【Redis】哨兵初始化和主观下线
在的redis启动函数main(server.c文件)中,对哨兵模式进行了检查,如果是哨兵模式,将调用initSentinelConfig和initSentinel进行初始化,initServer函数中会注册哨兵的时间事件,最后调用sentinelIsRunning运行哨兵实例,
int main(int argc, char **argv) {
// 省略...
// 检查哨兵模式
server.sentinel_mode = checkForSentinelMode(argc,argv);
initServerConfig();
// 省略...
if (server.sentinel_mode) {
initSentinelConfig(); // 初始化哨兵配置
initSentinel(); // 初始化哨兵
}
// 省略...
// 初始化服务
initServer();
// 省略...
if (!server.sentinel_mode) { // 非哨兵模式
// 省略...
} else {
ACLLoadUsersAtStartup();
InitServerLast();
// 运行哨兵实例
sentinelIsRunning();
if (server.supervised_mode == SUPERVISED_SYSTEMD) {
redisCommunicateSystemd("STATUS=Ready to accept connections\n");
redisCommunicateSystemd("READY=1\n");
}
}
// 省略...
aeMain(server.el);
aeDeleteEventLoop(server.el);
return 0;
}
哨兵初始化
哨兵模式校验
checkForSentinelMode
checkForSentinelMode函数在server.c文件中,用于校验是否是哨兵模式,可以看到有两种方式校验哨兵模式:
- 直接执行
redis-sentinel
命令 - 执行
redis-server
命令,命令参数中带有–sentinel
参数
int checkForSentinelMode(int argc, char **argv) {
int j;
// 直接执行redis-sentinel
if (strstr(argv[0],"redis-sentinel") != NULL) return 1;
for (j = 1; j < argc; j++)
if (!strcmp(argv[j],"--sentinel")) return 1; // 执行的命令参数中,是否有–sentinel
return 0;
}
初始化配置项
initSentinelConfig
initSentinelConfig函数在sentinel.c文件中,用于初始化哨兵配置:
- 将哨兵实例的端口号设置为
REDIS_SENTINEL_PORT
,默认值26379 - 将
protected_mode
设置为0,表示允许外部链接哨兵实例,而不是只能通过127.0.0.1本地连接 server
#define REDIS_SENTINEL_PORT 26379
void initSentinelConfig(void) {
server.port = REDIS_SENTINEL_PORT; /* 设置默认端口 */
server.protected_mode = 0; /* 允许外部链接哨兵实例 */
}
initSentinel
在看initSentinel函数之前,首先看下Redis中哨兵sentinel对象对应的结构体sentinelState:
- current_epoch:当前纪元,投票选举Leader时使用,纪元可以理解为投票的轮次
- masters:监控的master节点哈希表,Key为master节点名称, value为master节点对应sentinelRedisInstance实例的指针
struct sentinelState {
char myid[CONFIG_RUN_ID_SIZE+1]; /* sentinel ID */
uint64_t current_epoch; /* 当前的纪元(投票轮次)*/
dict *masters; /* 监控的master节点哈希表,Key为master节点名称, value为master节点对应的实例对象的指针 */
int tilt; /* TILT模式 */
int running_scripts;
mstime_t tilt_start_time;
mstime_t previous_time;
list *scripts_queue;
char *announce_ip;
int announce_port;
unsigned long simfailure_flags;
int deny_scripts_reconfig;
char *sentinel_auth_pass;
char *sentinel_auth_user;
int resolve_hostnames;
int announce_hostnames;
} sentinel;
sentinelRedisInstance是一个通用的结构体,在sentinel.c文件中定义,它既可以表示主节点,也可以表示从节点或者其他哨兵实例,从中选选出了一些主要的内容:
typedef struct {
int flags; /* 一些状态标识 */
char *name; /* Master name from the point of view of this sentinel. */
char *runid; /* 实例的运行ID */
uint64_t config_epoch; /* 配置的纪元. */
mstime_t s_down_since_time; /* 主观下线时长 */
mstime_t o_down_since_time; /* 客观下线时长 */
dict *sentinels; /* 监控同一主节点的其他哨兵实例 */
dict *slaves; /* slave节点(从节点) */
/* 故障切换 */
char *leader; /* 如果是master节点,保存了需要执行故障切换的哨兵leader的runid,如果是一个哨兵,保存的是这个哨兵投票选举的leader*/
uint64_t leader_epoch; /* leader纪元 */
uint64_t failover_epoch;
int failover_state; /* 故障切换状态 */
// 省略...
} sentinelRedisInstance;
initSentinel函数同样在sentinel.c文件中,用于初始化哨兵,由于哨兵实例与普通Redis实例不一样,所以需要替换Redis中的命令,添加哨兵实例命令,哨兵实例使用的命令在sentinelcmds中定义:
- 将server.commands和server.orig_commands保存的常规Redis命令清除
- 遍历sentinelcmds哨兵实例专用命令,将命令添加到server.commands和server.orig_commands中
- 初始化sentinel实例中的数据项
// 哨兵实例下的命令
struct redisCommand sentinelcmds[] = {
{"ping",pingCommand,1,"fast @connection",0,NULL,0,0,0,0,0},
{"sentinel",sentinelCommand,-2,"admin",0,NULL,0,0,0,0,0},
{"subscribe",subscribeCommand,-2,"pub-sub",0,NULL,0,0,0,0,0},
{"unsubscribe",unsubscribeCommand,-1,"pub-sub",0,NULL,0,0,0,0,0},
{"psubscribe",psubscribeCommand,-2,"pub-sub",0,NULL,0,0,0,0,0},
{"punsubscribe",punsubscribeCommand,-1,"pub-sub",0,NULL,0,0,0,0,0},
{"publish",sentinelPublishCommand,3,"pub-sub fast",0,NULL,0,0,0,0,0},
{"info",sentinelInfoCommand,-1,"random @dangerous",0,NULL,0,0,0,0,0},
{"role",sentinelRoleCommand,1,"fast read-only @dangerous",0,NULL,0,0,0,0,0},
{"client",clientCommand,-2,"admin random @connection",0,NULL,0,0,0,0,0},
{"shutdown",shutdownCommand,-1,"admin",0,NULL,0,0,0,0,0},
{"auth",authCommand,-2,"no-auth fast @connection",0,NULL,0,0,0,0,0},
{"hello",helloCommand,-1,"no-auth fast @connection",0,NULL,0,0,0,0,0},
{"acl",aclCommand,-2,"admin",0,NULL,0,0,0,0,0,0},
{"command",commandCommand,-1, "random @connection", 0,NULL,0,0,0,0,0,0}
};
/* 初始化哨兵 */
void initSentinel(void) {
unsigned int j;
/* 将常规的Redis命令移除,增加哨兵实例专用的命令 */
dictEmpty(server.commands,NULL);
dictEmpty(server.orig_commands,NULL);
ACLClearCommandID();
for (j = 0; j < sizeof(sentinelcmds)/sizeof(sentinelcmds[0]); j++) {
int retval;
struct redisCommand *cmd = sentinelcmds+j;
cmd->id = ACLGetCommandID(cmd->name);
// 添加到server.commands
retval = dictAdd(server.commands, sdsnew(cmd->name), cmd);
serverAssert(retval == DICT_OK);
// 添加到server.orig_commands
retval = dictAdd(server.orig_commands, sdsnew(cmd->name), cmd);
serverAssert(retval == DICT_OK);
if (populateCommandTableParseFlags(cmd,cmd->sflags) == C_ERR)
serverPanic("Unsupported command flag");
}
/* 初始化其他数据项 */
// current_epoch初始化为0
sentinel.current_epoch = 0;
// 监控的master节点实例对象
sentinel.masters = dictCreate(&instancesDictType,NULL);
sentinel.tilt = 0;
sentinel.tilt_start_time = 0;
sentinel.previous_time = mstime();
sentinel.running_scripts = 0;
sentinel.scripts_queue = listCreate();
sentinel.announce_ip = NULL;
sentinel.announce_port = 0;
sentinel.simfailure_flags = SENTINEL_SIMFAILURE_NONE;
sentinel.deny_scripts_reconfig = SENTINEL_DEFAULT_DENY_SCRIPTS_RECONFIG;
sentinel.sentinel_auth_pass = NULL;
sentinel.sentinel_auth_user = NULL;
sentinel.resolve_hostnames = SENTINEL_DEFAULT_RESOLVE_HOSTNAMES;
sentinel.announce_hostnames = SENTINEL_DEFAULT_ANNOUNCE_HOSTNAMES;
memset(sentinel.myid,0,sizeof(sentinel.myid));
server.sentinel_config = NULL;
}
启动哨兵实例
sentinelIsRunning
sentinelIsRunning函数在sentinel.c文件中,用于启动哨兵实例:
- 校验是否设置了哨兵实例的ID,如果未设置,将随机生成一个ID
- 调用sentinelGenerateInitialMonitorEvents向监控的主节点发送+monitor事件
void sentinelIsRunning(void) {
int j;
/* 校验myid是否为0 */
for (j = 0; j < CONFIG_RUN_ID_SIZE; j++)
if (sentinel.myid[j] != 0) break;
if (j == CONFIG_RUN_ID_SIZE) {
/* 随机生成ID */
getRandomHexChars(sentinel.myid,CONFIG_RUN_ID_SIZE);
sentinelFlushConfig();
}
serverLog(LL_WARNING,"Sentinel ID is %s", sentinel.myid);
/* 向监控的主节点发送+monitor事件 */
sentinelGenerateInitialMonitorEvents();
}
/* 向监控的主节点发布事件 */
void sentinelGenerateInitialMonitorEvents(void) {
dictIterator *di;
dictEntry *de;
// 获取监控的主节点
di = dictGetIterator(sentinel.masters);
while((de = dictNext(di)) != NULL) {
sentinelRedisInstance *ri = dictGetVal(de);
// 向主节点发送监控事件
sentinelEvent(LL_WARNING,"+monitor",ri,"%@ quorum %d",ri->quorum);
}
dictReleaseIterator(di);
}
哨兵时间事件
在initServer函数中,调用aeCreateTimeEvent注册了时间事件,周期性的执行serverCron函数,serverCron函数中通过server.sentinel_mode判断是否是哨兵模式,如果是哨兵模式,调用sentinelTimer执行哨兵事件:
void initServer(void) {
// 省略...
if (aeCreateTimeEvent(server.el, 1, serverCron, NULL, NULL) == AE_ERR) {
serverPanic("Can't create event loop timers.");
exit(1);
}
// 省略...
}
int serverCron(struct aeEventLoop *eventLoop, long long id, void *clientData) {
// 省略...
// 如果是哨兵模式,调用sentinelTimer执行哨兵事件
if (server.sentinel_mode) sentinelTimer();
// 省略...
}
sentinelTimer
sentinelTimer在sentinel.c文件中,sentinelTimer函数会周期性的执行:
void sentinelTimer(void) {
sentinelCheckTiltCondition();
// 处理RedisInstances,传入的参数是当前哨兵实例维护的主节点的哈希表,里面记录当前节点监听的主节点
sentinelHandleDictOfRedisInstances(sentinel.masters);
sentinelRunPendingScripts();
sentinelCollectTerminatedScripts();
sentinelKillTimedoutScripts();
// 调整sentinelTimer的执行频率
server.hz = CONFIG_DEFAULT_HZ + rand() % CONFIG_DEFAULT_HZ;
}
sentinelHandleDictOfRedisInstances
sentinelHandleDictOfRedisInstances函数中会对传入的当前哨兵实例监听的主节点哈希表进行遍历:
- 获取哈希表中的每一个节点,节点类型是sentinelRedisInstance
- 调用sentinelHandleRedisInstance检测哨兵监听节点的状态
- 如果sentinelHandleRedisInstance是主节点,由于主节点里面保存了监听该主节点的其他哨兵实例以及从节点,所以递归调用sentinelHandleDictOfRedisInstances对其他的节点进行检测
/* sentinelHandleDictOfRedisInstances */
void sentinelHandleDictOfRedisInstances(dict *instances) {
dictIterator *di;
dictEntry *de;
sentinelRedisInstance *switch_to_promoted = NULL;
di = dictGetIterator(instances);
// 遍历所有的sentinelRedisInstance实例
while((de = dictNext(di)) != NULL) {
// 获取每一个sentinelRedisInstance
sentinelRedisInstance *ri = dictGetVal(de);
// 调用sentinelHandleRedisInstance检测哨兵监听节点的状态
sentinelHandleRedisInstance(ri);
// 如果是sentinelRedisInstance是主节点,主节点里面保存了监听该主节点的其他哨兵实例以及从节点
if (ri->flags & SRI_MASTER) {
// 递归调用,处理从节点
sentinelHandleDictOfRedisInstances(ri->slaves);
// 递归调用,处理其他哨兵实例
sentinelHandleDictOfRedisInstances(ri->sentinels);
if (ri->failover_state == SENTINEL_FAILOVER_STATE_UPDATE_CONFIG) {
switch_to_promoted = ri;
}
}
}
if (switch_to_promoted)
sentinelFailoverSwitchToPromotedSlave(switch_to_promoted);
dictReleaseIterator(di);
}
检测哨兵监听的节点状态
sentinelHandleRedisInstance
sentinelHandleRedisInstance函数对传入的sentinelRedisInstance实例,进行状态检测,主要处理逻辑如下:
- 调用sentinelReconnectInstance对实例的连接状态进行判断,如果**连接中断尝试重新与实例建立连接 **
- 调用sentinelSendPeriodicCommands向实例发送PING INFO等命令
- sentinelCheckSubjectivelyDown判断实例是否主观下线
- 如果实例是master节点,调用sentinelCheckObjectivelyDown判断是否客观下线、是否需要执行故障切换
/* Perform scheduled operations for the specified Redis instance. */
void sentinelHandleRedisInstance(sentinelRedisInstance *ri) {
// 如果监听的节点连接中断,尝试重新建立连接
sentinelReconnectInstance(ri);
// 发送PING INFO等命令
sentinelSendPeriodicCommands(ri);
// 检查是否是TILT模式
if (sentinel.tilt) {
if (mstime()-sentinel.tilt_start_time < SENTINEL_TILT_PERIOD) return;
sentinel.tilt = 0;
sentinelEvent(LL_WARNING,"-tilt",NULL,"#tilt mode exited");
}
// 判断主观下线
sentinelCheckSubjectivelyDown(ri);
/* Masters and slaves */
if (ri->flags & (SRI_MASTER|SRI_SLAVE)) {
}
// 如果是master节点
if (ri->flags & SRI_MASTER) {
// 判断客观下线
sentinelCheckObjectivelyDown(ri);
// 是否需要启动故障切换
if (sentinelStartFailoverIfNeeded(ri))
sentinelAskMasterStateToOtherSentinels(ri,SENTINEL_ASK_FORCED);
// 执行故障切换
sentinelFailoverStateMachine(ri);
// 获取其他哨兵实例对master节点状态的判断
sentinelAskMasterStateToOtherSentinels(ri,SENTINEL_NO_FLAGS);
}
}
重新连接
sentinelReconnectInstance
sentinelReconnectInstance函数用于检测实例的连接状态,如果中断进行重连,主要处理逻辑如下:
-
检测连接是否中断,如果未中断直接返回
-
检查端口是否为0,0被认为是不合法的端口
-
从sentinelRedisInstance实例中获取instanceLink,instanceLink的定义在sentinel.c文件中,里面记录了哨兵和主节点的两个连接,分别为用于发送命令的连接cc和用于发送Pub/Sub消息的连接pc:
typedef struct instanceLink { int refcount; int disconnected; int pending_commands; redisAsyncContext *cc; /* 用于发送命令的连接 */ redisAsyncContext *pc; /* 用于发送Pub/Sub消息的连接 */ mstime_t cc_conn_time; /* cc 连接时间 */ mstime_t pc_conn_time; /* pc 连接时间 */ mstime_t pc_last_activity; mstime_t last_avail_time; /* 上一次收到实例回复PING命令(需要被认定为合法)的时间 */ mstime_t act_ping_time; /* 当收到PONG消息的时候会设置为0,在下次发送PING命令时设置为当前时间 */ mstime_t last_ping_time; /* 上次发送PING命令时间,在出现故障时可以通过判断发送时间避免多次发送PING命令 */ mstime_t last_pong_time; /* 上次收到PONG消息的时间 */ mstime_t last_reconn_time; /* 上次执行重连的时间 */ } instanceLink;
-
校验距离上次重连时间是否小于PING的检测周期SENTINEL_PING_PERIOD,如果小于说明距离上次重连时间过近,直接返回即可
SENTINEL_PING_PERIOD在server.c中定义,默认1000毫秒
#define SENTINEL_PING_PERIOD 1000
-
对用于发送命令的连接判断,如果连接为NULL,调用redisAsyncConnectBind函数进行重连
-
对用于处理发送 Pub/Sub 消息的连接进行判断,如果连接为NULL,调用redisAsyncConnectBind函数进行重连
void sentinelReconnectInstance(sentinelRedisInstance *ri) {
// 检查连接是否中断
if (ri->link->disconnected == 0) return;
if (ri->addr->port == 0) return; /* 检查端口是否为0,0被认为是不合法的端口 */
// 获取instanceLink
instanceLink *link = ri->link;
mstime_t now = mstime();
// 校验距离上次重连时间是否小于哨兵PING的周期设置
if (now - ri->link->last_reconn_time < SENTINEL_PING_PERIOD) return;
ri->link->last_reconn_time = now;
/* 处理用于发送命令的连接 */
if (link->cc == NULL) {
// 进行连接
link->cc = redisAsyncConnectBind(ri->addr->ip,ri->addr->port,NET_FIRST_BIND_ADDR);
if (link->cc && !link->cc->err) anetCloexec(link->cc->c.fd);
// 省略...
}
/* 处理用于发送 Pub/Sub 消息的连接 */
if ((ri->flags & (SRI_MASTER|SRI_SLAVE)) && link->pc == NULL) {
link->pc = redisAsyncConnectBind(ri->addr->ip,ri->addr->port,NET_FIRST_BIND_ADDR);
if (link->pc && !link->pc->err) anetCloexec(link->pc->c.fd);
// 省略...
}
if (link->cc && (ri->flags & SRI_SENTINEL || link->pc))
link->disconnected = 0;
}
发送命令
sentinelSendPeriodicCommands
sentinelSendPeriodicCommands用于向实例发送命令:
void sentinelSendPeriodicCommands(sentinelRedisInstance *ri) {
mstime_t now = mstime();
mstime_t info_period, ping_period;
int retval;
// 省略...
/* 向主节点和从节点发送INFO命令 */
if ((ri->flags & SRI_SENTINEL) == 0 &&
(ri->info_refresh == 0 ||
(now - ri->info_refresh) > info_period))
{
// 发送INFO命令
retval = redisAsyncCommand(ri->link->cc,
sentinelInfoReplyCallback, ri, "%s",
sentinelInstanceMapCommand(ri,"INFO"));
if (retval == C_OK) ri->link->pending_commands++;
}
if ((now - ri->link->last_pong_time) > ping_period &&
(now - ri->link->last_ping_time) > ping_period/2) {
// 向实例发送PING命令
sentinelSendPing(ri);
}
/* PUBLISH hello messages to all the three kinds of instances. */
if ((now - ri->last_pub_time) > SENTINEL_PUBLISH_PERIOD) {
// 发送PUBLISH命令
sentinelSendHello(ri);
}
}
主观下线
sentinelCheckSubjectivelyDown
sentinelCheckSubjectivelyDown函数用于判断是否主观下线。
标记主观下线的两个条件
- 距离上次发送PING命令的时长超过了down_after_period的值,down_after_period的值在sentinel.conf 配置文件中配置,对应的配置项为down-after-milliseconds ,默认值30s
- 哨兵认为实例是主节点(ri->flags & SRI_MASTE),但是实例向哨兵返回的角色是从节点(ri->role_reported == SRI_SLAVE) 并且当前时间-实例报告消息的时间role_reported_time大于down_after_period加上SENTINEL_INFO_PERIOD乘以2的时间 ,SENTINEL_INFO_PERIOD 是发送INFO命令的时间间隔,也就是说实例上次成功向哨兵报告角色的时间,已经超过了限定时间(down_after_period加上SENTINEL_INFO_PERIOD*2)
满足以上两个条件之一哨兵将会把sentinelRedisInstance判断为主观下线,flag标记会添加SRI_S_DOWN状态。
void sentinelCheckSubjectivelyDown(sentinelRedisInstance *ri) {
mstime_t elapsed = 0;
// 如果act_ping_time不为0
if (ri->link->act_ping_time)
elapsed = mstime() - ri->link->act_ping_time; // 计算距离上次发送PING命令的间隔时间
else if (ri->link->disconnected) // 如果连接断开
elapsed = mstime() - ri->link->last_avail_time; // 计算距离最近一次收到PING命令回复的间隔时间
if (ri->link->cc &&
(mstime() - ri->link->cc_conn_time) >
SENTINEL_MIN_LINK_RECONNECT_PERIOD &&
ri->link->act_ping_time != 0 &&
(mstime() - ri->link->act_ping_time) > (ri->down_after_period/2) &&
(mstime() - ri->link->last_pong_time) > (ri->down_after_period/2))
{
instanceLinkCloseConnection(ri->link,ri->link->cc);
}
if (ri->link->pc &&
(mstime() - ri->link->pc_conn_time) >
SENTINEL_MIN_LINK_RECONNECT_PERIOD &&
(mstime() - ri->link->pc_last_activity) > (SENTINEL_PUBLISH_PERIOD*3))
{
instanceLinkCloseConnection(ri->link,ri->link->pc);
}
/*
* 标记主观下线的两个条件(或的关系)
* 1) 距离上次发送PING命令的时长超过了down_after_period
* 2) 哨兵认为实例是主节点(ri->flags & SRI_MASTE),但是实例向哨兵返回的角色是从节点(ri->role_reported == SRI_SLAVE) 并且当前时间-实例返回消息的时间大于down_after_period加上SENTINEL_INFO_PERIOD*2的时间 */
if (elapsed > ri->down_after_period ||
(ri->flags & SRI_MASTER &&
ri->role_reported == SRI_SLAVE &&
mstime() - ri->role_reported_time >
(ri->down_after_period+SENTINEL_INFO_PERIOD*2)))
{
/* 主观下线 */
if ((ri->flags & SRI_S_DOWN) == 0) {
// 发送+sdown事件
sentinelEvent(LL_WARNING,"+sdown",ri,"%@");
ri->s_down_since_time = mstime();
ri->flags |= SRI_S_DOWN; // 更改状态
}
} else {
/* Is subjectively up */
if (ri->flags & SRI_S_DOWN) {
sentinelEvent(LL_WARNING,"-sdown",ri,"%@");
ri->flags &= ~(SRI_S_DOWN|SRI_SCRIPT_KILL_SENT);
}
}
}
客观下线
如果是主节点,将会调用sentinelCheckObjectivelyDown函数判断客观下线,之后调用sentinelStartFailoverIfNeeded判断是否需要执行故障切换。
总结
参考
Redis版本:redis-6.2.5