redis的maxmemory设置

redis的maxmemory参数用于控制redis可使用的最大内存容量。如果超过maxmemory的值，就会动用淘汰策略来处理expaire字典中的键。

关于redis的淘汰策略，很多文章都有描述，就不做解释了。

关于maxmemory的设置，如果redis的应用场景是作为db使用，那不要设置这个选项，因为db是不能容忍丢失数据的。

如果作为cache使用，则可以启用这个选项（其实既然有淘汰策略，那就是cache了。。。）

但是在集群环境下（尤其是有多个slavers的情形），maxmeomory的值并不是实际redis使用的内存，这个选项值并没有包括slaver的output buffer。

redis早期版本出过一个bug，在多个slaver的情形下，设置了maxmemory值，同时设定了淘汰策略，会造成master上的数据被渐渐擦除。

antirez先生给出了这个问题的原因：

The issue happens for the following reason:

Redis reached the configured limit, so it tries to expire keys.
Evicting keys turns into explicit DELs sent to slaves, since masters control the eviction of slaves for well known reasons.
But this way if there are enough slaves, emitting the protocol in the output buffers will actually take more memory than the amount freed removing keys...
So the key eviction process starts to enter into an infinite loop.
Up to a given point the fact that there is a static buffer part in the output queue of every client (including slaves) mitigate this in certain conditions, but once Redis can't use the output buffer but must use the queue of objects the infinite loop is triggered.

简单说来，删除过期键，需要产生del命令发送给slaver，如果slaver足够多，output buffer将会占用足够多的内存，导致更多的键过期，如此往复，陷入了无线循环。

解决方案有多种，比如output buffer可以不计入maxmemory。

因此，在3.0版本的配置说明中有了以下表述：

# WARNING: If you have slaves attached to an instance with maxmemory on,
# the size of the output buffers needed to feed the slaves are subtracted
# from the used memory count, so that network problems / resyncs will
# not trigger a loop where keys are evicted, and in turn the output
# buffer of slaves is full with DELs of keys evicted triggering the deletion
# of more keys, and so forth until the database is completely emptied.
#
# In short... if you have slaves attached it is suggested that you set a lower
# limit for maxmemory so that there is some free RAM on the system for slave
# output buffers (but this is not needed if the policy is 'noeviction').
#
# maxmemory 

由此可见，如果有slaver的情况下，建议适当调低maxmemory，给output buffer留出一定的可用空间是合理的。