Redis Study Notes ~ Redis implements multiple slave storage servers for read/write splitting

Back to directory

Redis, A nosql storage system, is generally deployed in Linux. we can regard it as a data server. When concurrency is high, we will use multiple servers to act as redis servers, at this time, redis is also distributed, and redis and WWW are also distributed. We do not need to intervene in the distribution between redis, it is linked by our redis client. The server to which you are linked is completely controlled by the client.

Next, let's take a look at the related configurations in redisconfig. I have added some instructions.

/// <Summary> /// configuration parameters of redis main information /// </Summary> Public sealed class redisconfiginfo: configurationsection {public static redisconfiginfo getconfig () {redisconfiginfo section = (redisconfiginfo) configurationmanager. getsection ("redisconfig"); Return section;} public static redisconfiginfo getconfig (string sectionname) {redisconfiginfo section = (redisconfiginfo) configurationmanager. getsection ("R Edisconfig "); If (Section = NULL) throw new configurationerrorsexception (" section "+ sectionname +" is not found. "); Return section ;}/// <summary> /// writable redis link address, separate multiple hosts with commas // </Summary> [configurationproperty ("writeserverlist", isrequired = false)] Public String writeserverlist {get {return (string) base ["writeserverlist"];} set {base ["writeserverlist"] = value ;}/// <summary> // readable The redis link address. Multiple hosts are separated by commas. // </Summary> [configurationproperty ("readserverlist", isrequired = false)] Public String readserverlist {get {return (string) base ["readserverlist"];} set {base ["readserverlist"] = value ;}} /// <summary> /// maximum number of write links /// </Summary> [configurationproperty ("maxwritepoolsize", isrequired = false, defaultvalue = 5)] public int maxwritepoolsize {get {int _ maxwritepoolsize = (INT) base ["maxwritepoolsize"]; return _ maxwritepoolsize> 0? _ Maxwritepoolsize: 5 ;}set {base ["maxwritepoolsize"] = value ;}} /// <summary> /// maximum number of read links /// </Summary> [configurationproperty ("maxreadpoolsize", isrequired = false, defaultvalue = 5)] public int maxreadpoolsize {get {int _ maxreadpoolsize = (INT) base ["maxreadpoolsize"]; return _ maxreadpoolsize> 0? _ Maxreadpoolsize: 5 ;}set {base ["maxreadpoolsize"] = value ;}} /// <summary> /// Automatic Restart /// </Summary> [configurationproperty ("autostart", isrequired = false, defaultvalue = true)] public bool autostart {get {return (bool) base ["autostart"] ;}set {base ["autostart"] = value ;}} /// <summary> /// expiration time of the local cache (timeout), in the unit of seconds /// </Summary> [configurationproperty ("localcachetime", isrequired = false, defaultvalue = 36000)] public int localcachetime {get {return (INT) base ["localcachetime"];} set {base ["localcachetime"] = value ;}} /// <summary> /// whether to record logs. This setting is only used to troubleshoot redis running problems, such as redis working properly, disable this item // </Summary> [configurationproperty ("recordelog", isrequired = false, defaultvalue = false)] public bool recordelog {get {return (bool) base ["recordelog"];} set {base ["recordelog"] = value ;}}}

In the configuration file, we can configure the redis read server and write server. multiple servers are separated by commas (,) (redis itself supports read/write splitting)

  <RedisConfig WriteServerList="192.168.2.71:6379,192.168.2.72:6379" ReadServerList="192.168.2.71:6379,192.168.2.72:6379" MaxWritePoolSize="60" MaxReadPoolSize="60" AutoStart="true" LocalCacheTime="180" RecordeLog="false">  </RedisConfig>

We can use multiple threads to test the function.

// Testing redis 100 threads for (INT I = 0; I <100; I ++) {task. run () => {using (VAR redisclient = redismanager. getclient () {using (VAR test = redisclient. gettypedclient <string> () {test. lists ["print"]. enqueue ("information added"); var S = test. lists ["print"]. dequeue (); response. write (s); response. write (redisclient. host); response. write ("<br> ");}}});}

The test result is that a redis server is randomly allocated for each client request. Let's take a look.

The figure above shows that our information is written to two redis servers, and the problem arises again. When the client retrieves data, you may retrieve the data in redisserver1, it is also possible to take the data in redisserver2, and the real complete data is the union of the two servers. How to merge the data of multiple servers is a problem we need to solve, if this problem cannot be solved, it will be meaningless for multiple servers. In the next section, we will mainly discuss the data synchronization problem of multiple redisservers.

Back to directory

Redis Study Notes ~ Redis implements multiple slave storage servers for read/write splitting