Shulou(Shulou.com)06/03 Report--

I. related concepts of Redis cluster

Redis has supported cluter since version 3.0, using hash slots, which can consolidate multiple Redis instances together to form a cluster, that is, data can be stored on multiple nodes in the cluster.

The cluster of Redis is an acentric structure. In the cluster, the identity of each master is equal, each node saves the data and the state of the entire cluster, and knows the slot that the other node is responsible for, and sends heartbeat information regularly, can timely detect the abnormal node in the cluster, and take the way of voting to determine whether the node is unavailable, if the number of votes is more than half of the nodes in the cluster. It is considered that the node is not available, and it is precisely because of this feature that a Redis cluster should be deployed with at least three or more nodes.

The cluster role has slots (slots) assigned between master and slave,master, and the slot number is 0-16383 (16384 in total).

By default, each cluster node has two TCP ports listening, one is 6379 (for listening for client access connections) and the other is 16379 (for node communication between clusters). Note that the firewall needs to release traffic from these two ports.

All data in Redis is stored in memory and then saved to disk asynchronously from time to time (this is called "semi-persistent mode"); each data change can also be written to an append only file (aof) (this is called "full persistence mode").

There are two ways of persistence provided by Redis, one is RDB persistence (the principle is that the database in memory of Redis records dump regularly to RDB on disk), and the other is AOF (append only file) persistence (the principle is that the operation log of Redis is written to the file in an additional way).

Advantages and disadvantages of RDB

Advantages of RDB semi-persistence:

Include only one file, which is good for file backup; disaster recovery is faster than aof persistence; and performance is maximized. For the Redis service process, at the beginning of persistence, the only thing it needs to do is to fork the child process, and then the child process completes the persistence work, which can greatly prevent the service process from performing IO operations. Compared with the AOF mechanism, if the dataset is too large, the RDB startup efficiency will be higher. Disadvantage of RDB semi-persistence: because RDB assists in data persistence through the help subprocess, if the dataset is large, it may cause the entire service to stop for hundreds of milliseconds, or even 1 second.

Advantages and disadvantages of AOF

Advantages of AOF full persistence:

It can ensure high availability of data; when downtime occurs during writing, it will not destroy the content already in the log file. If there is a downtime during writing, you can restart Redis and use redis-check-aof tool to solve this problem. If the log is too large, Redis can automatically enable rewrite mechanism to generate new files to store aof logs. This mechanism can bring higher data security and data persistence. Three synchronization strategies are provided in Redis, namely synchronization per second, synchronization per modification, and async. The disadvantage of AOF full persistence: AOF files are usually larger than RDB files for the same number of datasets. The recovery speed of RDB is faster than that of AOF when recovering large datasets; according to the different synchronization strategies, AOF tends to be slower than RDM in running efficiency. In a word, the efficiency of synchronization per second policy is relatively high, and the efficiency of synchronization disable policy is as efficient as RDB.

If both RDB and AOF exist at the same time, the AOF mode is preferred.

II. Deploy Redis clusters

1. The environment is as follows:

There are a total of six centos servers, with three master corresponding to one slave

In order to avoid the collapse of the entire cluster due to the downtime of the physical server, you can also configure multiple Redis instances on one server to achieve cross-master-slave replication. The so-called crossover is master in node01, but the corresponding slave master corresponds to slave in node03, while node03 corresponds to slave in node01.

The above is just a Redis cluster deployed on multiple servers.

But I am mainly deploying multiple nodes on a single server, so there will be multiple instances on node6

You can download the required software package before configuration, or you can download and prepare it by yourself. Link: https://pan.baidu.com/s/1mJGBD7b0QyR_II4mSpdD8Q

Extraction code: 72hi

2. Configure Redis instance

Here is an example of the configuration of node01. The configuration of other nodes is basically the same as that of node1.

[root@node1 /] # mkdir redis # personal habit [root@node1 /] # cd redis/ [root@node1 redis] # rz # upload the required software package [root@node1 redis] # ls # node node only needs the redis4.0 package The remaining two packages are redis-3.3.0.gem redis-4.0.14.tar.gz ruby-2.3.1.tar.gz [root@node1 redis] # tar zxf redis-4.0.14.tar.gz # unpacking [root@node1 redis] # mv redis-4.0.14 / usr/local/redis # Mobile and renaming [root@node1 redis] # cd / usr/local/ required for installing the ruby environment Redis/ # enter the directory [root@node1 redis] # make & & make install # compile and install [root@node1 redis] #. / utils/install_server.sh # initialize redis # all options for initialization remain default Enter all the way to confirm, is to confirm the listening port, configuration files, log files, pid storage Is this ok? Then press ENTER to go on or Ctrl-C to abort.Copied / tmp/6379.conf = > / etc/init.d/redis_6379Installing service...Successfully added to CHK configs successful added to runlevels 345 starting Redis server...Installation successful! Path and other information # when these messages appear, the initialization is successful [root@node1 redis] # echo "512" > / proc/sys/net/core/somaxconn # and then make some optimizations [root@node1 redis] # echo "vm.overcommit_memory = 1" > / etc/sysctl.conf [root@node1 redis] # sysctl-pvm.overcommit_memory = 1 [root@node1 redis] # echo "never" > / sys/ Kernel/mm/transparent_hugepage/enabled [root@node1 redis] # vim / etc/redis/6379.conf # Edit the configuration file bind 0.0.0.0 # find the line that is not commented Change it to 0.0.0.0daemonize yes # if there is a comment symbol, you need to delete the comment symbol in order to take effect cluster-enabled yes # if there are any comments You need to remove the comment symbol cluster-node-timeout 5000 # modify appendonly yes # and change it to yes [root@node1 redis] # / etc/init.d/redis_6379 restart # restart the service to make the configuration effective [root@node1 redis] # netstat-anput | grep 6379 # confirm that ports 6379 and 16379 tcp 00 0.0.0.0yes 6379 have been monitored 0.0.0.0 * LISTEN 44263/redis-server tcp 0 0 0.0.0.0pur16379 0.0.0.0purl * LISTEN 44263/redis-server

You can configure the above in turn on other node servers, except for node6

3. Configure multiple Redis instances of node06 hosts

On the node06 node, I will configure it to run multiple Redis database instances, so it is not exactly the same as the configuration of the first five nodes, so configure it carefully

[root@node6 /] # mkdir redis [root@node6 /] # cd redis/ [root@node6 redis] # rz [root@node6 redis] # tar zxf redis-4.0.14.tar.gz [root@node6 redis] # mv redis-4.0.14 / usr/local/redis [root@node6 redis] # cd / usr/local/redis/ [root@node6 redis] # make & & make install # compile and install After installation, there is no need to initialize [root@node6 redis] # redis-server # to start this command, there will be 3 warnings. These three warnings are the three options we need to optimize: [root@node6 redis] # echo "512" > / proc/sys/net/core/somaxconn [root@node6 redis] # echo "vm.overcommit_memory = 1" > > / etc/sysctl.conf [root@node6 redis] # sysctl-pvm.overcommit_memory = 1 [root@node6 redis] # echo "never" > / sys/kernel/mm/transparent_hugepage/ enabled [root @ node6 redis] # vim redis.conf # Edit configuration file bind 0.0.0.0port 7000 # the default port for redis is 6379 Here I changed it to 7000, in order to identify daemonize yes # to enable the background daemon to run cluster-enabled yes # in the background to turn on the timeout between cluster cluster-node-timeout 5000 # cluster nodes, in millisecond appendonly yes # whether to enable synchronization to disk appendfilename "appendonly-7000.aof" # aof log name, because there are multiple instances on this host So each port is different, so use the port to name the log name [root@node6 redis] # mkdir-p / usr/local/redis-cluster/ {7000.7004} # above is intended to run several Redis instances, just create a few directories Here I name the directory after the port number of the instance (temporarily planning to run 4 Redis instances) # the following is to copy the modified configuration file to the specified directory: [root@node6 redis] # cp redis.conf / usr/local/redis-cluster/7000/ [root@node6 redis] # cp redis.conf / usr/local/redis-cluster/7001/ [root@node6 redis] # cp redis.conf / usr/local/redis-cluster/7002/ [root @ node6 redis] # cp redis.conf / usr/local/redis-cluster/7003/ [root@node6 redis] # cp redis.conf / usr/local/redis-cluster/7004/# then change each copied configuration file to the port number corresponding to the Redis instance [root@node6 redis] # cd / usr/local/redis-cluster/ [root@node6 redis-cluster] # ls7000 7001 7002 7003 7004 [root@node6 redis-cluster] # sed-is / 7000and 7001 / g 7001/redis.conf [root@node6 redis-cluster] # sed-is / 7000is 7002 7002/redis.conf [root@node6 redis-cluster] # sed-is / 7000is 7003 root@node6 redis-cluster g 7003/redis.conf [root@node6 redis-cluster] # sed-is / 7004root@node6 redis-cluster 7004/redis.conf # cd 7000 / # enter the 7000 instance directory [root@node6 7000] # redis-server redis.conf # launch the instance 8142 Feb C 13 Feb 10 Feb 44 Feb 17.501 # Redis version=4.0.14 Bits=64, commit=00000000, modified=0, pid=8142, and just started8142:C 13 Feb 10 modified=0 44 17 50 1 # Configuration loaded# are entered into each instance directory in turn Start the instance, so I won't demonstrate it here.

4. Host node01 installs and configures the running environment of ruby to facilitate the management of Redis clusters

[root@node1 /] # yum-y install rpm-build openssl openssl-devel # installation dependency package [root@node1 /] # cd / redis/ [root@node1 redis] # lsredis-3.3.0.gem redis-4.0.14.tar.gz ruby-2.3.1.tar.gz [root@node1 redis] # tar zxf ruby-2.3.1.tar.gz # Unpack [root@node1 redis] # cd ruby-2.3 .1 / [root@node1 ruby-2.3.1] #. / configure-- prefix=/usr/local/ruby & & make & & make install # this environment takes a long time to compile [root@node1 ruby-2.3.1] # ln-s / usr/local/ruby/bin/* / usr/local/bin/ # softlink for commands [root@node1 ruby-2.3.1] # ln-s / usr/local / redis/src/redis-trib.rb / usr/local/bin/ [root@node1 ruby-2.3.1] # ln-s / usr/local/ruby/bin/* / usr/bin/ [root@node1 ruby-2.3.1] # cd. [root@node1 redis] # gem install redis-3.3.0.gem # return to the directory with the gem file Install Successfully installed redis-3.3.0Parsing documentation for redis-3.3.0Installing ri documentation for redis-3.3.0Done installing documentation for redis after 0 seconds1 gem installed# range if the above information indicates that the installation is successful

5. Configure each node in the cluster

[root@node1 /] # redis-cli-p 6379 # Log in to the ben'di local instance 127.0.0.1 CLUSTER MEET 192.168.171.134 6379 # add each node to the cluster OK127.0.0.1:6379 > CLUSTER MEET 192.168.171.135 6379OK127.0.0.1:6379 > CLUSTER MEET 192.168.171.136 6379OK127.0.0.1:6379 > CLUSTER MEET 192.168.171.137 6379OK127 .0.0.1: 6379 > CLUSTER MEET 192.168.171.138 7000 # this is because node6 modified the port number OK127.0.0.1:6379 > set test aaa # because no hash slot is assigned So you cannot add data (error) CLUSTERDOWN Hash slot not served 127.0.0.1 CLUSTER INFO 6379 > CLUSTER INFO # check the status of the cluster cluster_state:fail # and you will find that here is the cluster_slots_assigned:0127.0.0.1:6379 of fail failure > exit# below we assign slot points, be careful to configure # although you can replace "add" with "del" in the command But I didn't succeed # must allocate 0 to 16383 completely, preferably equally # just assign it to the node as master My node01 to node03 is master [root@node1 /] # redis-cli-h 192.168.171.133-p 6379 cluster addslots {0.5461} OK [root@node1 /] # redis-cli-h 192.168.171.134-p 6379 cluster addslots {5462... 10922} OK [root@node1 /] # redis-cli-h 192.168.171.135-p 6379 cluster addslots {10923. 16383} OK [root@node1 /] # redis-cli-p 6379-c # enter the cluster Need to add-c option 127.0.0.1 CLUSTER NODES # to view cluster node information # the next step is to bind each slave slave node to master # node04 as the slave node of node01, node05 as the slave node of node02, and 6379 instance of node06 as the slave node of node03 # you need to log in to which instance 127.0.0.1 node04 6379 > CLUSTER NODES # can execute this command first View the ID of the corresponding node So that you can next specify [root@node1 /] # redis-cli-h 192.168.171.136-p 6379 # Log in to node4192.168.171.136:6379 > CLUSTER REPLICATE 4397bb3e5e43fdf1700551d0a2587f502e6c02a8OK# and add the node id192.168.171.136:6379 > exit # of node1 to exit the node [root@node1 /] # redis-cli-h 192.168.171.137-p 6379192.168.171.137 > CLUSTER REPLICATE 1eca51f415ba5194a95d5400176daa5713d64990OK# Is the node id192.168.171.137:6379 > exit [root@node1 /] # redis-cli-h 192.168.171.138-p 7000192.168.171.138redis-cli 7000 > CLUSTER REPLICATE ee71a2fd995f35992ce69328096ccd50fc68f89eOK#. Above is the node id of node3.

At this point, the cluster can read and write data normally, as follows:

[root@node1 /] # redis-cli-h 192.168.171.133-p 6379-c 192.168.171.133 located at 192.168.171.134:6379OK192.168.171.134:6379 6379 > set test aaa- > Redirected to slot [6918] located at 192.168.171.134:6379OK192.168.171.134:6379 > get test "aaa" 192.168.171.134Redirected to slot 6379 >

6. Use the commands installed by ruby to manage Redis clusters

[root@node1 /] # redis-trib.rb check 127.0.0.1:6379

7. Add the 7001 instance of node06 to the Redi cluster

Add nod

[root@node1 /] # redis-trib.rb add-node 192.168.171.138master 7001 192.168.171.133master when adding nodes, no other configuration is added. After joining the cluster, the role is by default

After adding nodes, you need to assign the corresponding slot points.

Because for a cluster to function properly, all slot points must be allocated, so when new nodes join, you need to reassign slot points to new nodes, as follows:

[root@node01 /] # redis-trib.rb check 127.0.0.1 master 6379 # execute this command to confirm the newly joined node, master [root@node01 /] # redis-trib.rb reshard 192.168.171.133master 6379 # specify the cluster address and port How many slots do you want to move (from 1 to 16384)? 409 master, then the average is 4096, so enter 4096What is the receiving node ID? Eca1db0392f69207c4fc3ad4ff290d5a94ea5f99# specifies the ID Source node # 1:all # of the receiving node, that is, the instance of 7001 on the newly joined node06 host, specifies the slot allocation from which node, enter "all" to select all nodes Do you want to proceed with the proposed reshard plan (yes/no)? Yes # enter "yes" to confirm

At this point, the new node is added and the corresponding slot point is assigned, but there is no slave node yet, so the new master is assigned a slave node next.

8. Assign slave nodes to the newly joined master

There are two ways to assign slave nodes, one is to automatically bind to the master that does not specify which master the slave node is, and the other is to specify which master to bind to directly, both of which are written down here

Method 1:

[root@node1 /] # redis-trib.rb add-node-- slave 192.168.171.138root@node1 7002 192.168.171.138Viru 700mm add 7002 instances on node06 to the cluster as slave # Note: the returned information must not be red, which means there is an error [root@node1 /] # redis-trib.rb check 192.168.171.1336379

Method 2:

[root@node1 /] # redis-trib.rb add-node-- slave-- master-id eca1db0392f69207c4fc3ad4ff290d5a94ea5f99 192.168.171.138Viru 7003 192.168.171.138slave 7001

# the above id is the id of 7001master

[root@node1 /] # redis-trib.rb check 192.168.171.133: 6379 # check the cluster status and find that 7001 has two slave nodes

9. Delete the master node operation

To delete the master node is to reverse the operation of adding the master node. The slot points on the master node to be deleted need to be assigned to other master before the deletion operation can be performed. After deleting the master node, the corresponding slave of the master will also be transferred to the new master along with the slots slot.

Take the removal of 7001 instances on node06 as an example

[root@node1 /] # redis-trib.rb reshard 192.168.171.133:6379How many slots do you want to move (from 1 to 16384)? 4096 # specify how many slots to delete What is the receiving node ID? after the id added above is used to delete the master node, to whom the remaining slots will be assigned. Here I add the idPlease enter all the source node IDs of node2. Type 'all' to use all the nodes as source nodes for the hash slots. Type 'done' once you entered all the source nodes IDs.Source node # 1:eca1db0392f69207c4fc3ad4ff290d5a94ea5f99# the node specified above is the node to be deleted. Here is 7001Source node # 2:done # on node6 for ending Do you want to proceed with the proposed reshard plan (yes/no)? After the yes# removes the slot, remove the 7001 instance from the cluster as follows: [root@node1 /] # redis-trib.rb del-node 192.168.171.133 eca1db0392f69207c4fc3ad4ff290d5a94ea5f99 > Removing node eca1db0392f69207c4fc3ad4ff290d5a94ea5f99 from cluster 192.168.171.133 Removing node eca1db0392f69207c4fc3ad4ff290d5a94ea5f99 from cluster specifies an id of 7001 > Sending CLUSTER FORGET messages to the cluster... > SHUTDOWN the node.

At this point, 7001 instances have been completely removed from the cluster, and the corresponding slave has also become the slave of node02 as the slot point is transferred to node02. Now if you look at the cluster information, the master of node02 should correspond to three slave

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.