macro nodes should be checked for
* number of items: not a great CPU / space tradeoff. */
if ((n = rdbSaveLen(rdb,s->length)) == -1) return -1;
nwritten += n;
/* Save the last entry ID. */
if ((n = rdbSaveLen(rdb,s->last_id.ms)) == -1) return -1;
nwritten += n;
if ((n = rdbSaveLen(rdb,s->last_id.seq)) == -1) return -1;
nwritten += n;
/* The consumer groups and their clients are part of the stream
* type, so serialize every consumer group. */
/* Save the number of groups. */
size_t num_cgroups = s->cgroups ? raxSize(s->cgroups) : 0;
if ((n = rdbSaveLen(rdb,num_cgroups)) == -1) return -1;
nwritten += n;
if (num_cgroups) {
/* Serialize each consumer group. */
raxStart(&ri,s->cgroups);
raxSeek(&ri,"^",NULL,0);
while(raxNext(&ri)) {
streamCG *cg = ri.data;
/* Save the group name. */
if ((n = rdbSaveRawString(rdb,ri.key,ri.key_len)) == -1)
return -1;
nwritten += n;
/* Last ID. */
if ((n = rdbSaveLen(rdb,cg->last_id.ms)) == -1) return -1;
nwritten += n;
if ((n = rdbSaveLen(rdb,cg->last_id.seq)) == -1) return -1;
nwritten += n;
/* Save the global PEL. */
if ((n = rdbSaveStreamPEL(rdb,cg->pel,1)) == -1) return -1;
nwritten += n;
/* Save the consumers of this group. */
if ((n = rdbSaveStreamConsumers(rdb,cg)) == -1) return -1;
nwritten += n;
}
raxStop(&ri);
}
} else if (o->type == OBJ_MODULE) {
/* Save a module-specific value. */
RedisModuleIO io;
moduleValue *mv = o->ptr;
moduleType *mt = mv->type;
/* Write the "module" identifier as prefix, so that we'll be able
* to call the right module during loading. */
int retval = rdbSaveLen(rdb,mt->id);
if (retval == -1) return -1;
io.bytes += retval;
/* Then write the module-specific representation + EOF marker. */
moduleInitIOContext(io,mt,rdb,key);
mt->rdb_save(&io,mv->value);
retval = rdbSaveLen(rdb,RDB_MODULE_OPCODE_EOF);
if (retval == -1)
io.error = 1;
else
io.bytes += retval;
if (io.ctx) {
moduleFreeContext(io.ctx);
zfree(io.ctx);
}
return io.error ? -1 : (ssize_t)io.bytes;
} else {
serverPanic("Unknown object type");
}
return nwritten;
}
最后的rdbSaveObject才是最核心的保存数据成员的函数。最终在：
static int rdbWriteRaw(rio *rdb, void *p, size_t len) {
if (rdb && rioWrite(rdb,p,len) == 0)
return -1;
return len;
}
static inline size_t rioWrite(rio *r, const void *buf, size_t len) {
if (r->flags & RIO_FLAG_WRITE_ERROR) return 0;
while (len) {
size_t bytes_to_write = (r->max_processing_chunk && r->max_processing_chunk < len) ? r->max_processing_chunk : len;
if (r->update_cksum) r->update_cksum(r,buf,bytes_to_write);
if (r->write(r,buf,bytes_to_write) == 0) {
r->flags |= RIO_FLAG_WRITE_ERROR;
return 0;
}
buf = (char*)buf + bytes_to_write;
len -= bytes_to_write;
r->processed_bytes += bytes_to_write;
}
return 1;
}
通过RIO的接口，直接调用系统接口，进行写操作。
2、AOF
首先，将redis命令写入aof_buf缓冲区。
然后，根据备份方式将缓冲区中数据写入日志文件。
再后，如果aof的膨胀过大，将根据配置的策略重写AOF文件，实现压缩数据，减少磁盘占用空间。
最后，redis在重启时会加载重写成功的新的AOF文件，保证数据安全。
看一下相关源码：
struct redisServer{
...
sds aof_buf; /* 第1239行：AOF buffer, written before entering the event loop */
...
}
//Redis会同步操作propagate函数对此变量写入操作
void propagate(struct redisCommand *cmd, int dbid, robj **argv, int argc,
int flags)
{
//检查AOF选项，打开则调用
if (server.aof_state != AOF_OFF && flags & PROPAGATE_AOF)
feedAppendOnlyFile(cmd,dbid,argv,argc);
if (flags & PROPAGATE_REPL)
replicationFeedSlaves(server.slaves,dbid,argv,argc);
}
void feedAppendOnlyFile(struct redisCommand *cmd, int dictid, robj **argv, int argc) {
sds buf = sdsempty();
robj *tmpargv[3];
/* The DB this command was targeting is not the same as the last command
* we appended. To issue a SELECT command is needed. */
if (dictid != server.aof_selected_db) {
char seldb[64];
snprintf(seldb,sizeof(seldb),"%d",dictid);
buf = sdscatprintf(buf,"*2\r\n$6\r\nSELECT\r\n$%lu\r\n%s\r\n",
(unsigned long)strlen(seldb),seldb);
server.aof_selected_db = dictid;
}
if (cmd->proc == expireCommand || cmd->proc == pexpireCommand ||
cmd->proc == expireatCommand) {
/* Translate EXPIRE/PEXPIRE/EXPIREAT into PEXPIREAT */
buf = catAppendOnlyExpireAtCommand(buf,cmd,argv[1],argv[2]);
} else if (cmd->proc == setexCommand || cmd->proc == psetexCommand) {
/* Translate SETEX/PSETEX to SET and PEXPIREAT */
tmpargv[0] = createStringObject("SET",3);
tmpargv[1] = argv[1];
tmpargv[2] = argv[3];
buf = catAppendOnlyGenericCommand(buf,3,tmpargv);
decrRefCount(tmpargv[0]);
buf = catAppendOnlyExpireAtCommand(buf,cmd,argv[1],argv[2]);
} else if (cmd->proc == setCommand && argc > 3) {
int i;
robj *exarg = NULL, *pxarg = NULL;
/* Translate SET [EX seconds][PX milliseconds] to SET and PEXPIREAT */
buf = catAppendOnlyGenericCommand(buf,3,argv);
for (i = 3; i < argc; i ++) {
if (!strcasecmp(argv[i]->ptr, "ex")) exarg = argv[i+1];
if (!strcasecmp(argv[i]->ptr, "px")) pxarg = argv[i+1];
}
serverAssert(!(exarg && pxarg));
if (exarg)
buf = catAppendOnlyExpireAtCommand(buf,server.expireCommand,argv[1],
exarg);
if (pxarg)
buf = catAppendOnlyExpireAtCommand(buf,server.pexpireCommand,argv[1],
pxarg);
} else {
/* All the other commands don't need translation or need the
* same translation already operated in the command vector
* for the replication itself. */
buf = catAppendOnlyGenericCommand(buf,argc,argv);
}
//下面是写入和重写两个动作，前面介绍过
/* Append to the AOF buffer. This will be flushed on disk just before
* of re-entering the event loop, so before the client will get a
* positive reply about the operation performed. */
if (server.aof_state == AOF_ON)
server.aof_buf = sdscatlen(server.aof_buf,buf,sdslen(buf));
/* If a background append only file rewriting is in progress we want to
* accumulate the differences between the child DB and the current one
* in a buffer, so that when the child process will do its work we
* can append the differences to the new append only file. */
if (server.aof_child_pid != -1)
aofRewriteBufferAppend((unsigned char*)buf,sdslen(buf));
sdsfree(buf);
}
在自动触发（3中会介绍）会将写入的数据刷进文件：
void flushAppendOnlyFile(int force) {
ssize_t nwritten;
int sync_in_progress = 0;
mstime_t latency;
if (sdslen(server.aof_buf) == 0) {
/* Check if we need to do fsync even the aof buffer is empty,
* because previously in AOF_FSYNC_EVERYSEC mode, fsync is
* called only when aof buffer is not empty, so if users
* stop write commands before fsync called in one second,
* the data in page cache cannot be flushed in time. */
if (server.aof_fsync == AOF_FSYNC_EVERYSEC &&
server.aof_fsync_offset != server.aof_current_size &&
server.unixtime > server.aof_last_fsync &&
!(sync_in_progress = aofFsyncInProgress())) {
goto try_fsync;
} else {
return;
}
}
if (server.aof_fsync == AOF_FSYNC_EVERYSEC)
sync_in_progress = aofFsyncInProgress();
if (server.aof_fsync == AOF_FSYNC_EVERYSEC && !force) {
/* With this append fsync policy we do background fsyncing.
* If the fsync is still in progress we can try to delay
* the write for a couple of seconds. */
if (sync_in_progress) {
if (server.aof_flush_postponed_start == 0) {
/* No previous write postponing, remember that we are
* postponing the flush and return. */
server.aof_flush_postponed_start = server.unixtime;
return;
} else if (server.unixtime - server.aof_flush_postponed_start < 2) {
/* We were already waiting for fsync to finish, but for less
* than two seconds this is still ok. Postpone again. */
return;
}
/* Otherwise fall trough, and go write since we can't wait
* over two seconds. */
server.aof_delayed_fsync++;
serverLog(LL_NOTICE,"Asynchronous AOF fsync is taking too long (disk is busy?). Writing the AOF buffer without waiting for fsync to complete, this may slow down Redis.");
}
}
/* We want to perform a single write. This should be guaranteed atomic
* at least if the filesystem we are writing is a real physical one.
* While this will save us against the server being killed I don't think
* there is much to do about the whole server stopping for power problems
* or alike */
if (server.aof_flush_sleep && sdslen(server.aof_buf)) {
usleep(server.aof_flush_sleep);
}
latencyStartMonitor(latency);
nwritten = aofWrite(server.aof_fd,server.aof_buf,sdslen(server.aof_buf));
latencyEndMonitor(latency);
/* We want to capture different events for delayed writes:
* when the delay happens with a pending fsync, or with a saving child
* active, and when the above two conditions are missing.
* We also use an additional event name to save all samples which is
* useful for graphing / monitoring purposes. */
if (sync_in_progress) {
latencyAddSampleIfNeeded("aof-write-pending-fsync",latency);
} else if (hasActiveChildProcess()) {
latencyAddSampleIfNeeded("aof-write-active-child",latency);
} else {
latencyAddSampleIfNeeded("aof-write-alone",latency);
}
latencyAddSampleIfNeeded("aof-write",latency);
/* We performed the write so reset the postponed flush sentinel to zero. */
server.aof_flush_postponed_start = 0;
if (nwritten != (ssize_t)sdslen(server.aof_buf)) {
static time_t last_write_error_log = 0;
int can_log = 0;
/* Limit logging rate to 1 line per AOF_WRITE_LOG_ERROR_RATE seconds. */
if ((server.unixtime - last_write_error_log) > AOF_WRITE_LOG_ERROR_RATE) {
can_log = 1;
last_write_error_log = server.unixtime;
}
/* Log the AOF write error and record the error code. */
if (nwritten == -1) {
if (can_log) {
serverLog(LL_WARNING,"Error writing to the AOF file: %s",
strerror(errno));
server.aof_last_write_errno = errno;
}
} else {
if (can_log) {
serverLog(LL_WARNING,"Short write while writing to "
"the AOF file: (nwritten=%lld, "
"expected=%lld)",
(long long)nwritten,
(long long)sdslen(server.aof_buf));
}
if (ftruncate(server.aof_fd, server.aof_current_size) == -1) {
if (can_log) {
serverLog(LL_WARNING, "Could not remove short write "
"from the append-only file. Redis may refuse "
"to load the AOF the next time it starts. "
"ftruncate: %s", strerror(errno));
}
} else {
/* If the ftruncate() succeeded we can set nwritten to
* -1 since there is no longer partial data into the AOF. */
nwritten = -1;
}
server.aof_last_write_errno = ENOSPC;
}
/* Handle the AOF write error. */
if (server.aof_fsync == AOF_FSYNC_ALWAYS) {
/* We can't recover when the fsync policy is ALWAYS since the
* reply for the client is already in the output buffers, and we
* have the contract with the user that on acknowledged write data
* is synced on disk. */
serverLog(LL_WARNING,"Can't recover from AOF write error when the AOF fsync policy is 'always'. Exiting...");
exit(1);
} else {
/* Recover from failed write leaving data into the buffer. However
* set an error to stop accepting writes as long as the error
* condition is not cleared. */
server.aof_last_write_status = C_ERR;
/* Trim the sds buffer if there was a partial write, and there
* was no way to undo it with ftruncate(2). */
if (nwritten > 0) {
server.aof_current_size += nwritten;
sdsrange(server.aof_buf,nwritten,-1);
}
return; /* We'll try again on the next call... */
}
} else {
/* Successful write(2). If AOF was in error state, restore the
* OK state and log the event. */
if (server.aof_last_write_status == C_ERR) {
serverLog(LL_WARNING,
"AOF write error looks solved, Redis can write again.");
server.aof_last_write_status = C_OK;
}
}
server.aof_current_size += nwritten;
/* Re-use AOF buffer when it is small enough. The maximum comes from the
* arena size of 4k minus some overhead (but is otherwise arbitrary). */
if ((sdslen(server.aof_buf)+sdsavail(server.aof_buf)) < 4000) {
sdsclear(server.aof_buf);
} else {
sdsfree(server.aof_buf);
server.aof_buf = sdsempty();
}
try_fsync:
/* Don't fsync if no-appendfsync-on-rewrite is set to yes and there are
* children doing I/O in the background. */
if (server.aof_no_fsync_on_rewrite && hasActiveChildProcess())
return;
/* Perform the fsync if needed. * /
if (server.aof_fsync == AOF_FSYNC_ALWAYS) {
/* redis_fsync is defined as fdatasync() for Linux in order to avoid
* flushing metadata. */
latencyStartMonitor(latency);
redis_fsync(server.aof_fd); /* Let's try to get this data on the disk */
latencyEndMonitor(latency);
latencyAddSampleIfNeeded("aof-fsync-always",latency);
server.aof_fsync_offset = server.aof_current_size;
server.aof_last_fsync = server.unixtime;
} else if ((server.aof_fsync == AOF_FSYNC_EVERYSEC &&
server.unixtime > server.aof_last_fsync)) {
if (!sync_in_progress) {
aof_background_fsync(server.aof_fd);
server.aof_fsync_offset = server.aof_current_size;
}
server.aof_last_fsync = server.unixtime;
}
}
ssize_t aofWrite(int fd, const char * buf, size_t len) {
ssize_t nwritten = 0, totwritten = 0;
while(len) {
nwritten = write(fd, buf, len);
if (nwritten < 0) {
if (errno == EINTR) continue;
return totwritten ? totwritten : -1;
}
len -= nwritten;
buf += nwritten;
totwritten += nwritten;
}
return totwritten;
}
一顿操作猛如虎，其实细看也没啥。最终通过宏定义：
#ifdef __linux__
#define redis_fsync fdatasync
#else
#define redis_fsync fsync
#endif
调用系统fdatasync来实现数据的落盘。
3、自动触发
这个在老地方，定时器处理函数：
int serverCron(struct aeEventLoop *eventLoop, long long id, void *clientData) {
...
/* Start a scheduled AOF rewrite if this was requested by the user while
* a BGSAVE was in progress. */
if (!hasActiveChildProcess() &&
server.aof_rewrite_scheduled)
{
rewriteAppendOnlyFileBackground();
}
/* Check if a background saving or AOF rewrite in progress terminated. */
if (hasActiveChildProcess() || ldbPendingChildren())
{
checkChildrenDone();
} else {
/* 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;
/* Save if we reached the given amount of changes,
* the given amount of seconds, and if the latest bgsave was
* successful or if, in case of an error, at least
* CONFIG_BGSAVE_RETRY_DELAY seconds already elapsed. */
if (server.dirty >= sp->changes &&
server.unixtime-server.lastsave > sp->seconds &&
(server.unixtime-server.lastbgsave_try >
CONFIG_BGSAVE_RETRY_DELAY ||
server.lastbgsave_status == C_OK))
{
serverLog(LL_NOTICE,"%d changes in %d seconds. Saving...",
sp->changes, (int)sp->seconds);
rdbSaveInfo rsi, *rsiptr;
rsiptr = rdbPopulateSaveInfo(&rsi);
rdbSaveBackground(server.rdb_filename,rsiptr);
break;
}
}
/* Trigger an AOF rewrite if needed. */
if (server.aof_state == AOF_ON &&
!hasActiveChildProcess() &&
server.aof_rewrite_perc &&
server.aof_current_size > server.aof_rewrite_min_size)
{
long long base = server.aof_rewrite_base_size ?
server.aof_rewrite_base_size : 1;
long long growth = (server.aof_current_size*100/base) - 100;
if (growth >= server.aof_rewrite_perc) {
serverLog(LL_NOTICE,"Starting automatic rewriting of AOF on %lld%% growth",growth);
rewriteAppendOnlyFileBackground();
}
}
/* AOF postponed flush: Try at every cron cycle if the slow fsync * completed. */
if (server.aof_flush_postponed_start) flushAppendOnlyFile(0);
/* AOF write errors: in this case we have a buffer to flush as well and
* clear the AOF error in case of success to make the DB writable again,
* however to try every second is enough in case of 'hz' is set to
* an higher frequency. */
run_with_period(1000) {
if (server.aof_last_write_status == C_ERR)
flushAppendOnlyFile(0);
}
}
...
}
其实就是对配置信息和当前状态不断进行判断，然后进行数据的存储。其它的触发，可以根据调用栈来寻找，包括主从复制，其实也启动bgsave通过RDB文件来传输，这里就不再过多的介绍了。
四、总结
REDIS持久化，目前看来，醉翁之意不在此。不过做为数据存储选项，多一个后手，总是有胜于无。
