Ruby on rails Redis:：TimeoutError:连接超时错误-Rails缓存

我们正在使用redis（Elasticache）作为缓存存储。由于redis连接超时，我们的应用程序出现崩溃。我们有三台应用程序服务器，使用phusion passenger。在崩溃时，大约有200个活动连接

/home/chillr/deploy/chillr-api/shared/vendor/bundle/ruby/2.2.0/gems/redis-3.2.1/lib/redis/connection/ruby.rb:55:in `rescue in _read_from_socket'
/home/chillr/deploy/chillr-api/shared/vendor/bundle/ruby/2.2.0/gems/redis-3.2.1/lib/redis/connection/ruby.rb:48:in `_read_from_socket'
/home/chillr/deploy/chillr-api/shared/vendor/bundle/ruby/2.2.0/gems/redis-3.2.1/lib/redis/connection/ruby.rb:41:in `gets'
/home/chillr/deploy/chillr-api/shared/vendor/bundle/ruby/2.2.0/gems/redis-3.2.1/lib/redis/connection/ruby.rb:273:in `read'
/home/chillr/deploy/chillr-api/shared/vendor/bundle/ruby/2.2.0/gems/redis-3.2.1/lib/redis/client.rb:248:in `block in read'
/home/chillr/deploy/chillr-api/shared/vendor/bundle/ruby/2.2.0/gems/redis-3.2.1/lib/redis/client.rb:236:in `io'
/home/chillr/deploy/chillr-api/shared/vendor/bundle/ruby/2.2.0/gems/redis-3.2.1/lib/redis/client.rb:247:in `read'
/home/chillr/deploy/chillr-api/shared/vendor/bundle/ruby/2.2.0/gems/redis-3.2.1/lib/redis/client.rb:112:in `block in call'
/home/chillr/deploy/chillr-api/shared/vendor/bundle/ruby/2.2.0/gems/redis-3.2.1/lib/redis/client.rb:217:in `block (2 levels) in process'
/home/chillr/deploy/chillr-api/shared/vendor/bundle/ruby/2.2.0/gems/redis-3.2.1/lib/redis/client.rb:353:in `ensure_connected'
/home/chillr/deploy/chillr-api/shared/vendor/bundle/ruby/2.2.0/gems/redis-3.2.1/lib/redis/client.rb:207:in `block in process'
/home/chillr/deploy/chillr-api/shared/vendor/bundle/ruby/2.2.0/gems/redis-3.2.1/lib/redis/client.rb:292:in `logging'
/home/chillr/deploy/chillr-api/shared/vendor/bundle/ruby/2.2.0/gems/redis-3.2.1/lib/redis/client.rb:206:in `process'
/home/chillr/deploy/chillr-api/shared/vendor/bundle/ruby/2.2.0/gems/redis-3.2.1/lib/redis/client.rb:112:in `call'

SlOWLOG命令

1) 1) (integer) 6867375
2) (integer) 1486701507
3) (integer) 62008
4) 1) "eval"
   2) "for i, name in ipairs(redis.call('KEYS', 'cache:user_transaction_logs:*:9008245678')) do redis.call('DEL', name); end"
   3) "0"
 2) 1) (integer) 6867374
    2) (integer) 1486701507
    3) (integer) 61989
    4) 1) "KEYS"
       2) "cache:user_transaction_logs:*:9008245678"
 3) 1) (integer) 6867373
    2) (integer) 1486701507
    3) (integer) 61026
    4) 1) "eval"
       2) "for i, name in ipairs(redis.call('KEYS', 'cache:user_transaction_logs:*:8888662136')) do redis.call('DEL', name); end"
       3) "0"
 4) 1) (integer) 6867372
    2) (integer) 1486701507
    3) (integer) 61006
    4) 1) "KEYS"
       2) "cache:user_transaction_logs:*:8888662136"
 5) 1) (integer) 6867371
    2) (integer) 1486701507
    3) (integer) 63070
    4) 1) "eval"

我们使用

• 弹性缓存
• t2.3中等客运量

崩溃时的连接数175..200

宝石 redis（3.2.1） redis rails（5.0.1）

---

可以做什么

监视您的redis服务器的当前资源使用情况。 暂时增加连接超时redis connection

先试试这个检查表

---

在生产中不要使用密钥：但在这种情况下我们无法避免。我们希望从缓存中获取用户的所有事务。任何其他方法都可以尝试扫描或更改数据结构，以便直接查找用户事务，就像使用一个将用户id链接到事务id的集合一样，我们已经升级了我们的elasticache服务器