sysctl.conf

Add to /etc/sysctl.conf

nano /etc/sysctl.conf

In the end of file these lines:

net.ipv4.conf.default.rp_filter = 1
net.ipv4.conf.default.accept_source_route = 0
kernel.sysrq = 0
net.ipv4.tcp_syncookies = 0
kernel.msgmnb = 65536
kernel.msgmax = 65536
kernel.shmmax = 68719476736
kernel.shmall = 4294967296
vm.swappiness = 0
net.ipv4.tcp_fack = 1
net.ipv4.tcp_sack = 1
net.ipv4.tcp_mem = 8388608 12582912 16777216
net.ipv4.udp_mem = 8388608 12582912 16777216
net.ipv4.udp_rmem_min = 16384
net.ipv4.udp_wmem_min = 16384
net.core.wmem_max = 8388608
net.core.rmem_max = 8388608
net.ipv4.tcp_rmem = 8192 87380 8388608
net.ipv4.tcp_wmem = 8192 87380 8388608
net.ipv4.tcp_timestamps = 0
net.ipv4.tcp_window_scaling = 1
net.core.somaxconn = 300000
net.core.netdev_max_backlog = 8192
net.ipv4.tcp_max_syn_backlog = 2048
net.ipv4.tcp_keepalive_time = 180
net.ipv4.tcp_keepalive_probes = 5
net.ipv4.tcp_keepalive_intvl = 30
net.ipv4.tcp_tw_reuse = 1
net.ipv4.tcp_tw_recycle = 0
net.ipv4.tcp_max_tw_buckets = 1000000
net.ipv4.ip_local_port_range = 1024 65535
net.nf_conntrack_max = 1000000

Also, if you have in the redis.log error like:

# Can't save in background: fork: Cannot allocate memory

then add these line:

vm.overcommit_memory = 1

Count of opened files

Sometimes you can see in the redis.log that error:

#Too many open files

It signals you that you need to increase count of opened files in your server settings.

Look at the current settings of count of simultaneously open files:

cat /proc/sys/fs/file-max

Look at the current state of opened files:

cat /proc/sys/fs/file-nr

If you need to increase this value, add to the end of /etc/sysctl.conf something like that:

fs.file-max = 6545570

 

limits.conf

Add to /etc/security/limits.conf

nano /etc/security/limits.conf

In the end of file these lines:


* soft nproc 65000
* hard nproc 1000000
* - nofile 1048576
root - memlock unlimited

AOF fsync problem

When you get error:

#Asynchronous AOF fsync is taking too long

You have only one possible solution — get server with better disks performance. But if you can optimize your data, just do it 😉

UPD: Some interesting configuration changes

More information can be found here:

https://github.com/antirez/redis/issues/990

http://m.blog.chinaunix.net/uid-20776139-id-5739743.html

How Redis Took Us Offline And What We Did About It 

If you don’t worry about losing some data, look at redis config at these settings:

#no-appendfsync-on-rewrite yes
#appendfsync no

but be carefully and read documentation in marked comments: Redis stable config with documentation

Links

System Settings for server with high load
Sysctl.conf example with comments