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

Redis

It's a key-value storage system. Similar to Memcached, it supports relatively more value types, including string (string), list (linked list), set (collection), zset (sorted set-ordered collection), and hash (hash type). These data types support push/pop, add/remove, and take intersection union and difference sets, and richer operations, and these operations are atomic. On this basis, redis supports a variety of different sorting methods. Like memcached, data is cached in memory for efficiency. The difference is that redis will periodically write updated data to disk or modify operations to additional record files, and on this basis to achieve master-slave (master-slave) synchronization.

[1] case one phenomenon:

The production system has just started to run, and the system is stable. However, after running for a period of time, it is found that the response of the system interface becomes slower in part of the time. The following error often occurs when viewing the client log:

Redis.clients.jedis.exception.JedisConnectionException:java.net.SocketTimeoutException:Read time out

Problem location: execute slowlog to view the slow query log and find a large number of keys command operations. The performance of keys command is very poor in a large number of concurrent cases. In the production environment, try to avoid using keys, and then find out the code that uses keys to optimize until the time out problem is solved.

192.168.17.46 slowlog get 1) 1) (integer) 22 2) (integer) 1563344158 3) (integer) 10193 4) 1) "SET" 2) "getBatchChapterFiles" 3)\ x0b\ xfa\ 529:\ t489761532B\ x02-1J\ t48976181... "1) (integer) 21 2) (integer) 1545403066 3) (integer) 10915 4) 1)" GET "2)" getVolumeChapters#data "

[2] case 2 phenomenon:

After the production environment has been running for a long time, there is often a failure of the interface to return data, or it is found that the pressure on the database increases sharply at a certain time from the monitoring. Check the client log and find the following error:

Redis.clients.jedis.exceptions.JedisConnectionException:Cloud not get a resource from the pool

Found an error in the redis log:

[2489] 02 Jun 10:43:42 # Error allocating resoures for the client

Problem location: execute the client list command and find that the idle time of a large number of client is very long. Check the configuration and find that both timeout and tcp-keepalive (heartbeat detection) are enabled (both 0), and the Redis server has no effective mechanism to ensure whether the server connection has failed. When a flash disconnect occurs between the server and the client network, resulting in the interruption of tcp, the client in this case will always be held by the redis server, and there will be client connections with long idle (idle) time.

Solution: set up timeout and tcp-keepalive to clean up failed connections.

Redis/bin > redis-cli-h 192.168.17.46-p 6386 info Clients# Clientsconnected_clients:5000-larger client_longest_output_list:0client_biggest_input_buf:0blocked_clients:0192.168.17.46:6386 > CONFIG GET timeout 1) "timeout" 2) "0" 192.168.17.46 info Clients# Clientsconnected_clients:5000 6386 > CONFIG GET tcp-keepalive1) "tcp-keepalive" 2) "0" 192.168.17.46Viru 6386 > client listid=612260747 addr=192.168.17.92:53069 fd=806 name= age=114 idle=21 flags=N db=0 sub=0 psub=0 multi=-1 qbuf=0 qbuf-free=0 obl=0 oll=0 omem=0 events=r cmd=pingid=612260593 addr=192.168.41.44:38248 fd=381 name= age=131 idle=61 flags=N db=0 sub=0 psub=0 multi=-1 qbuf=0 qbuf-free=0 obl=0 oll=0 omem=0 events=r cmd=get

Field definition

Addr: address and port of the client

Fd: the file descriptor used by the socket

Age: connected duration in seconds

Idle: idle time in seconds

Flags: client flag

Db: the database ID that the client is using

Sub: number of subscribed channels

Psub: number of subscribed patterns

Multi: the number of commands executed in the transaction

Qbuf: the length of the query buffer (in bytes, 0 means no query buffer is allocated)

Qbuf-free: the length of the remaining space in the query buffer (in bytes, 0 means there is no remaining space)

Obl: the length of the output buffer (in bytes, 0 means no output buffer is allocated)

Oll: the number of objects contained in the output list (when there is no space left in the output buffer, the command reply will be queued in the form of a string object)

Omem: total memory consumed by output buffers and output lists

Events: file descriptor event

Cmd: the last command executed

[3] case three phenomena:

Redis suddenly cannot be accessed, and the following error is returned:

Redis.client.jedis.exception.JedisDataException:MISCONF Redis is configured to save RDB snapshots

But is currently not able to persist on disk.Commands that may modify the data set are disabled.

Please check Redis logs for details about the error

Problem location: check the redis log and find the following error: when Cant save in background:fork:Cannot allocate memory Redis saves the data in memory to disk, in order to prevent the main thread from faking death, it will Fork a child process to complete the save operation. The child process of this Fork needs to allocate the same memory as the main process, which is equivalent to doubling the required memory. If there is not enough available memory to allocate the required memory at this time, the Fork child process will fail and the data cannot be persisted to disk. The problem can be solved by modifying the Linux kernel parameter vm.overcommit_memeory=1, which means that the kernel allows all physical memory to be allocated, regardless of the current memory state.

192.168.17.46 home/redis02/redis/log/6386.log 6386 > CONFIG GET logfile1) "logfile" 2) / home/redis02/redis/log/6386.log "

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.