关于 Redis 的 Persistence, 最好的入门材料应该是:
它有两种形式,AOF 和 RDB:RDB 相对来说是一个数据库的“snapshot”,通过 SAVE 或者 BGSAVE 存储,在这个贴 指出 Redis 有一个 redis.conf,存储了具体的dbfilename , 它的默认值是dump.rdb。具体来说,Redis 是靠 fork, 然后利用 fork 的 COW, 在子进程写一个 tmp rdb 文件,然后写入完成后 replace 原来的 rdb 文件
AOF 又叫 append-only file. 它以一定粒度的范围来写入。这玩意需要配置 appendonly yes. 相当于一个顺序写入的文件Redis 每次受到 SET 等更改命令的时候,会 apply 到 AOF
restart redis的时候,它会重放日志
append 本身写盘是肯定比 random access 快的,但是每次 fsync 仍然要付出很高代价,所以 Redis 允许指定不同的 async 形式:设置appendfsync参数,可以always 总是刷新
everysec 一秒一次
appendfsync no: Never fsync. 这不代表不刷新,只是交给操作系统处理
Normally Linux will flush data every 30 seconds with this configuration, but it's up to the kernel exact tuning.
(我以为几秒刷一次,竟然这么慢...)
redis 官网建议一秒刷新一次,反正我也觉得如果做缓存的话是不是可以容忍这种不刷新...AOF 暗示它会无限增长,那么肯定需要某种策略处理:像 Raft 一样的系统支持将一批 Log 构建成一个 snap, 这个策略暗示 AOF 可能整个再度成为一个 rdb
BGREWRITEAOF 会重建 AOF。Redis 能够 check aof, 检查 aof 数据的完整性是否有问题
AOF 结构和 REPL 协议是一样的
AOF 的实现是通过:会有一个 appendonly.aof 文件,这个在 redis.conf的 dir 和 appendfilename 选项中被定义
一个进程被 fork 出来,承担功能,把新的 AOF 写入 文件,同时父进程一边写入一边在内存buffer 写
所以:AOF 和 RDB 是两种不同的策略(这和我之前了解到的不一样),AOF 逻辑类似 Raft Snapshot 那一套;RDB 是对整个系统备份,相对来说 RDB 结构肯定会小一些。
RDB 在同步数据方面有很大的价值,通常 RDB 的备份会从 master 被发送给 slave,作为一种初始化
RDB 和 AOF 实际 BGSAVE 或者 AOFREWRITE 的时候互相不会干扰
下面是几张和上述流程有关的图,从书上抄下来的。
RDB
rdb 的定义部分在 rdb.c 和 rdb.h 两个文件中,rdb.h 定义了很多 msb 及其他相关的宏。同时 server.h 本身也承担了一部分 RDB 的逻辑。
在这个版本的 1196行之后,标注了对应的逻辑:
/* RDB persistence */
long long dirty; /* Changes to DB from the last save */
long long dirty_before_bgsave; /* Used to restore dirty on failed BGSAVE */
pid_t rdb_child_pid; /* PID of RDB saving child */
struct saveparam *saveparams; /* Save points array for RDB */
int saveparamslen; /* Number of saving points */
char *rdb_filename; /* Name of RDB file */
int rdb_compression; /* Use compression in RDB? */
int rdb_checksum; /* Use RDB checksum? */
time_t lastsave; /* Unix time of last successful save */
time_t lastbgsave_try; /* Unix time of last attempted bgsave */
time_t rdb_save_time_last; /* Time used by last RDB save run. */
time_t rdb_save_time_start; /* Current RDB save start time. */
int rdb_bgsave_scheduled; /* BGSAVE when possible if true. */
int rdb_child_type; /* Type of save by active child. */
int lastbgsave_status; /* C_OK or C_ERR */
int stop_writes_on_bgsave_err; /* Don't allow writes if can't BGSAVE */
int rdb_pipe_write; /* RDB pipes used to transfer the rdb */
int rdb_pipe_read; /* data to the parent process in diskless repl. */
connection **rdb_pipe_conns; /* Connections which are currently the */
int rdb_pipe_numconns; /* target of diskless rdb fork child. */
int rdb_pipe_numconns_writing; /* Number of rdb conns with pending writes. */
char *rdb_pipe_buff; /* In diskless replication, this buffer holds data */
int rdb_pipe_bufflen; /* that was read from the the rdb pipe. */
int rdb_key_save_delay; /* Delay in microseconds between keys while
* writing the RDB. (for testings) */
int key_load_delay; /* Delay in microseconds between keys while
* loading aof or rdb. (for testings) */dirty 和 dirty_before_bgsave 对应两套逻辑,但是都是表示save对应的写入的 cnt.
由于之前说的 fork, 所以需要 child_pid
saveparams 是 save 相关的参数
同时,server.h 定义了 save 相关的 info, 表示 rdb 的 metadata.
/* This structure can be optionally passed to RDB save/load functions in
* order to implement additional functionalities, by storing and loading
* metadata to the RDB file.
*
* Currently the only use is to select a DB at load time, useful in
* replication in order to make sure that chained slaves (slaves of slaves)
* select the correct DB and are able to accept the stream coming from the
* top-level master. */
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;
rdbSave 这个函数具体流程不长,可以看看:
/* Save the DB on disk. Return C_ERR on error, C_OK on success. */
int rdbSave(char *filename, rdbSaveInfo *rsi) {
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;
}
rioInitWithFile(&rdb,fp);
startSaving(RDBFLAGS_NONE);
if (server.rdb_save_incremental_fsync)
rioSetAutoSync(&rdb,REDIS_AUTOSYNC_BYTES);
if (rdbSaveRio(&rdb,&error,RDBFLAGS_NONE,rsi) == C_ERR) {
errno = error;
goto werr;
}
/* 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;
}
其中,通过对 全局 server 的读写,完成操作。
/* Global vars */
struct redisServer server; /* Server global state */
volatile unsigned long lru_clock; /* Server global current LRU time. */
同时,这里有 rdbSaveRio, 这是最主要的代码段:
/* Produces a dump of the database in RDB format sending it to the specified
* Redis I/O channel. On success C_OK is returned, otherwise C_ERR
* is returned and part of the output, or all the output, can be
* missing because of I/O errors.
*
* When the function returns C_ERR and if 'error' is not NULL, the
* integer pointed by 'error' is set to the value of errno just after the I/O
* error. */
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;
/* 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);
if (rioWrite(rdb,&cksum,8) == 0) goto werr;
return C_OK;
werr:
if (error) *error = errno;
if (di) dictReleaseIterator(di);
return C_ERR;
}
屏蔽 rio 相关的逻辑,然后这一段可以参考:snprintf(magic,sizeof(magic),"REDIS%04d",RDB_VERSION); 在 rdbWriteRaw 写入 REDIS 和 version
写入 redis 对应的 flag
针对每个 DB 写入
AOF
在beforeSleep 中,有:
flushAppendOnlyFile(0);
其中 0 表示 not force.
server.aof_buf 存储着对应的 aof 数据,出现 set 指令的时候,aof_buf 会被记录。
AOF 具体逻辑如下:
/* Write the append only file buffer on disk.
*
* Since we are required to write the AOF before replying to the client,
* and the only way the client socket can get a write is entering when the
* the event loop, we accumulate all the AOF writes in a memory
* buffer and write it on disk using this function just before entering
* the event loop again.
*
* About the 'force' argument:
*
* When the fsync policy is set to 'everysec' we may delay the flush if there
* is still an fsync() going on in the background thread, since for instance
* on Linux write(2) will be blocked by the background fsync anyway.
* When this happens we remember that there is some aof buffer to be
* flushed ASAP, and will try to do that in the serverCron() function.
*
* However if force is set to 1 we'll write regardless of the background
* fsync. */
#define AOF_WRITE_LOG_ERROR_RATE 30 /* Seconds between errors logging. */
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;
}
}